I used to be a soccer fan, a long time ago, when my hair was less grey.

Now, I only look at some matches when a World Cup is on.

Which is today, right on the day of the SpaceX IPO.

Let’s take a look at a company that profits from these kinds of events, and who recently dropped 20% in price.

In this article, we are going to look at a company that plays a key and lucrative role in the industry that brings live sports events into the homes of billions of fans.

This company created its own ecosystem and built a trusted brand for more than a decade, with gross margins over 70% and net margins over 15% while reaccelerating revenue.

Whereas most businesses in sports are barely profitable, this company found its niche and is looking to transform itself to capture most of the biggest infrastructure upgrade cycles in decades.

Company Snapshot

• Company: EVS Broadcast Equipment

• Ticker: EVS

• Exchange: NYSE EURONEXT

• Industry: Communications Equipment

• Market Cap: $400M

• Founding year: 1994

• Gross Margins: 70%

• ROIC: 20%

• Guidance for 2026: $230M (11% growth)

Executive Summary: The architecture of live production

The value chain

EVS Broadcast Equipment designs and sells the hardware and software layers that facilitate the capture, management, and engagement of live video content. At its core, EVS solves the problem of instantaneous storytelling. It allows operators to capture dozens of high-resolution video feeds, synchronize them, and create a broadcast-ready slow-motion highlight or tactical review within seconds of the event. EVS is essentially the Adobe of live production, providing the essential, high-end toolkit that professionals rely on to turn raw footage into a finished masterpiece in seconds.

Let’s use the Champions League quarter-final between Bayern Munich and Real Madrid as a real-world example.

Rights owners: UEFA controls the rights to the Champions League and sells them to Broadcasters for billions. The rights are being auctioned ahead of every season.

Broadcasters and streamers: Sky, Amazon, and DAZN pay enormous sums for those rights and monetize via subscriptions and advertising. They then, most of the time, outsource production to…

OB van operators: Companies like NEP Group and Gravity Media own the trucks, hire the crews, and run the entire live production. They bill broadcasters per event or on long-term contracts. They are operationally complex with lower margins.

Broadcast Equipment: Unlike many technology vendors in the broadcast sector that occupy peripheral or replaceable segments of the value chain, EVS provides the fundamental recording, replay, and media-processing architecture for the world’s biggest live events.

Broadcasters are volatile, OB operators are margin-constrained, and EVS is the high-margin, mission-critical, locked-in supplier that wins regardless of who holds the rights.

As customers, we pay for a subscription and get to see these games live. We get to see dozens of camera angles on the pitch, which are the input data that EVS captures. Replay operators use EVS to edit the game while it’s still happening. Adding slow motion and choosing the best angles to show you exactly what happened. Once the game is over, TV stations around the world can instantly access the footage and show it to fans on their phones or in news broadcasts. It also powers the VAR Assistant.

Business Model

EVS segments its revenue into three categories:

• Live Audience Business (LAB): These are customers who have built their own infrastructure to manage their content, such as Sky, ESPN, or stadiums like Real Madrid’s Santiago Bernabéu. EVS can score higher margins in this segment through its recently launched SaaS model and its ecosystem bundle.

• Live Service Providers: These are clients like NEP Group and Gravity Media that own a fleet of vans and are contracted to broadcast events. They are the largest buyers of EVS hardware, often purchasing dozens of servers to equip a single large fleet of trucks at lower margins than LAB customers.

• Big-event rental: These are large contracts for events like the Olympics and the World Cup, where top-quality equipment is needed to ensure reliable broadcasting.

The growth story here is LAB customers who are increasingly broadcasting their own content.

Competitive Advantages

EVS positioned itself as the default operating system of live sports production. This positioning comes from deep product entrenchment, operator lock-in, and a proven track record at the world’s highest-stakes live events. EVS built a sticky ecosystem around its own hardware and software that a software-only player can’t match. As mentioned before, their products are on the same level as Adobe’s professional products, but in the broadcasting space.

They developed operator familiarity over decades. There are over 3,000 operators globally who have spent 10-20 years mastering the EVS remote. Switching from EVS to a competitor would require retraining the entire freelance crew, introducing unacceptable operational risk for the broadcaster. The maintenance and support renewal rate is exceptionally high, estimated at over 95% for Tier-1 broadcasters. This gives EVS pricing power, as it has an intangible asset that justifies premium pricing due to its high reliability.

Subscribe now

The SaaS Transition

Historically, software was a free or low-cost add-on to sell servers or hardware in general. Now the company is pivoting its value proposition from high upfront Capex with long-milestone payments to multi-year subscriptions with software at its center. This recurring revenue mode is expected to exceed 50% of the mix by 2026–2027. This is important for them because it means that they can decouple their base business from its cyclicality.

Strategic Roadmap: PLAYForward 2030

The SaaS Transition is also underscored by management’s stated “BHAG” (Big Hairy Audacious Goal): to be the #1 solution provider in the live video industry by 2030, with a revenue target of $350mn. To achieve this, they plan on executing their strategy via three pillars:

• From product provider to ecosystem:

  • from pure hardware sales (CAPEX) to software and SaaS models (OPEX)

  • from OB van operators to broadcast centers

  • from sports-only to entertainment, news, and digital

  • from a replay specialist to a live production ecosystem in multi-tier markets

• Strategic M&A Track Record:

  They recently started ramping up their M&A activity to accelerate their technology through:

  • Axon (2020) is now part of the MediaInfra division with its control (Cerebrum) and IP-gateway (Neuron) technology.

  • MOG Technologies (2024) was acquired to add to the MediaCeption portfolio, with its cloud-based transcoding and ingest capabilities.

  • Telemetrics & XD Motion (2025) is now the T-Motion robotics division, used to expand the ecosystem to include automated camera capture.

  Those businesses are expected to contribute around €30 million in revenue, accounting for 14% of total revenue in 2025.

• Geographic Strategy “Double Down North America”:

  North America is EVS’s fastest-growing region, contributing 37,5% of revenue in 2025. North America is the largest and most competitive market, driven by massive sports media rights deals (NFL, NBA, MLB) and a concentration of Big Tech (Amazon, Apple) entering the live arena.

Management

The founder is still on board as the CTO of the hardware department (the largest insider owner,>6%). The CEO is an industry veteran who joined in 2019. He pushed the PLAYForward strategy and the aggressive move into IP and North America.

The live broadcast equipment industry

The broadcast equipment industry is currently passing through its most significant structural pivot since the transition from analog to digital. Historically defined by proprietary hardware and quadrennial sporting cycles, the sector is evolving into a software-defined, IP-native, and AI-augmented ecosystem.

Addressable Market

EVS’s own estimated TAM is the best approximation for what the company could address since it does not compete across the entire live broadcast market. It targets tier-1 sports broadcasters and OB van operators for whom downtime is simply not an option. From where we are now, the company projects the TAM will double by 2030, which is ambitious but justified by their product and market expansion.

Demand-side

It is important to note that EVS’s demand has historically followed a power-law distribution of spend. The top 10 global broadcasters and sports federations (e.g., NBCUniversal, BBC, Sky, FIFA) account for more than half of enterprise spend on high-end production kits. This gave them bargaining power over EVS, but it also meant that they chose EVS because of an existing trust relationship.

Streaming Platforms Entering Live Production

In recent years, however, the entry of Amazon, Netflix, Apple, and DAZN into live sports rights has brought substantial fresh capital into the market. They spent a combined $12.5 billion on sports rights in 2025, representing 20% of the global total sports rights spend of $64 billion. These new entrants need to either build, rent, or partner to ensure a good broadcast experience for their customers. This structurally expands the addressable customer base for EVS and OB operators alike.

The best part is that EVS doesn’t care whether it’s Sky, Amazon, or DAZN buying the rights. As long as high-end live production is happening, EVS gets paid. And since it is the de facto default operating system of live sports production, all professionals needed will demand EVS’s products.

Supply-side

The overall market share for broadcast technology is fairly evenly distributed, with no player exceeding 12%, according to EVS’s own management. Historically, the sector struggled with declining revenues due to the decline of TV. This led to consolidation pressured by PE and the expectation of synergies.

Established Players

• EVS stayed independent and kept a tight focus on live production, which is their bread and butter. In recent times, however, they have begun to strengthen their share in other pillars, such as Infrastructure and Media Asset Management, through M&A. This is strategically important, as current trends demand an overall package with greater control across the value chain.

• Grass Valley was taken private by Black Dragon Capital in 2021. Grass Valley has emerged as EVS’s most direct competitor, with comprehensive end-to-end solutions and leadership in cloud platforms. The company competes aggressively by leveraging bundled pricing strategies in major sports contracts.

• Avid Technology was taken private by the PE firm STG in November 2023 in an all-cash deal valued at roughly $1.4 billion. It overlaps with EVS in the newsroom and Media Asset Management adjacency. Avid’s competitive pressure on EVS comes from owning the creative and editorial layer of the value chain that EVS is increasingly trying to reach through MediaCeption.

• Evertz Technologies is listed on the Toronto Stock Exchange (TSX: ET). Its centre of gravity is the broadcast infrastructure. Therefore, it competes with EVS primarily in the MediaInfra pillar rather than in live replay, making it an adjacent infrastructure specialist more than a direct threat to EVS’s core.

• Sony’s broadcast business sits inside the professional solutions arm of Sony Group Corporation, the publicly traded Japanese conglomerate. Sony R&D depth and balance-sheet scale can’t be matched by any pure-play vendor. Sony dominates cameras, sensors and image quality, but it has been deliberately building toward an end-to-end position that competes with EVS’s own ambitions to own more of the chain. The result is a “frenemy” dynamic with EVS because their systems are routinely deployed alongside Sony cameras.

A risk worth monitoring for me is the entry of PE capital trying to consolidate the market to increase pricing power, like in the case of Grass Valley. For EVS, this could result in margin pressure fueled by seemingly infinite money that distorts competition, as seen in the case of Chelsea Football Club and professional football. On the other hand, it could also result in a takeover bid for EVS with a premium between 25% and 40% over today’s price (Avid at 32% over the share price at that time).

Risks

EVS operates in a growing market driven by streaming, but is transforming due to the cloud. More live sports events and more streaming buyers expand the volume of production needed, but cloud-native production workflows threaten to gradually commoditize the hardware layer that EVS has historically monetized at 70%+ gross margins. They redraw the competitive map and hand EVS both its biggest long-term risk and its clearest strategic imperative. This transformation can be categorized into two simultaneous shifts.

The architectural shift: from on-premise to cloud-native production

The industry is moving past the binary choice between on-premises and cloud, with compute, switching, and playout occurringhappen in the cloud rather than in a physical OB truck. Grass Valley’s AMPP platform, AWS Elemental MediaLive and Adobe’s Frame.io integrations are the clearest articulations of this vision. For a vendor whose heritage is purpose-built hardware, this is the most serious threat as it questions not the product, but the premise.

The business-model shift: CAPEX → OPEX

Broadcasters and OB operators are under structural pressure to move away from large upfront capital expenditures, such as buying servers outright, toward operating-expense models. The shift is double-edged. Done well, it converts lumpy, event-driven hardware sales into predictable recurring revenue. Done poorly, it cannibalizes the high-margin hardware business before the software revenue stream is mature enough to replace it. The classic SaaS J-curve.

EVS’s answer: Balanced Computing

EVS’s response is a deliberate bet against the binary. Rather than choosing on-prem or cloud, it is positioning itself around a hybrid model in which high-performance on-prem hardware handles what genuinely requires it, while cloud modules handle everything else in a more elastic way.

Conclusion

The company’s bet is that ultra-low-latency live replay is the last function to move to the cloud, that the 2-second requirement for a Super Bowl replay is a hard physical constraint that keeps dedicated on-site hardware relevant for longer than skeptics assume. So far, the market has validated that bet.

Whether it holds through 2030 is the kicker question!

The opportunity - why the shift to IP matters for EVS

One of the most underappreciated drivers behind EVS’s growth is the industry’s migration from SDI (Serial Digital Interface) to IP (Internet Protocol) infrastructure. Think of SDI as dedicated cables each carrying a single video signal point-to-point, and IP as a flexible, software-defined network. For decades, live production ran on SDI.

Crucially, this is a structural transition over time, making it a long, staggered investment cycle for equipment vendors.

For EVS, that cycle reshapes what it can sell. In the SDI world, EVS sold replay servers, which it pioneered and still dominates. In an IP world, feeds, metadata, replay, routing, and control all run on a single network, and that complexity is an addressable market: EVS can now sell the control systems, orchestration software, and infrastructure layer that tie the whole production together. The prize is the shift from selling one box to capturing a far larger share of each customer’s infrastructure budget.

But IP is not pure upside. The same shift onto standard hardware and networks that expands EVS’s market also lowers the barrier that protected purpose-built broadcast hardware, letting software and cloud players into the same workflows.

For an investor, this is ultimately a revenue-mix story where EVS is trying to turn a cyclical, hardware-driven business into one with more software and recurring revenue. If it works, EVS re-rates from a hardware vendor toward a stickier platform business; if the hardware base erodes faster than software can scale, it becomes the J-curve risk.

The Financials - why EVS stood still for a decade

We use fiscal.ai for all our charting. Grab a 15% discount through this link.

GRAB 15% DISCOUNT

In EVS’s 2023 annual report, the company acknowledged a plateau in revenues since 2012. The cause was structural. Even though EVS dominated live replay servers, the market matured as the company only replaced outdated equipment. And for years, EVS simply failed to build a second pillar because its replay franchise was so dominant that it became a comfortable trap. To counter this inertia, the new CEO announced the PlayForward Strategy mentioned earlier.

On the margin side, it looks similar, as the company’s margins are in a slightly downward trend over a 10-year perspective. This is partly due to North American expansion, which increased SG&A costs.

The ROIC and ROE are moving on a similar plateau between 17-25%, which is similar to Evertz.

How much does EVS invest in the future?

EVS is, at its core, an engineering company where R&D is the real investment in the future and almost as big as SG&A. The business itself is high-margin, asset-light, and cash-generative, with excellent existing economics but historically mediocre incremental returns. This should change as EVS revenue becomes more predictable, as evidenced by order backlog visibility. Entering 2026, EVS has a record order book of €182.2 million. This provides coverage for nearly 80% of the 2026 revenue guidance (€220-240 million).

EVS pays a €1.10 dividend at a ~35% earnings payout ratio, while still accumulating net cash. This tells me that EVS has a hard time finding enough high-return projects to absorb them. That is the classic signature of a high-quality but low-reinvestment-runway business.

The counter-signal is management’s public ambition of reaching €350m in annual revenue by 2030. If they mean it, the reinvestment rate should rise from here, and that is the number to watch, not the current margin.

So, although the past looks bad and the future looks promising, is this company a BUY or not?

In the end, it’s all about valuation, as a great company can still be a bad investment.

So what is EVS worth?

Read more

Subscribe now

Today, we’re going to talk about edges. And why it’s important to find your own.

Bill Miller, a renowned value investor, mentioned in the past that an investor’s edge can be sourced in 3 areas:

1. Information

2. Your process

3. Your behavior

What I consider my edge can be broken down into 3 parts:

1. Choosing the right arena: Microcaps

2. Having a unique lens: Grounded in physics

3. Amplifying my entire process (ideas, research) through AI

This may surprise you, but I do not consider myself a growth, value, or quality investor. (although growth is always needed)

I don’t want to be locked inside a category.

I hunt for mispricings in the market.

Mispricings mean the market prices a stock below its intrinsic value.

Why?

• Because of an overall market crash, the easiest

• Because of a sector downturn: a bit harder → check our analysis of the Saaspocalypse

• Because of a company downturn, the hardest

Behavior is important, but information asymmetry is often overlooked.

Information asymmetry means:

• Having information that the market does not (very rare)

• Having information that the market only recognizes slowly (rare)

• Processing information in a unique way to create a different view (more common)

That asymmetry is most pronounced in the microcap space. Constructing a different view is easier thanks to my background in nuclear physics.

Let’s go through Bill Miller’s framework and how you can think about this for yourself.

The informational edge

Information is what we discussed in the introduction. The fact that you have information the market does not have, or has not recognized yet.

When Buffett was browsing through the Moody’s manuals, going through thousands of pages all those decades ago, he was exploiting an informational edge.

As you can see, a Moody’s page is not very sexy.

Imagine browsing through thousands of these.

In that time, you did not have an online broker. Quotes for stocks were displayed in the newsletter. If you wanted to buy a stock, you had to call your bank, and it took time.

Time flowed more slowly back then.

But going through those manuals, Buffett did something few others did. He took the time and effort to search for a needle in a haystack.

He created information asymmetry for himself.

And when he started out, he hunted for small companies, sometimes obscure, with low liquidity. Focusing on their balance sheet, buying companies at 25% of their cash value. Coca-Cola and Apple came later, when he had significant capital to deploy.

Buffett is a very smart man. And a shrewd investor. That’s without a doubt.

Last year in Omaha, I was impressed by how sharp he still is at 94.

But what he did in his early days had nothing to do with being smart.

• He chose a specific arena: Micro and small caps

• He hunted for information asymmetry: Going through the Moody’s manuals

And this asymmetry was even more pronounced when he became known as a smart businessman looking to buy companies.

He did not only have to hunt for ideas as we do.

Ideas came to him.

He got a deal flow that is not accessible to most investors.

That again is an example of information asymmetry.

The processing edge

The beauty (and difficulty) of the stock market is that 2 analysts, with exactly the same information, can come to different conclusions.

Making money in the markets is about creating a variant view of what the market has priced in.

And a variant view is created by having a unique process (or unique information).

Did Buffett have a unique process?

I don’t think so.

But what he did do was be very picky about what he was processing. He chose his arena carefully.

Before trying to be the best at a certain game, be mindful of what game you want to play.

He only looked at things he deeply understood.

By gaining a deep understanding, you gain conviction.

Through conviction, you withstand volatility in the markets.

That’s how you make money. It avoids you selling at lows.

So even if Buffett did not have a unique process, that doesn’t mean you can’t.

My process is grounded in physics. Yours could be completely different based on your background.

What helps is writing it down. (for me)

Buffett has never written down his process. (Others have tried in numerous books, though.)

The behavioral edge

If you’re a quant, all decisions are based on rules.

Buy on X, sell on Y.

I’m not a quant, and I am envious of their strong process. The human element is involved only in constructing the process itself.

So why don’t all quants win, even if their backtesting shows outperformance over multiple decades?

Because there will be years when a quant process underperforms, while the rest of the market does well. (you can see this in backtesting).

That’s hard.

Imagine your quant portfolio is down 15%, and the market is ripping.

You start doubting your process. Do I need to change it? And most investors will intervene, possibly breaking a great quant strategy, because they could not stomach a temporary loss.

Buying a position is easy. It’s the easiest decision one can make. Behavior still matters, though. Don’t chase a stock because of FOMO.

Buy because you did the work. Have a variant view of the market. No emotion.

But behavior is a much bigger factor when volatility kicks in. It comes down to the question:

When do you sell a position?

• My stock is down 30%, what do I do?

• My stock is up 100%, what do I do?

• My stock has done nothing for a whole year. What do I do?

The perfect setup is when your business keeps performing, and the stock price tracks performance perfectly.

But that never happens.

Many investors say their edge is taking a long-term view. Decades.

It’s easier said than done. And most companies don’t earn your attention for a full decade. Most falter. Only a few outperform.

Our solution to solve the behavioral problem is the 2 last two parts of our edge.

Our process, grounded in physics, aims to eliminate emotional decision-making. Like a quant, we want our process to be so strong that it becomes the primary reason for a decision.

Secondly, one of the prime uses of AI is to counter our behavioral biases. It’s our bullshit detector. AI is designed to conform to you. But you need to set it up so that it disagrees with you.

You can do some incredible things with Claude: build advanced screeners and automate portfolio tracking, but the number one use case remains pushing back on your ideas.

We’ve written about different behavioral biases in the past. Check them out in our archive.

How to find your edge?

Start by thinking about it.

Take a pencil and a piece of paper. Use the Bill Miller framework.

Now write down what you believe your edge is.

If you struggle at first, don’t sweat. It’s an iterative process. You write down what you know. And you come back to it once a year. Over time, your edge will surface.

One thing is important: Notice that our edge is very specific. We choose an arena. We use a unique lens. We state how we want to tackle our behavior.

If you write down: My edge is: ‘I think long-term’.

That’s not enough.

What actions are you taking to support that claim? Apply first principles. Try to break it down to its bare bones.

Get away from the abstract and write down specifics.

A unique edge could look like this. Let’s say you’re a medical doctor.

• Arena: We only invest in the healthcare sector

• Lens: Biology. You analyze companies with a unique framework. (There’s a great book: What I learned about investing from Darwin)

Every investor’s process is unique. It is based on past experiences and learnings.

The goal is to keep learning and refining your process every year.

It took me 15 years to find my edge. And maybe in 5 years, it will have changed again.

It’s a journey, not a destination.

So, a final question: Care to share your edge?

Leave a comment

May the markets be with you, always!

Kevin (Atomical Capital)

P.S.: Write-up almost ready on a new company + update on some of the holdings. Stay tuned. Got delayed a bit due to sickness.

Let’s take a look at all our positions in both portfolios.

Since we started tracking both portfolios at the start of 2025, we have a total return since inception of 35.3%, or, today, a CAGR of 19.8%. (versus our current 15% hurdle rate).

We are sitting on an outperformance of about 9% versus the S&P 500 since we started sharing at the start of last year.

Are we happy with that?

Yes.

It meets our hurdle rate, and based on our information today, we believe our companies will continue to create value over the coming years, leading to an even greater rerating of their market prices. (especially if our biggest position starts performing)

But humility is important.

Keep refining the process. Keep underwriting the positions.

I’m working on a couple of new ideas. I’ll also write about what I finally did with those 15 pitches from Omaha. Which ones survived my investing funnel, and do I plan to take a new position?

If you haven’t read those, you can find all pitches here:

Best pitches from Omaha part 1

Best pitches from Omaha part 2

Best pitches from Omaha part 3

Now, let me explain what happened with each company and what my plan is for each.

I use a tool we custom-built called Tracker to help me keep track of companies. It saves me hours each week. I talked about it in this article.

Here’s the analysis of why I think our portfolios are well-positioned for strong returns towards the end of the year.

All charts used sources from fiscal.ai

Read more

Find our initial thesis on NUAI here.

Price action was wild:

But we know this was going to be a bumpy ride.

A speculative bet, in a speculative stock.

So I ran my tracker agent to read the transcript and go through the business update. Then it compares this new information to our initial thesis.

It concluded: Thesis intact.

I then got the report and analyzed it.

Here’s what matters:

The good

54 acres of land

They entered into a binding agreement on the 54 acres of additional land they wanted to purchase. That doesn’t mean the deal is closed, but it’s on its way.

We flagged it as a “kill condition” in our thesis because, based on our understanding, it was an explicit demand from their hyperscaler client.

So this is a big milestone in the development of this project.

And just to show you how our tracker surfaces the kill conditions, it looks like this:

All other kill conditions remain on the table.

This company is a free sample. If you want to continue receiving these kinds of articles + deep dives and everything else, please consider taking a subscription.

Subscribe now

Phase 1: bypassing the ERCOT queue

This was already hinted at in a previous update. Their first 200 MW is not a behind-the-meter solution. That means they need a grid connection, but they remain vague on how this will be implemented: By using existing infrastructure.

This is an assumption, but the only way they can pull this off, and effectively bypass the ERCOT queue for grid interconnection:

1. Negotiate a power purchase agreement with Vistra or Calpine plants

2. Because they are adjacent: They can probably connect to an existing substation (and not directly into the grid)

3. The 54-acre corridor allows them to build this connection

No new power means no air permits needed for phase 1.

This makes everything a lot simpler.

Phase 2: 450 MW

They showed an updated picture of what the Texas data center would look like, and you immediately see what they plan to do with half of the new land they are acquiring:

It’s where the Behind-The-Meter power generation will come in.

Cash burn is manageable

Cash sits at >$80M as of April 30, confirmed on the earnings call. SharonAI note is fully prepaid.

Burn rate sits at $7 to 8M/quarter. The current runway extends well past the Oct 8 deadline (the loan covenant). Management states that current liquidity is sufficient for NUAI's phase one equity share without material near-term dilution.

Sharon AI: No problem

During the Sharon AI earnings call, they implied that the hyperscaler had walked away from the deal with NUAI.

The rebuttal from the CEO:

This is a thing from the past, and it was not the “hyperscaler deal”.

Sharon AI is now fully off the cap table. (paid back in cash).

Moving along.

The bad

Well, it’s not exactly bad. But the company keeps talking about “building a platform” and now building an 8.4 GW total campus.

And I admit, they did refrain from showing the atom, modular design in this recent business update.

I get it. Big Hairy Audacious Goals. (BHAG)

But in business, the saying is: Sell 1 dollar first. Then sell some more.

Sell that 200 MW first. Then sell some more.

Now, if you compare NUAI to other companies like IREN, APLD, etc., and assume they can actually reach that number, on a pure peer-to-peer comparison, you would be looking at a stock price between $50 and $100. (I considered a 17% of the total platform value because of their GP/LP structure).

The future

The most important task at the moment is to finalize the Joint Venture Agreement with Stream.

Why?

Here’s part of the transcript with Charlie Nelson, who was leading the business update:

This means that:

• A hyperscaler wanted to buy the land in Texas, but NUAI refused

• They agreed to maybe partner up, but only on 1 condition: they had to bring in Stream

That means a final contract with Stream is key to signing on the hyperscaler.

Here’s their slide:

It’s true that they are working on all three in parallel, but the thesis now becomes clearer:

• Agree on the final joint venture with Stream

• Get those power purchase agreements for 200 MW

The hyperscaler can come on board for phase 1.

The market is hungry for compute

I bought into NUAI because I believe powerful forces are at work, creating unique scarcity in the market.

• Companies are hungry for tokens

• But token prices are rising because of the lack of compute

• More computing needs mean more power requirements

If NUAI can pull it off, it would be an incredible feat. This is startup land governed by experienced executives.

So my read of the current developments is positive.

But make no mistake. This is still a speculative play.

The downside risk decreases with each passing month. But if that hyperscaler does not sign, this company will be dead in the water.

You’ll receive an update on all the other positions in the coming week.

May the markets be with you, always

Kevin

This is part 3 of the best stock pitches from Omaha.

You can read part 1 here and part 2 here.

Each company will have:

• A short description of what they do

• The type of company they are

• The pitch

Only 24 hours left!

If you hesitated about taking a paid subscription, we’re running a special Omaha sale for the next week. You’ll get 25% off an annual subscription. Prices of the annual subscription will go up permanently after the sale. This is your chance to lock in the current rate.

GET 25% OFF

Now let’s look at the last 5 stocks. The best ones are at the end of the article.

5 Wise- Ticker WSE on the Nasdaq)

We wrote a deeper article about Wise here.

Description: Wise is a cross-border payments company that moves money between countries for people and businesses more cheaply and easily than their banks.

Their customer: A Polish nurse works at a London hospital. She earns about £2,800 a month and sends £400 home to her mother in Krakow on the first of every month. She walks into her bank, asks for an international transfer, and they tell her it'll cost a £20 wire fee. What they don't tell her is that the exchange rate they're giving her is roughly 3% worse than the rate you'd see on Google. So, on a £400 transfer, she's actually paying about £20 in visible fees plus another £12 hidden in the exchange rate. Total costs: about £32 to move £400.

Now she tries the same thing with Wise. She gets the mid-market exchange rate (the real one, the one banks use among themselves), and pays one small upfront fee of roughly £2. The money arrives in her mother’s account in seconds.

Wise built its own payment rails instead of using the bank network. When HSBC sends money abroad, it goes through a system called SWIFT, which is basically a chain of banks passing your money along, each taking a cut. Slow and expensive. Wise instead set up real local bank accounts in roughly 10 countries and connected directly to local payment systems like ACH in the US and Faster Payments in the UK. When you “send money from London to Krakow,” nothing actually crosses a border. Wise just takes pounds from its UK account and pays out euros from its Polish account. The money doesn’t move.

This is the moat. Building those direct rails takes years and a banking license in every country. Once you have them, your unit cost per transfer is a fraction of what a bank pays. Wise then deliberately cuts prices every year, making it impossible for competitors to catch up. The take rate has fallen from 67 bps to 53 bps in a single year. Volumes grow faster than prices fall. Revenue keeps going up.

Type: Economies of scale growth machine

The pitch: Global cross-border payments are a £32 trillion market. Wise has roughly 5% of personal, 1% of SMB, and almost nothing of the £16 trillion enterprise segment. That last bucket is the real motherload. Banks still own it because they make fat margins on FX. It will take years to gain a foothold, but the runway is huge.

The competition isn’t really competition. Remitly is a consumer remittance app. Revolut is a super-app.

Here’s the data from fiscal.ai:

The FCF per share is larger than the revenue per share, but that’s because of the deposits.

Wise is deliberately under-earning today to build a payments network so cheap and so wide that nobody else can profitably catch up, and the market is punishing them for it.

4 Organo (6368 on Tokyo Stock Exchange)

Description: Organo is a Japanese water treatment company that builds the ultra-clean water systems chipmakers like TSMC use to wash silicon wafers.

Customer: To make a chip with circuits 3 nanometers wide (about 25,000 times thinner than a human hair), the engineers need water that is insanely clean. Not "bottled water" clean. We're talking water so pure it has almost no dissolved minerals, no metals, no microbes, and zero particles bigger than 10 nanometers. This is called ultrapure water (UPW). A single fab can use tens of thousands of tonnes of it a day.

As of FY03/25, Organo controlled an estimated 70 to 80% of TSMC's water treatment plants and roughly 100% of plants supplying fabs at 10 nanometers or smaller. TSMC alone is 14% of their revenue. Their core customer base is the world's most advanced semiconductor fabs, plus a long tail of pharma plants, power stations, and food refineries that also need very clean water.

They make money by building the water treatment system on site and then running it forever. This creates enormous switching costs.

A “boring” 80-year-old Japanese plumbing company that quietly became a chokepoint in the global semiconductor supply chain, paid first to build the water plant and then paid forever to keep it running.

Type: Slow growing but high operational leverage

The pitch: Organo has a dominant market share. Their margins have expanded over the last few years, but it’s not by accident. There are 3 reasons: the service business is growing as a share of revenue (now 40-50% of sales). Tight UPW supply lets Organo charge premium prices on new builds. High-margin specialty chemicals and resins are made in-house. This is why earnings have grown faster than revenue.

TSMC is building fabs in the US (Arizona), Japan (Kumamoto), and Germany. Organo is following them. Their US business is being built out alongside TSMC's CHIPS Act fabs. The pharma side, especially injectable drug water, is also accelerating.

This is a strong company with a moat, where EPS is growing faster than revenue. The market has taken notice and priced it higher.

Subscribe now

3 Samsara (Ticker IOT on NYSE)

Description: Samsara is a software and sensors business that helps trucking, construction, and field service companies see what their trucks, equipment, and workers are actually doing.

The perfect customer: Meet Maria. She’s the dispatch manager at a 90-truck regional trucking outfit in Memphis. Her job is to know where every truck is, whether her drivers are tailgating or falling asleep, when the next oil change is due, and why fuel costs jumped last month. In practice, she has a whiteboard, a stack of paper logbooks, and a phone that won’t stop ringing. When a truck misses a delivery, she calls the driver. When the driver doesn’t pick up, she calls his wife. That’s her tracking system.

Last year, she tried calling one of the big incumbent telematics vendors (think Verizon Connect, Geotab, Trimble). They told her: Six months of implementation and a six-figure quote. And software that still required three separate screens to locate one truck.

Samsara is a subscription software business.

• A trucking or construction company signs a 3-year contract.

• Samsara ships gateways, AI dashcams, and tags. Each device plugs in or sticks on in minutes.

• The hardware streams data to Samsara’s cloud-based software.

• The customer pays roughly $27 to $33 per vehicle per month for basic GPS, climbing to $40 to $60+ once you add the AI dashcam and driver coaching.

What’s interesting is the per-application, per-asset pricing.

If a customer has 50 trucks and wants both GPS tracking AND the video safety camera, that's 100 subscriptions, not 50. Every new product Samsara ships (equipment tags, site cameras, maintenance, workflows) is sold as a separate per-asset line item to the same fleet. Over 95% of their $100K+ customers buy two or more products, and that cross-sell is the growth engine.

Here’s a chart you don’t see every day. Revenue and free cash flow per share are growing, and earnings yield to. That means the stock price is dropping while the business seems to be getting stronger and stronger. (stock is down 40% over the past year)

Type: Fast grower

The pitch: The market puts Samsara in the SaaS bucket and judges it on SaaS multiples. That's the mistake. Samsara is not a SaaS company. It's the company digitizing the physical operations of the working world, and the SaaS label undersells both what they do today and what's coming next.

Once Samsara's hardware is wired into a customer's 10,000-truck fleet or bolted onto 200 construction job sites, switching vendors is not a software decision. It's ripping off the most sticky band-aid, breaking compliance reporting, driver coaching workflows, insurance arrangements, and dispatch routines.

The next bet is AI for the physical world: robots, autonomous trucks, smart job sites, and electrified fleets. Samsara processes over 25 trillion data points a year through GPS telemetry, AI dash cams, and sensors, the kind of dataset you cannot scrape from the internet.

They have seen strong performance over the past years, but missed their last quarter (revenue, not EPS!).

If you want full access, consider the current 25% discount to continue reading.

Let’s dive in:

Read more

This is part 2 of the best stock pitches from Omaha.

You can read part 1 here.

Each company will have:

• A short description of what they do

• The type of company they are

• The pitch

Only 48 hours left!

If you hesitated about taking a paid subscription, we’re running a special Omaha sale for the next week. You’ll get 25% off an annual subscription. Prices of the annual subscription will go up permanently after the sale. This is your chance to lock in the current rate.

GET 25% OFF

Now let’s look at the next 5 stocks. The best ones are at the end of the article.

5 Plover Bay Technologies (1523 on the Hong Kong stock exchange)

Description: Plover Bay is a Hong Kong hardware-and-software company that makes routers and subscriptions for customers who need their internet to never go down.

Their customer: A firefighter captain rolling out of the station in Los Angeles during the January 2024 wildfires. He needs maps, weather feeds, a live video link to dispatch, and a connection to the hospital if a firefighter gets hurt. The cell tower three miles away is on fire. The cell tower behind her is congested because half the city is calling family. His satellite dish works, but only when the sky is clear of smoke.

He calls a normal enterprise networking vendor like Cisco. Cisco sells beautiful gear for a corporate office. They don’t really sell a single box that takes one Verizon SIM, one AT&T SIM, one Starlink dish, and one satellite phone, and welds them together into a connection that just keeps working when any one of them dies.

They use a land-and-expand business model.

Three revenue lines, stacked on top of each other.

1. Sell the box. A router with multiple slots for SIM cards, ethernet cables, and satellite uplinks. This is the chunky upfront purchase. Hardware was about 71% of revenue in 2024.

2. Sell the subscription. The router’s magic only works if you pay an annual fee for SpeedFusion (their bonding software, which ties multiple internet links into one) and InControl (cloud software that lets you manage a fleet of routers from a dashboard).

3. Sell the warranty. Annual support contract. Renewable.

It’s a capital-light business as all manufacturing is outsourced.

Type: Steady growing cash printer

The pitch: It's not that this company keeps growing earnings by 25% each year. Or that it gives back most of the cash through dividends. All those things are true.

The pitch is that almost 60% of their business comes from the US, and they have announced a spin-off to list on the Nasdaq in 2026.

Compared to US peers like Ubiquiti or Digi International (with much lower margins), that could lead to a 50% upside just from the rerating.

Here’s the overview: look at the FCF margin!

The spin-off is on track to be completed towards the end of the year.

As you might have noticed, I use the same charts from fiscal.ai in every pitch. All per-share metrics are combined with the forward earnings yield. You can set this up once by creating a metric template:

Fiscal is running their annual 25% discount. You can use this link to get access.

25% OFF FISCAL.AI

4 Tasmea (TEA on ASX)

I wrote about Tasmea in a previous article.

Description: Tasmea is an Australian industrial services group that fixes, maintains, and upgrades the equipment inside big mining and utility plants for the giant companies that own them

Customer: A plant manager at a BHP iron ore site in the Pilbara, a remote corner of Western Australia. Her plant moves rocks, crushes them, separates iron ore, and loads it onto trains. The plant runs 24/7. Every hour it sits idle, the company loses about $500,000 in shipments. One morning, a giant electrical switchboard fails. She needs specialist high-voltage electricians on a plane today, with the right safety credentials, the right insurance, the right relationship with BHP’s procurement team, and the willingness to live in a camp 1,500 km from anywhere.

She calls a big generalist engineering firm. They build new plants. They don’t fly six tradespeople into a desert tonight to fix a switchboard. She calls a local electrician. He doesn’t have the high-voltage tickets or the BHP-approved safety system. She calls a global services giant. They want a 12-month contract negotiation. None of them are built for what she actually needs: specialist trades on call, sleeping in a camp, fixing the plant by morning. That’s the customer Tasmea was built for.

The actual customer base is what Tasmea calls “blue-chip essential asset owners”: large miners (iron ore, gold, copper, lithium), electricity transmission companies, water utilities, and oil and gas operators. In FY25, around 94% of Tasmea’s revenue came from repeat customers, and a single customer accounted for about 24% of total revenue.

The business model is simple: It’s labor arbitrage. Tasmea bills out skilled tradespeople (electricians, boilermakers, riggers, fitters) at a markup over what it pays them. That alone is a normal staffing business.

What makes Tasmea different is this:

The customer doesn’t care what Tasmea costs; they care that the plant keeps running. When a $4 billion iron ore mine is losing half a million dollars an hour, nobody is haggling over the electrician’s hourly rate. That’s why Tasmea earns roughly 25% gross margins and 15% EBITDA margins, while peer construction and engineering firms earn closer to 10%. Combine that with 70+ Master Service Agreements (multi-year contracts that lock Tasmea in as a preferred supplier), and you get a record $600 million secured order book heading into FY26.

Type: Roll-up (serial acquirer)

The pitch: It’s not complicated. You’re buying a high earnings growth roll-up (30%/year) in a boring industry at a forward P/E of 19. Tasmea profits from an electrification tailwind in Australia. Their electrical segment grew 66% in 2025 and contributes 44% to net profit. Management guides towards 110M in NPAT by 2028-2029. At today’s multiple, that’s a 3.7B company compared to a current multiple of 1.5B.

You can see from the chart that revenue is growing fast. Their capex seems low, but that’s because many acquisitions are cash, not reflected in the capex number. So the actual number per share is double or triple what is shown in the chart.

I like Tasmea; I liked it even more 2 months ago when I first wrote about it 😉 (when the price was 50% lower). But the company keeps performing, and more and more analysts have taken notice, which explains the multiple expansion.

Subscribe now

3 Amadeus IT Group (AMS on Bolsa de Madrid)

Description: Amadeus IT Group is a Spanish travel technology company that runs the booking plumbing between airlines, hotels, and travel agencies.

Customer: TAP Air Portugal. A medium-sized airline with a decent network out of Lisbon, mostly flying Europeans to Brazil. They want their flights to show up when a travel agent in Singapore types “Lisbon to Rio” into a screen. They also want to be bookable by a corporate travel desk in Frankfurt and an online travel site in Dubai.

So they call Singapore. “Hey, can you integrate with our reservation system?” Singapore says no. Singapore already has 700 airlines they could integrate with one by one. Nobody is building 700 custom phone lines. Dubai says the same thing. Frankfurt too.

That’s the customer Amadeus was built for.

Airlines plug into Amadeus once. The moment they do, they are bookable by around 190,000 travel agencies and online sellers worldwide. The same model works in reverse for the travel agents.

Amadeus runs three businesses, and the mechanism is different for each. Total group revenue in 2025 was €6.5 billion.

1. Air Distribution (€3.1 billion in 2025). This is the GDS, short for Global Distribution System. A giant electronic marketplace. Airlines list their seats. Travel agencies search for them. When the agency books a ticket, the airline pays Amadeus a fee per booking, roughly €5-€6 per ticket. The travel agency pays nothing. Amadeus often pays the agency to use Amadeus instead of a competitor. Those payments are called incentives, and they are a major cost line item.

2. Air IT Solutions (€2.3 billion in 2025). This is software-as-a-service for airlines. Amadeus runs the airline’s reservation system, inventory, check-in, and boarding. Their flagship product is called Altéa. The airline pays a fee for every passenger boarded. If Altéa goes down, the airline cannot fly.

3. Hospitality and Other Solutions (€1.1 billion in 2025). Same model, but for hotels. Marriott and Accor both pay Amadeus to run their central reservation systems. Add some payment processing and travel data on top.

This is a two-sided network. The more airlines on Amadeus, the more useful it is to travel agents. The more travel agents on Amadeus, the more useful it is to airlines. Sabre and Travelport are the only real competitors. That is the whole market. Three players. The switching costs are enormous. When Ryanair runs on Amadeus’s low-cost-carrier system Navitaire, ripping it out and replacing it is a multi-year, multi-million-euro project that nobody wants to greenlight. So they renew. Ryanair just renewed for another stretch, marking 25 years on the system.

Type: Monopoly in a 3-player oligopoly (mature growth)

The pitch: Amadeus has a 40% market share of global airline bookings, 20%+ net margins, zero net debt risk, and multi-year recurring contracts with the world's biggest airlines and hotel chains.

It is the picks-and-shovels play on global air travel growth (which has outpaced GDP for 50+ years), it grows revenue mid-to-high single digits, returns cash via dividends and a €1.3bn buyback program, and it just signed Marriott, Accor, British Airways, Air France-KLM, and Lufthansa Group to long-term deals. The stock is down 30% over the past year, trading at the lowest multiple in a decade.

The stock is down on AI fears. But they seem to be overblown.

Booking a flight is not the same as searching for a recipe. It requires real-time inventory, fare rules, seat maps, payment processing, and PNR creation (PNR = the booking record). AI agents that try to book directly still need to hit somebody's infrastructure. That infrastructure is Amadeus, Sabre, or the airline's own system (which is often, you guessed it, Amadeus's Altéa).

Amadeus has already signed British Airways, Air France-KLM, and Lufthansa Group as launch customers for Nevio, their AI-native airline retailing product. So far, the data says AI is making them more relevant, not less.

The beat expectations on their latest Q1:

The bet is achieving about 15% annual growth in earnings, dividends, and buybacks. If the stock reverts to its prior multiples, you’ll earn more.

Now let’s get to the final 2 on this list.

If you want full access, consider the current 25% discount to continue reading.

Let’s dive in:

Read more

I’m a sucker for improving my workflows and testing new tools.

Here are the tools we use in our investing process.

Let me know in the comments what your tech stack looks like!

Leave a comment

Our stockbroker: Interactive Brokers

The number 1 question we receive: It looks like a great company, but I cannot buy it with my current broker.

That’s where IBKR comes in.

It allows us to invest easily and at low cost worldwide. We trade stocks in the US, Canada, Europe, Asia, Australia, and the UK.

It’s the only broker I know that has this breadth of possibilities.

GO TO IBKR

Our financial software: Fiscal.ai

I wrote about fiscal.ai in depth here.

I think there’s about 8 hours left to grab their 25% discount if you’re interested.

We have been using fiscal for several years now.

We mainly use it to quickly have a first look at a company in the first hour of our due diligence.

Some advantages of fiscal.ai I really like:

• The main dashboard gives you a snapshot of a company

• Creating custom financial templates (we like the per-share metrics)

• Create custom KPIs

• A solid visual interface

GET 25% OFF FISCAL

Our favorite AI LLM: Claude.ai

We use Claude Code to build tools and analyze investments. You can find more details about how we use and master Claude in our second newsletter, "Invest with AI."

• Claude Chat and projects can be used for analyzing a stock with the right prompts and data → See an example here

• Claude cowork and skills can be used for repeat work and analyzing a series of companies at once → See how that is done here

• Claude code can be used to build custom tools, often Python scripts with a financial API that can do a lot of work for you on auto → See how that is done here

GO TO CLAUDE

Our financial data API: EODHD

EODHD provides the backbone for our personal investing tools. It provides up to 30 years of worldwide financial data for over 100,000 public companies. The value-to-price ratio is unparalleled in the data it provides.

You can check it out here.

Easy money transfer: Wise

Not only are we a shareholder, but we have also been using Wise for multiple years, since we trade in multiple currencies worldwide.

When you use the Wise app for a cross-border money transfer, and see how seamless and transparent it works, you really wonder what the legacy banks are doing.

Do they want to get disrupted?

If you’re interested, this link allows you to quickly set up an account:

That’s the main tech stack. Seems quite simple, right?

What’s yours? Is it more complex? Why?

Leave a comment

May the markets be with you, always!

Kevin

Note: I do have a Mac mini, which runs multiple jobs and manages my second brain. But I would need multiple articles to explain how this is set up. It’s not that simple 😉

And I only recommend tools to others that I actually like and use. (full tech stack incoming in about a week).

So here are 5 incredible features why I think fiscal.ai should be part of your stock research workflow.

Note: Fiscal is running their 25% off sale. It ends tomorrow. So if you want to grab a discounted subscription, you have 1 day left to profit from their promotion.

25% OFF FISCAL

1 Financials: Custom Metrics

Fiscal.ai allows you to create your own custom metrics. Each business should be viewed with a different lens.

Take a business like MELI for example where most investors say the name in the same breath as Amazon. (Meli is a different beast though).

But one thing they have in common: The business is not optimized for profits. Management invests and takes a long-term approach.

The result is that a combination of popularity, and not optimizing for profits, make it so that their P/E multiple is always high.

But is the P/E ratio the right ratio to look at?

Meli combines online retail with finance. It invests for the future. Net income might not be the right metric. You could consider cash flows, but same thing. Free cash flow will vary with investments and be distorted by cash deposits on the balance sheet (which you can’t count upon).

Also, you’re not taking into account the growth of the company. Static, trailing ratios have no real use when you have a company growing at 30%+ each year.

So here’s what fiscal allows you to do.

Create a custom ratio.

This ratio is the EV/GP/G ratio.

• Correct the Total Enterprise Value for the cash deposits

• Divide by the gross profits

• Divide by the gross profits growth (I took 1 year, but you can take 3 years)

When we plot this in fiscal, we get this (blue line):

• The P/E ratio tells you, it’s still high, but the cheapest ratio at which MELI has ever traded

• Our EV/GP/G ratio tells you it’s priced back as in the downturn in 2022

We took gross profits as so much is going on below that line in the income statement, that it’s the “cleanest representation” to the past.

So we love this: Custom Metrics

2 Business owner mode

When we buy companies in the public market, we should approach it as if we’re buying a private company.

When you’re interested in buying a local company, you don’t know what the price is at the start of your due diligence. You cannot be influenced by the price.

That’s not the case in public markets.

But fiscal has a solution: owner mode.

Here’s the normal view you get when you select a company:

Now when you select “owner’s mode” in the settings.

You get the following view:

Metrics related to pure price are masked.

I don’t know which other tool can do this.

Subscribe now

3 Awesome Customer Service

They still have real customer service.

Why real, because yesterday, I was battling with an AI Chatbot from Paypal and wanted to shoot myself in the head. (stuff for another discussion, but creating a business account on Paypal was an absolute nightmare.)

And when I use my Custom Metric “per share” template

This is what fiscal shows me:

The past metrics don’t look that bad. I in fact you have growth at 10% in diluted EPS, but earnings yield growing at 15%. So the business has become more profitable AND “cheaper” at the same time.

But cheaper doesn’t mean cheap.

Forward earnings yield sits at 12%, but Paypal seems to have a growth problem. These are quarterly results. Look at the Q1 FCF/share result for each year:

• Q1 2024: 2.3

• Q1 2025: 2.2

• Q1 2026: 2.3

But i’m straying of course. That’s not what I wanted to talk about.

I had a question about using those custom metrics, and within 12 hours, I received a reply by mail:

And it seemed it was me who was at fault by the way. I was using it wrong.

The cherry on the cake.

Even the CEO replies to questions.

This is not an invitation to start asking questions like this to the CEO. I admit, I was a bit too forthcoming 😉

4 Breath of data

We invest in microcap companies worldwide. That means we want worldwide data.

We are invested in the US, Canada, Australia, the UK and Japan.

And even in the UK, on the small AIM market, fiscal has the data.

It has never let us down in the breath of data it has. Over 50,000 companies (and ETFs- worldwide).

5 The one thing that bothered me

When I use software, I want speed. It needs to feel snappy.

Even waiting several seconds for a company to load is too much for me.

And in the past, fiscal.ai was not the fastest.

But again, they listened to their customers.

You can try out their free plan, and test it for yourself.

Conclusion

This is just a small glimpse of what fiscal has to offer.

Their 25% off sale is ending in about a day.

You can still profit from their offer. And if not, which tool are you using and why?

Leave a comment

May the markets be with you, always!

Kevin

Atomical Capital

Note: Full disclosure, the link above is an affiliate link. If you decide this software to be useful and buy a subscription with fiscal, that will help this newsletter.

I made a selection of the 15 I consider to be the best.

Since I want to provide you with a first analysis, the 15 best pitches will be split into 3 parts, as otherwise this article would become way too long.

Each company will have:

• A short description of what they do

• The type of company they are

• The pitch

I’ll need to do more due diligence on several of these names, but I think it’s worth getting them to you as quickly as possible. (After all, I did have 2 stocks go up over 100% while sitting in my research pipeline. So I do plan to publish ideas more quickly, so you might benefit from them.

I pitched our last deep dive, which had a great quarter and is up 20% since March.

If you hesitated about taking a paid subscription, we’re running a special Omaha sale for the next week. You’ll get 25% off an annual subscription. Prices of the annual subscription will go up permanently after the sale. This is your chance to lock in the current rate.

GET 25% OFF

Now let’s look at the first 5 stocks. The best ones are at the end of the article.

5 AMCORI Group (AMCO on AIM in the UK)

Description: Amcom buys small, boring British engineering companies from retiring owners, makes them run better, then bolts more of them together.

Imagine a 68-year-old guy in Sheffield who runs a valve repair shop. He started it in 1987. Forty employees. Nice profits. His customers are power stations and oil refineries, and they’ve been buying from him for 30 years. He wants to retire, but his kids don’t want the business. He’s too small for private equity. Selling to a competitor would gut his loyal team. He’s stuck.

That’s the seller Amcomri targets.

Type: Tiny serial acquirer

The pitch: They have a long runway to reinvest capital profitably (the ability to buy small, family-owned businesses with little competition at less than 5x EBITDA). They have a very experienced management team with deep operational and deal-structuring expertise. The underlying businesses are resilient and mission-critical (high recurring revenue, low customer-concentration risk, less susceptible to AI disruption).

Now, fiscal has a great feature that shows revenue per share along with earnings per share.

Which means, as you can see, despite overall revenue growing, revenue per share dropped, indicating they issued capital to finance acquisitions.

Fiscal.ai has a revenue and earnings beats/misses view now. We’ll use that for each company. If you’re interested in a 25% discount on a fiscal.ai subscription, they are running a sale at the moment.

Here’s what it looks like for AMCO.

I think this could be an interesting idea. But I would need to dig a lot deeper.

The price of the stock has seen only strength, which I like a lot:

The bet for this company would be:

• The founders keep finding deals at 4-6x EBITDA over the next 5-10 years

• Their own equity multiple stays at 9-10x or rerates higher

• They avoid blowing up an integration

• The UK SME succession wave keeps producing supply

• They eventually scale to £30-50 million EBITDA, at which point institutional money piles in and rerates the stock

So, despite the long runway, the only potential problem is that a lot has already been priced in.

Now let’s go from a microcap to a bigger company.

4 Primoris (PRIM on NYSE)

Description: Primoris is the company that physically builds and maintains the boring but essential infrastructure that keeps America running. The high-voltage line, the gas main, the solar farm out past the highway, the long-distance oil pipeline crossing Texas. None of that installs itself.

They are a specialty contractor: they don’t own the wires or pipelines, they get hired to put them in the ground and keep them working by utilities, oil and gas operators, solar developers, and state highway departments.

The business splits into three buckets.

• Utilities is the steady recurring half (gas, electric, and telecom distribution, mostly under long-term Master Service Agreements).

• Energy/Renewables is the growth half (utility-scale solar, battery storage, refinery work, and highway projects for state DOTs).

• Pipeline Services is the legacy half going back to the company’s 1960 roots as a California pipeline outfit (big oil and gas trunk lines, pump and compressor stations).

Although it’s a bit big for my taste, with a market cap of almost 6B USD, I wanted to take a look at this as a data center infrastructure play. And since the company missed its quarterly estimates, the stock dropped 40%.

I mean, it went from a 12B USD market cap to 5.8 in a matter of days…

Now, before jumping into this stock, it did trade at a forward P/E of 38 before the drop, which is staggering for an infrastructure services company. They missed their renewables segment and had to cut their year-end guidance.

And we’ll need to dig, but it has almost always beaten expectations.

Now, what’s great about these kinds of companies is that their backlog gives us some visibility into the future.

From their latest report:

About 5.2 billion of that management states should be converted into revenue over the next 12 months, which sets a floor for upcoming revenue. (The Master Service Agreement part) (compared to 7.6 billion in revenue in 2025).

The pitch: The infrastructure supercycle is driving massive spending in the utilities and energy sectors; Primoris has an enormous backlog. There is a significant uptick in energy prices, datacenter construction, and electric utility projects. Their proximity to Texas energy companies and long-term relationships with utilities.

So the final question for this company is: Is their renewables setback a one-off? If so, and they resume growth in 2027, you could buy this company at an attractive price.

Based on the new guidance from management, the stock is trading at a forward P/E of 24.

Subscribe now

3 Lesaka Technologies (LSAK on Nasdaq)

Description: Lesaka is a South African fintech for people and small businesses that regular banks ignore. In townships across Southern Africa, most commerce happens in spaza shops, taverns, and gas stations. The owners and their customers mostly never had bank accounts. They live on cash. Cash gets stolen and can’t be tracked, so no bank will lend to them. Lesaka fills that gap from both sides.

For shop owners, it installs card terminals, sells software to run the business, and lends working capital based on the cash flowing through the till.

For consumers, it offers a bank account, small loans, and affordable insurance, such as funeral cover.

Lesaka earns a cut on every card swipe, interest on every loan, a premium on every policy, and a margin on every prepaid airtime and electricity top-up.

Previous earnings and revenue beats/misses from fiscal.ai.

And revenue and earnings per share per quarter (as this company had to raise capital to acquire other companies)

Type: Fast grower. Inflected Earnings.Consolidator.

The pitch: Lesaka is a Fintech displacing cash transactions in South Africa, serving the underbanked with high incremental margins. It just shifted to positive earnings based on their earnings report on the 7th of May. Based on expected earnings in 2027, it sits at a forward P/E ratio of 14, which is lower than that of emerging-market peers.

However, this forward P/E is based on non-GAAP EPS numbers (see figure below). Because Lesaka has made several acquisitions, amortization is not included.

3 forces working together could make this an interesting setup:

1. The amortization will naturally increase GAAP EPS over time (but management must refrain from another big acquisition). But the business is framed as a consolidator, so they will keep acquiring companies.

2. The company’s market cap is still small. As the market cap increases, Emerging Markets funds can come in, and this will have a positive impact on the multiple

3. The business growth itself, where management guides towards 30% operating margins (from the current 21%)

Final note: The local devaluation of the currency (the ZAR) will reduce the final ROI, but if the management base case holds, it should be more than offset by the company’s growth.

Now we get to the best. The ones I regard as the highest potentials for multibagger returns.

If you want full access, consider the current 25% discount to continue reading.

Let’s dive in:

Read more

I consider myself lucky.

Lucky to travel to Omaha for the second time, accompanied by friends/stock nerds. Some of them also run stock investing newsletters like from Potential Multibaggers and Pieter from .

It was my second time in Omaha, and even though Buffett no longer runs Berkshire Hathaway, and the overall turnout was lower than last year, the ecosystem around it is still thriving.

Lots of events. Stock pitches by some of the best investors I know. AI workshops by startup founders from Silicon Valley. And of course, hanging out in the evening with friends.

And the serendipity is strong.

During the day, you randomly bump into some juggernauts in the industry, like Bill Ackman:

My goal is always the same for these kinds of events:

• Surround myself with super smart people

• Try to learn as much as I can, be a sponge

• Apply those lessons to my stock analysis (and life)

The most important lesson I learned from nuclear engineering, when applied to stock investing, is to always stay humble. Always keep learning.

If you know me, you know my favorite slogan:

I never lose. Either I win or learn.

- Nelson Mandela

Let’s get into some of the highlights of the past week

(you can find the best stock pitches in a separate article in the coming days).

Subscribe now

Lesson 1: It’s about the people

Imagine you’re like my son. You collect Pokémon cards.

One of his favorite pastimes?

To go to a Pokémon con to talk and interact with like-minded people.

Buffett and Charlie have created something similar.

And it’s not just about investing in stocks.

It’s about values. About integrity and respect.

40,000 people don’t show up to hear what the holding company Berkshire Hathaway is all about.

They want to seek out the present-day philosophers. They want to feel connected to like-minded people.

And that’s why I believe the introduction video they chose at the beginning of the AGM was so fitting. To emphasize that even though Buffett is no longer at the helm, those values and virtues will be kept alive.

I’ve watched this video multiple times but those last 10 seconds still give me goosebumps:

Lesson 2: Maybe there is an AI bubble forming

As you know, I’m knee-deep in learning how to use Claude for stock analysis. I created a whole new newsletter dedicated to it.

A workshop was organized by 5 university graduates from Munich, who declined high-paying consultancy roles and moved to Palo Alto in San Francisco. There, they got sponsorship from Andreessen Horowitz. So now, they spend 16 hours a day getting up to speed on AI and figuring out which startups to launch.

When I asked them what the current bottleneck is to launch a great startup within the AI space, the answer was:

• Raising capital is not a problem (VC’s back the team)

• The bottleneck is finding great problems to solve

AI allows us to easily build software. In a world where everyone can code, the most important question is:

What should one build?

The fact that they have no problem raising a couple of million USD to launch a startup, even without a clear idea, smells like bubble territory.

This workshop was incredibly inspirational. I will write more on this topic in a future article over on “the other newsletter”.

Lesson 3: Even AI disruption will take time

I attended a fireside chat with none other than Bill Ackman as the main speaker. Pershing Square recently did an IPO. They now have 3 vehicles you can invest in. (PSUS on the New York stock exchange, PSH on the London Exchange and PS on the NYSE)

The fireside chat was organized by UBS, and quickly became a Q&A between the audience and Bill, who was accompanied by his Chief Investment Officer, Ryan Israel.

I must admit, I did not expect it, but the quality of the answers they gave blew me away.

Pershing invests in quality companies that they can grab at a great price following some kind of problem. Some examples from the past:

• They bought Chipotle after the food contamination crisis in 2016

• They bought Alphabet in 2023 when it dipped due to AI fears

• They bought Restaurant brands in 2020 when COVID struck

And for just a moment, I thought: Maybe I should buy Pershing Square 😉

But I think I’m sticking with our microcaps.

What I learned the most was their holistic approach to investing. When asked about AI and how fast it would disrupt all industries, Ryan mentioned a historical study by Carlota Perez, who wrote: Technological Revolutions and Financial Capital. (I ordered it on Amazon)

Carlota’s delve into history shows that, over and over again, total disruption will take 20 to 30 years to achieve full market penetration.

Especially physical AI, (confirmed by the AI workshop I attended) will take a lot more time to develop.

Physical AI: AI systems that operate in, interact with, and understand the physical world, moving beyond software-only tools to control hardware via sensors and actuators

The main reason: It’s a lot harder to gather quality training data, and create the feedback loop of continuous improvement within physical AI. Basically, you need to put a robot to work, measure everything and feed it back into the system (like Tesla FSD)

Lesson 4: The time to live is now

Vitaliy Katsenelson, from the intellectual investor, invites everyone for breakfast each year.

More than 1,000 registered.

He takes about 2 and a half hours to do a Q&A where people can ask questions about investing, life, and everything in between.

His responses are thoughtful, and more importantly, when he doesn’t know, he says so.

Some of the highlights I picked up: (and I’m using the Feynman technique here to simplify, I hope I’m not butchering Vitaliy’s responses…)

• When asked about how to achieve financial independence, remember that life happens now. Not after you’ve hit your number

• How to raise your children when you have money, so that, as a man said, “They don’t become pricks”. Use mindful scarcity. Scarcity leads to struggle and innovation. It helps people grow.

• On when to sell: The better the company, the more optionality it has, the less you should be inclined to sell. It’s difficult to create proper DCFs for these kinds of companies because of that optionality.

• What he likes about Watches of Switzerland (Ticker WOSG). WOSG looks like a retailer, but it’s not. It’s a second derivative bet on Rolex. Rolex determines whether you can add an additional store. Rolex’s job is to maintain the scarcity of its product. Half of the sales are from the UK, the other half from the US. But based on population, if the US matched the UK’s sales, revenue would increase 6-fold.

For a quick research report, just use the research function within fiscal.ai

Oh, and, if you’re interested in fiscal.ai, they offer an exclusive discount of 25% at the moment.

Lesson 5: The time to live is now

Important lessons merit repetition.

And I repeat this one because I often forget it myself.

When you’re passionate about stocks and building new things, you sometimes lose yourself in the frenzy of work.

We did our best to enjoy our stay in Omaha and live in the moment.

This closing picture is a testament to that:

So now that this article comes to its end, do one thing for me.

Close your computer, or put down your smartphone.

Hug your children and family. Go grab a bite to eat with a friend.

Enjoy the moment.

We’ll talk about stocks again in the coming days.

May the markets be with you, always!

Kevin

Atomical Capital

Because of the busy schedule over here, I won’t be writing my usual articles this week.

So I figured it’s time to add some much-needed structure to what already exists.

It has taken more than 2 years to write it all, but Substack is not the best for easily navigating it.

This page will help you with that!

Articles with a * are paid. If you want access to the full library ⬇️

Subscribe now

Categories: Start Here · Portfolio Performance · Stock Deep Dives · AI Tools for Investors · Investment Ideas · Frameworks · Lessons from Great Investors · Earnings Updates

Start Here

New to 100 Bagger Hunting?

These are the foundations of how I think and invest.

• How to hunt for 100-baggers, an investment check-list — A full checklist for hunting 100-baggers — the original.

• 100-Baggers, where are they born? — Where 100-baggers actually come from — the common traits.

• How to hunt for 100-baggers — Updated 100-bagger hunt with 50 companies from Joel Greenblatt.

• My 100-bagger investing philosophy — The core philosophy behind every pick on this newsletter.

• The 8-step AtmosInvest Investment process — My 8-step investment process, start to finish.

Portfolio Performance (most recent first)

Annual, half-year, quarterly, and monthly updates on how the portfolio is doing.

• NEW April 2026 portfolio review* — Atomical Capital reveal — and a candid portfolio update.

• Quarterly video Update* — Q4 2025 quarterly video update for subscribers.

• 2025 was crazy — A wild 2025 reviewed, with lessons carried into 2026.

• Up 71%, End of year Portfolio Update (Jan 26)* — +71% for the year — what worked, what didn’t.

• Pro Q3 2025 Update* — Q3 2025 portfolio review for Pro members.

• Monthly Portfolio Update (Dec 25)* — December 2025 — up 18% on the month.

• Monthly Portfolio Update* — Monthly portfolio check-in with position-sizing commentary.

• Quarterly Call Q2 2025* — Q2 2025 quarterly call for Pro members.

• Half Year in: 6 to 25% YTD* — Mid-2025: from +6% to +25% YTD, plus a 20-bagger I missed.

• Q1 2025 Quartely call* — Q1 2025 quarterly call recording.

• 100-bagger hunting - 5 takeaways from 2024 — 5 takeaways from 2024 — mistakes to avoid.

• Half-year review - 11 companies - still investible? — 11 portfolio companies reviewed — still investible?

• 2023 Annual review: Investment mistakes and goals for 2024 — 2023 in review: mistakes made, goals for 2024.

Subscribe now

Stock Deep Dives (most recent first)

Individual company analyses, destination analyses, and new buy pitches.

• Data infrastructure 5x play* — A data infrastructure play derisking into a potential 5x.

• This software company, impervious to AI, has never been cheaper* — A software company AI can’t kill — never been cheaper.

• A double or triple, hidden from the markets* — A base business with a hidden future-growth business layered on top.

• Important info on the 10x opportunity* — Addendum to the 10x opportunity — engineering data matters.

• A data infrastructure 10x opportunity* — The setup for a potential 10x in data infrastructure.

• 172% upside for this hidden serial acquirer* — CareCloud: 172% upside if the turnaround delivers.

• Industry roll-up at a 20% free cash flow yield (and poised for growth) — Industry roll-up at a 20% FCF yield and still growing.

• Weight watchers, value trap or multibagger? — Weight Watchers: value trap or multibagger? My take.

• If Constellation Software and SAP had an Italian child...* — An Italian serial-acquirer that looks like baby Constellation + SAP.

• Update on one of the best risk-reward setups out there* — Update on one of my highest-conviction risk/reward setups.

• Hidden Earnings power about to be unleashed* — Tiny, quality, and an earnings inflection coming.

• Poised for 2,500% returns like Costco? — Long runway, Costco-style compounder potential.

• A data center liquid cooling play — A data center liquid cooling pick with real tailwinds.

• Towards a multibagger with the twin engines* — Toward a multibagger with twin engines — Deep Dive part 1.

• Is Fiserv an easy double? — Fiserv — is it an easy double from here?

• Hidden serial acquirer up 125% in 6 months* — A hidden serial acquirer up 125% in 6 months.

• A small medtech company growing at 50%/year* — A small medtech growing 50%/year with purpose and profit.

• A double in 5 years with the potential for more* — A microcap with unique growth — double in 5 years, maybe more.

• David vs Goliath, David is winning and growing fast* — David vs Goliath: margins expanding, I’m buying.

• From a 200M to a 1 billion dollar Business: A roll-up in security growing at 43% per year* — Security roll-up growing 43% from $200M toward $1B.

• A fast-growing, under-the-radar (30% ROIC) serial acquirer trading at a P/E of 10* — 30% ROIC serial acquirer at a P/E of 10 — too good to be true?

• 55% revenue growth, 158% expected return. I’m buying.* — 55% revenue growth, 158% expected return — I’m buying.

• A small fintech accelerating earnings (+40% yoy)* — A small fintech with +40% YoY earnings — operating leverage kicking in.

• The n°1 stock pitch from Omaha* — The #1 stock pitch from Omaha — not what you’d expect.

• A mini-Palantir growing fast* — A mini-Palantir growing fast, with a smaller crowd watching.

• A microcap monopoly accelerating growth* — A microcap monopoly accelerating into a faster growth phase.

• A secretly growing net-net at an EV/EBIT of 3 — A net-net at EV/EBIT of 3 — secretly growing.

• Rearm Europe: A defense drone producer that goes kaboom (100% growth)* — A European defense drone producer with 100% growth.

• Fast-growing at 25%/year and a forward EV/EBIT of 3.3* — 25%/year growth at a forward EV/EBIT of 3.3. I’m buying.

• Another Pharma becoming a 70-bagger?* — Another pharma setup with 70-bagger potential.

• This microcap won’t stay micro for long…* — A microcap that won’t stay micro much longer.

• Dino Polska is accelerating! — Dino Polska is accelerating — the setup.

• The hidden monopoly is getting stronger* — Update: the hidden monopoly is getting stronger.

• A hidden monopoly promising multi-bagger returns — A hidden monopoly pivoting into something wonderful.

• One 100-bagger for you! — One potential 100-bagger — skill plus a little luck.

• An under-the-radar 20-bagger — An under-the-radar 20-bagger — what’s the next leg?

• A high gross margin software company trading at a PE of 7* — High gross margin software at a P/E of 7.

• Inmode Updated 30-page Deep Dive — InMode — full 30-page deep dive, updated.

• Epsilon Net - A 10-bagger. Can it repeat that feat? — Epsilon Net — a 10-bagger. Can it repeat the feat?

• Retail adventures in Poland: Is Dino still on track + a short position? — Is Dino still on track? Plus a short position.

• Lovesac - The fastest growing company in the furniture industry- A destination analysis — Lovesac — the fastest growing furniture company.

• Inmode Update in 7 steps: Beautiful devices and nasty lawsuits — InMode in 7 steps: beautiful devices, nasty lawsuits.

• Till death do us part: A multi-bagger in the funeral industry? — Fu Shou Yuan — a multibagger in the funeral industry.

• Kaspi – The future Wechat of Central Asia? – A destination analysis — Kaspi — the future WeChat of Central Asia?

• Tinybuild (TBLD) dropped 96% in price since its IPO. Is this the beginning of a turnaround? — Tinybuild down 96% — is this the turnaround?

• Adyen: A destination analysis, where is it headed after Investor Day? — Adyen — destination analysis post investor day.

• Inmode: destroy the long thesis, 7 risks to consider, selling rules, and Q3 update — InMode — 7 risks, selling rules, and a Q3 update.

• Dino Polska: Where is it headed? A destination analysis — Dino Polska — destination analysis.

• Deep dive into Inmode: Multi-bagger or value trap? — InMode — the original deep dive: multibagger or value trap?

• Maison Solutions Inc. - a road to multi-bagger delights? — Maison Solutions — a road to multibagger delights?

• CD Projekt: Taking the company to the next level? — CD Projekt — taking the company to the next level?

AI Tools for Investors

How I use AI (GPT, Claude, Perplexity, Grok) to accelerate stock research. Love AI for investing?

➡️I have a dedicated newsletter on the topic: Invest with AI.

• Better Call Grok: 5 prompts to analyze a stock in less than a minute — 5 Grok prompts to analyze a stock in under a minute.

• 5 steps to accelerate your stock research with Perplexity AI — 5 steps to accelerate stock research with Perplexity AI.

• Perplexity Finance VS GPT-5—here’s what truly wins — Perplexity Finance vs GPT-5 — which wins for research?

• 3 prompts to find hidden winners — 3 prompts to find hidden winners.

• GPT5.2 for stock analysis is here… — GPT-5.2 for stock analysis — plus a descent into cybersecurity.

• GPT5: Will it supercharge your stock analysis?* — GPT-5 applied to The Trade Desk (TTD).

• Measure pricing power in 1 minute with a simple prompt — A one-minute prompt to measure pricing power.

• A 4-step AI enhanced stock analysis process — A 4-step AI-enhanced stock analysis process.

• 📊5 investing tools and the dangers of bad data quality — 5 investing tools — and the dangers of bad data quality.

• Is AI going to kill your favourite SaaS company? — An AI-resilience checklist — is your favorite SaaS safe?

• Will your portfolio companies die by AI? — Will your portfolio companies die by AI?

Investment Ideas

Screens, lists, and roundups — raw idea flow to dig into.

• NEW A death-focused serial acquirer: recession-proof — Death and taxes — a serial acquirer in death care.

• 72 quality companies generated using Buffett’s checklist — 72 quality companies generated with Buffett’s checklist.

• Peter Lynch’s golden rules: 41 Fast Growers on the cheap — 41 fast growers on the cheap — Peter Lynch’s rules applied.

• 50 Wide MOAT Companies, which one reigns supreme? — 50 wide-moat companies — which one reigns supreme?

• Top 4 favorite stocks to own forever — Top 4 favorite stocks to own forever.

• 3 companies inflecting towards high ROIC — 3 companies inflecting toward high ROIC.

• 20 Founder-Led Companies: The one skill you need to look for! — 20 founder-led companies and the one skill that matters.

• 5 watchlist Companies and a Quality Investing Strategy — 5 watchlist companies and a quality investing strategy.

• 1-foot hurdles: 192 companies with an FCF yield > 100% — 192 companies with an FCF yield > 100%.

• TOP 2 “Lynch” fast growers to buy* — The top 2 Lynch-style fast growers to buy now.

• Blooming Ideas: Springtime Investment Collaboration — 7 investors, 7 companies of all sizes — a collaboration.

• 100 Software companies tested for AI resilience — 100 software companies stress-tested for AI resilience.

• 3 Tollway companies - Buffett’s favorite type of business — 3 tollway companies — Buffett’s favorite business model.

• 1 potential multi-bagger and a Zebra — 1 potential multibagger and a turnaround Zebra.

• 🥢Japan, found 3 gems in the rubble? 💎 — Japan — 3 gems found in the rubble.

• 142 serial acquirers to discover* — The database — 142 serial acquirers to dig into.

Frameworks

The reusable checklists, mental models, and tools I use to evaluate companies.

Finding Great Companies

• The alchemy of multibaggers - Which factors can help us find them? — The factors that help find multibaggers — from theory to money.

• Hunting for global stocks that return 1,000% — Hunting globally for stocks that return 1,000%.

• Why unit economics matter when searching for multi-baggers — Why unit economics matter for multibaggers.

• Picking the right wave for better returns — A 7-step checklist for picking the right wave.

• How a Modified Lynch Strategy Could Transform Your Portfolio (While You Do Less Research) — A modified Lynch strategy — more edge, less research.

• 5 steps to pick a software company — 5 steps to pick a software company.

• A 3-step ranking system to beat the market — A 3-step ranking system to beat the market.

• 10 key questions to discover the most profitable yet overlooked companies — 10 questions to uncover overlooked cash-rich companies.

• A ranking system to beat the market — A market-beating ranking system: S for Sales Strength.

• Fantastic MOATS and how to find them — Fantastic MOATs — and how to find them.

Valuation & Multiples

• To boldly go, beyond ROIC: what to do? — Beyond ROIC — what to do when it isn’t enough.

• Stop using the P/E multiple — Stop using the P/E — here’s the one multiple to rule them all.

• Are discounted cash flow models useless? — Are DCF models useless? The competitive advantage period.

• Unlocking 633 Billion USD - The cash conversion cycle and why you should care as an investor — Cash conversion cycle — why it matters, with $633B at stake.

• 6 things I learned about ROIC and how it can help identify a 100-bagger (applied to Inmode Ltd) — 6 things about ROIC — applied to InMode.

• 11 takeaways by applying Musk’s ‘first principles thinking’ to the P/E-ratio — 11 takeaways applying Musk’s first principles to P/E.

• The 6 dangers of Return On Equity (example: Dino Polska) — The 6 dangers of ROE — Dino Polska example.

• The valuation of a fine dining restaurant - Hof Van Cleve — Valuing a fine dining restaurant — Hof Van Cleve case.

Management & Capital Allocation

• A Definitive Spinoff Checklist towards 1500% IRR — A definitive spinoff checklist — Marriott case study.

• Show me the incentives, and I’ll show you the outcome, a checklist to evaluate management — A checklist to evaluate management incentives.

• A 10-step checklist you can use to evaluate management capital allocation to make better investment decisions — A 10-step checklist to evaluate management’s capital allocation.

Buying, Selling & Position Sizing

• Position sizing for massive returns — Position sizing — the most underused lever for returns.

• 5 steps to avoid stupid mistakes before buying in the markets — 5 steps to avoid stupid mistakes before you buy.

• A stock investor’s n°1 secret weapon for excess returns — An investor’s #1 secret weapon for excess returns.

• When to sell a position and boost returns? — When to sell a position — and how it boosts returns.

• A strategy to increase your holding periods and get better returns: The Reactor Portfolio — The Reactor Portfolio — a strategy to increase holding periods.

Mindset & Mistakes

• I missed a 10-bagger in 6 months (IRR = 9900%) — I missed a 10-bagger in 6 months — make sure you don’t.

• 5 easy steps to multibagger returns — 5 easy steps to multibagger returns — wisdom from India.

• The one mistake 99% of investors make — The one mistake 99% of investors make — I made it too.

• 10 lessons from Darwin for 20% yearly returns — 10 lessons from Darwin for 20%/year returns.

• From SpaceX to Einstein and how to gain an edge in investing — From SpaceX to Einstein — how to think for yourself.

• Quality traps and how to avoid them to protect returns — Quality traps — and how to avoid them.

• 7 things to AVOID in the stock market (in 7 minutes) — 7 things to avoid in the stock market — in 7 minutes.

• 1 mistake I made and 7 easy tactics to make better decisions in investing — 1 mistake I made and 7 tactics for better decisions.

• 5 ways to tame our stone age brain — 5 ways to tame our stone-age brain for better returns.

Lessons from Great Investors

Takeaways from the masters: Buffett, Lynch, Greenblatt, Karman, Kantesaria and more.

• A simple strategy that beats the market — A simple strategy that beats the market — Polen Capital.

• 5 takeaways from Lynch’s 29% CAGR to find multibaggers — 5 takeaways from Lynch’s 29% CAGR — life cycle investing.

• How to break the rules to find 100 Baggers — Break the rules to find 100-baggers — from one of the best books ever.

• 6 lessons from a 1000-bagger — 6 lessons from a 1,000-bagger. Yes, a thousand.

• Growth, value, or quality, is all useless. The road to a 29% CAGR — Growth, value or quality — useless. The brilliance of Peter Lynch.

• Stop Index Investing! — Stop index investing — 5 lessons from Seth Karman.

• Gates & Buffett: How to invest, when you have little time — Gates and Buffett — Return on Invested Time.

• 5 lessons on “next level quality investing” towards a 15% CAGR — 5 lessons on next-level quality investing — Dev Kantesaria.

• Divesting, Buffett’s most powerful capital allocation skill — Divesting — Buffett’s most powerful capital allocation skill.

• Finding multi-baggers: 7 lessons from breaking down Paul Andreola’s investment strategy — 7 lessons from Paul Andreola’s multibagger strategy.

• 9 lessons for investors: How farmers yield higher returns — 9 lessons for investors — how farmers yield higher returns.

• 9 takeaways from the deals of Warren Buffett before his retirement at the age of 38. — 9 takeaways from young Buffett’s deals.

Earnings Updates

Quarterly and half-year check-ins on portfolio holdings.

• SAAS Software at a PE of 7 — H1 2025 update — SaaS at a P/E of 7.

• Our second-best bet, up 50% this year* — Our #2 bet — Q2 2025 update, up 50% YTD.

• A market-leading security roll-up aiming for a 5x in revenue* — IMB update — aiming for a 5x in revenue.

• From heaven to hell: $ADYEY and $INMD — Adyen and InMode — from heaven to hell, welcome to purgatory.

• 3Y Sales CAGR of 48%: Code Cracked* — Q2 2025 — the small fast-grower with 48% 3Y sales CAGR.

• Pharmx and sigma sign strategic partnership* — PharmX x Sigma partnership — what it means for PHX holders.

• Adyen H2 earnings ($ADYEN/$ADYEY) — Adyen H2 earnings — expanding margins.

• A banger quarter for CareCloud* — CareCloud Q3 2025 — a banger of a quarter.

• The security roll-up, is it a buy?* — IMB Q1 2026 — the security roll-up, is it a buy?

• Pharmx earnings update - Thesis aligment review* — PharmX H1 FY26 — earnings analysis.

• Adyen update in 3 minutes + Do you rank your stocks? — Adyen update in 3 minutes — buy, hold or sell?

• IMB update - thesis alignement review* — IMB H1 FY26 — thesis alignment review.

• A solid quarter, time to buy more of this high ROIC serial acquirer?* — A high-ROIC serial acquirer — solid quarter, buy more?

• Robot Payment - Addendum to quarterly update* — Robot Payment — addendum with IR feedback.

• A prime example of operational leverage* — Robot Payment Q3 2025 — operational leverage in action.

May the markets be with you, always!

Kevin

I present you:

Atomical Capital

You can check out our site here.

We’ve been hard at work behind the scenes to set this all up. And both 100 Bagger Hunting and Invest with AI will be 2 publications that fall under the Atomical Capital vision.

Once the rebranding is done, I will talk more about what Atomical Capital is.

But regarding why we did it.

Because we had to.

We are driven by always making things better. And that is what we hope to do here.

And as you noticed, I said WE in the introduction, because I’m no longer alone in this endeavor.

• Harvey is our software engineer. He helps build the backbone of Atomical Capital.

• Lars is our analyst. This allows us to look at more companies and find more opportunities.

But we’ll go deeper into this in a later article.

I want to introduce you to something we built in secret (well, kind of) over the past 2 months, and you’ll see the results of that in this post.

Our investment process looks like this:

It’s a closed loop: one agentic tool produces an output for a human, who judges, adapts, and provides input to the next tool.

We want to zoom in on a tool called Tracker.

Tracker is an agentic tool that:

• Reads our deep dive thesis on file

• Gathers ALL data on ALL our companies every week

• Then tells us if the thesis is:

  • INTACT

  • UNDER STRESS

  • BROKEN

It does not judge.

Judging whether we hold or buy more is our job. But the tracker helps us save hours tracking all our companies.

It delivers a dashboard like this:

And it writes a weekly report which highlights:

• If certain companies are under stress or not

• What happened that week with each company

• What we should look out for

We’re still iterating on the design, but we’re glad with how it turned out.

Subscribe now

The thesis the tracker uses must have a specific structure and quality. We spent a lot of time rewriting our originals to make sure the tracker can use them.

You can download the full thesis for Pharmx Technologies on this page.

Microcap Thesis

72.3KB ∙ PDF file

Download

Download

Ok, so enough about Atomical Capital.

Any feedback would be appreciated! Especially if you think Tracker would be useful to you!

Leave a comment

Now let’s get the meat: Our portfolio

Portfolio update

Our portfolios reached an absolute peak at the end of 2025 (remember we trimmed our biggest position)

But since then, the first 4 months of 2026 have not been kind.

Our outperformance over the S&P 500 has decreased from more than 50% to about 9% today.

Are we worried? No.

Because our biggest positions, those companies have only gotten stronger.

But let me show you the actual output from Tracker and what it showed us.

It’s pretty amazing…

Read more

1. A short overview of the markets

2. Rumbu Holdings, a tiny serial acquirer

3. How to measure dilution risk in microcap companies

1. Bubble meter

“We look at some metrics to gauge what the market temperature is.”

Note: Take the skyscraper metric with a grain of salt

After war troubles, the market has jumped once more. Shiller PE has risen, and the Buffett indicator has briefly touched 200%. Margin debt has declined a little bit, and IPO’s are still flowing (96 this year in the US).

You know I follow the macro from afar. If a tariff storm comes in, it’s easy to check your companies. I did do a quick portfolio check with the Hormuz blockage, but nothing to worry about.

2. Stock in the funnel

Peter Lynch always said to go look at boring businesses. Well, I found one, and it sits in our serial acquirer database.

Rumbu Holdings (Ticker: RMB.V) is a microcap Canadian serial acquirer that buys up funeral homes. Here’s my first assessment of the company.

Description: Never heard of it? Neither had I if I hadn’t built the database. Rumbu Holdings buys funeral homes in Western Canada. That’s the business. Death care. This business is recession-proof. People die. They’ll keep dying. (don’t want to sound too morbid, but it’s true).

Type: Early-stage micro-cap roll-up

Why it’s interesting: Canada’s death-care industry generates C$1.5–1.8 billion a year. More than 70% of it is still family-owned. Not chains. Not private equity. Grandpa built it from the ground up. His children grew up watching him work and decided they wanted a different life. Now, grandpa is 72 and has nobody to hand the keys to.

That’s the supply side of this story.

CEO Daryl Lockyer spent 30 years in the funeral industry. He built 17 homes himself before taking Rumbu public via a reverse takeover in late 2023. When he calls a family owner about selling, they know he’s one of them. His pitch is easy: keep your name on the door, keep your staff, take your cheque, and go enjoy retirement.

The business model is simple math. Rumbu buys these homes at 3–6× EBITDA. It trades publicly at 8–10× EBITDA. That difference is where the money is. Every acquisition is theoretically accretive on day one.

And Rumbu has a tailwind. Statistics Canada projects 40% more deaths per year by 2040 as the baby boomers age.

This playbook has worked in Canada before. Park Lawn Corporation ran this exact roll-up and eventually got acquired by a US firm at a meaningful premium ($1.2 billion; it owned 170 funeral homes). Western Canada alone has roughly 300 independent homes still available. Rumbu has six. The runway is long.

Current Status (in the WATCHLIST). The road is clear. But the execution is early.

Run-rate revenue is around C$7M. Operating cash flow margin sits at 12% (well below their 20–26% targets). A full management team, TSXV listing costs, and audit fees are being spread across only six locations, which explains the difference. Those costs don’t scale with funeral volumes. If they reach 12–15 locations, the same overhead becomes a fraction of revenue. The business would start funding its own growth. But right now, every new acquisition still leans on the BMO credit facility (C$9.96M, roughly 40% still unused), and net leverage sits at 4.5× EBITDA, which is elevated for a company this size.

Although they wanted to grow purely on debt, Rumbu announced a C$2M equity raise in early 2026 at C$1.00/share. The reason: They needed to satisfy bank covenants and unlock the next deal. The placement got delayed because the stock had slid below C$1. Nobody wants to subscribe at C$1.00 when the stock trades below it. So the next acquisition is on hold. The flywheel isn’t spinning yet.

The business model is sound: Make sure you have the cash to reach a certain scale, so that margins increase, more operational cash can be used for growth, and debt can be repaid. Grow to 170 funeral homes, and get bought by private equity like Park Lawn in the past.

What I like: Zero analyst coverage. Average daily volume is around 17,000 shares, so pretty illiquid.

So the question becomes: Is this an opportunity or not?

I see two possible outcomes:

1. The raise closes. Rumbu gets back to 2–3 acquisitions per year and reaches 12–15 locations by 2027–2028. Their overhead shrinks as a share of revenue. Annual cash flows cross C$1.5–2M. The business starts compounding on its own. A US consolidator eyeing Canada shows up and pays a premium. Early investors do very well.

2. The raise stays stuck. BMO tightens covenants. Acquisitions stall at six locations. The cash flows cover debt service and not much else. No growth. No re-rating. The thesis doesn’t break, but the stock goes nowhere for years.

I like this company because the setup is easy to understand. But it’s too early for my taste. That’s why I added it to my watchlist. To track it, see how it evolves.

Subscribe now

3. Best article of the week

When you’re diving into small companies, as we saw with Rumbu, one thing to consider is dilution risk.

They can have a great business model and a market tailwind, but if, for some reason, they cannot get the cash to grow, you as an investor might be in trouble.

Ryan Telford introduces the dilution risk scorecard

It’s a way to assess a company’s dilution risk.

And he tested this on the US microcaps universe and got some interesting results.

Check out the full article below ⬇️

https://microcapclub.com/how-to-avoid-dilution-in-microcaps-a-dilution-risk-scorecard/

That’s it for this week.

Portfolio analysis incoming soon!

May the markets be with you, always

Kevin

The setup is clear.

They want to build a powered shell, lease it to a hyperscaler. Microsoft, Google, Meta, or Amazon will bring in the GPU racks and other infrastructure.

Hyperscaler negotiations are ongoing.

And if they pull it off, the company won’t be a microcap for much longer.

But in the meantime, they have a $50M debt to repay. With an ultimate date set on the 30th of June.

That means they need money, and there are only 2 ways to procure it: by equity issuance or debt financing.

And yesterday, they ripped off the band-aid, and I believe they are now in a great position to execute.

Let me explain in detail why I think this play still has a 5x upside from today’s price.

Read more

Some were classified as Fortress-like, meaning it would be difficult for AI to disrupt these companies. Others were classified as exposed, in which AI poses a much greater risk.

One company caught my eye in particular, and with its price down so much over the past 6 months, it was time to dive deeper.

For example, here’s what our AI agent, called “Hunter,” showed us for a company like Adobe:

A final score of 55% and a neutral rating.

Although the financials on the right side of the table look attractive, a neutral rating is tricky. That means margins might erode in the future. And we’re looking for exactly the opposite: a company that could benefit from AI, where margins might increase in the future.

In our article on software companies, we concluded that the impact is more nuanced than at first glance and that the playing field has fundamentally changed. Some companies' multiples have been corrected, rightfully so.

Others who benefit from this transition have seen the same contraction without it affecting their business. And this is the case with the following company. Misunderstood and wrongfully sold off by more than 40% from the ATH, even though:

• Their pricing is tied to data volume and capacity, so as AI models generate massive amounts of new data, the taxable surface area grows.

• They operate in a space where systems have to be deterministic and are not allowed to hallucinate

• They launched an Agent Cloud, making them the coordination hub for AI safety and data integrity

On the financial side, the company has:

• An average subscription dollar-based net revenue retention of > 120%.

• Yoy revenue growth of >40% at over a billion $ scale

• Net promoter score of >80

• Gross Margins of 80%, thanks to its transition to a software pure play

at an EV/Sales below 8!

The company we’re diving into today received a Fortress rating. Only 8 out of 100 companies received such a high score.

After the downturn over the last 9 months, this company is now up 10% over the past week.

Is the company finally inflecting?

Is it time to hop on the train?

Let’s dive in.

Read more

It’s a tool we call Hunter. You give it a framework and a list of tickers, and it will use financial data and annual reports to assess each company.

We gave Hunter a SAAS scoring framework and 100 software companies that had significant drawdowns over the past 12 months to evaluate.

This took us many days, trial and error, and iteration. You can download the full Excel at the end of this article.

Here are the steps we used:

1. Refining the framework

2. Choosing the right model

3. Choosing financial metrics

Let’s go.

Refining the framework

Our initial framework looked like this:

Tier 1: Survival

• Does the company charge for an outcome or per seat? If per seat, the company might have a problem.

• How accurate does the software need to be? If an error does not lead to a big problem, AI might displace it.

Survival is the most important aspect of the 4.

Tier 2: Defence

• How many system integrations does the software control? The more integrations, the bigger its defensive position.

• How big is the switching cost? How much accumulated value would a customer lose by switching?

• Who are the clients? Enterprise is better than SMB, which is better than individuals

• Are there network effects? If the product becomes better with more users, AI might not be a problem

Tier 3: Offence

• How effectively is the company integrating AI into its own systems to strengthen its product?

Tier 4: Secondary moats that compound over time

• How much proprietary data is involved? The more the better

• Do legal or compliance barriers protect the business?

This is where we stopped.

We don’t want a framework that is simple but not too simple.

And we applied this to a couple of companies, like Rubrik, which gave us the following table:

And a final score of 10.5 out of a max 13.5, which is very high.

But then one of you asked about agentic interaction with this software. If the future is agentic, will some software companies actually benefit from the AI revolution?

So we changed the framework and added this to Tier 3:

• Is this company’s product likely to be called by autonomous AI agents at scale?

In other words, we try to look into the future. The strongest companies will provide infrastructure for the agentic era.

Choosing the right model

So we have a framework and a scoring system.

We need an LLM to evaluate the information we provide about the company (from the annual and quarterly reports) and return a score.

So we used Claude Sonnet 4.6 and Opus 4.6, with and without thinking mode, to see which would provide the best result.

For ServiceNow (NOW), we got the following results:

So first off, ServiceNow scores high. But the scoring is off if you don’t apply the reasoning mode to the model, which makes sense.

In our analysis, although Opus 4.6 can capture much more nuance, it made the scoring very bullish (with a 100% perfect score).

After evaluating about 10 companies, we decided that Sonnet 4.6 with the thinking mode would be the best model from both performance and cost perspectives. (Sonnet is 5x cheaper but seems to be more grounded in facts)

Subscribe now

The scoring system

Just a quick recall of the scoring system used.

For each statement, it will say:

• Anti-fragile = +1

• Neutral = 0

• Exposed = -1

And we’ve assigned weights to the different parts of the framework:

So the perfect company would score 13.5 points or 100%.

Based on this model, here’s what the different ranges will be for the companies:

So we can aim for the resilient, fortress companies in the bunch.

We can debate this framework. What matters is the assessment and reasoning Hunter will deliver. That will allow us to draw our own conclusions.

Financial metrics

But a resilient score on itself means nothing.

If a company is classified as Fortress, but still commands a high price in the market, it might still be a bad investment.

So we add specific SaaS metrics alongside each company’s overall AI resilience score. This would give us a better picture.

We added different metrics, but some of them in particular:

• Gross margins

• Rule of 40

• Rule of 40 (Stock-Based Compensation (SBC) corrected)

• Price / Free Cash Flow

• Price / (FCF - SBC)

• Revenue growth

Maybe some metrics are missing. We tried not to put too much in. Which do you think we should add?

Leave a comment

The challenger

You might have heard of the council model by Andrej Karpathy. You let different models do an analysis and then compare them, or let them debate their own results.

The final result is higher quality output and fewer hallucinations. We didn’t build a full council model.

We added an additional challenge layer to Hunter. The challenger has only one goal: to poke holes in Hunter’s assessment of a company. And a brief text analysis will be included in the final output.

You could see it as a form of reliability of the final output. Now, in the end result, the challenger will give the following output:

• Agree: Only minor challenges with the assessment

• Challenged: Some challenges

• Very Challenged: Significant challenges

You can see it as a form of reliability.

But no company has scored “Very Challenged” in the final output. The reason is that the Sonnet in thinking mode is already very conservative in its assessments.

The result

Now let me show you the detailed output on a company like Adobe.

Adobe trades at around 240 USD today, down from its previous 634 USD peak almost 2 years ago.

Here’s what Hunter gives us:

The financials look great. But it also shows that Adobe is not in a bad spot, but not in a great spot either. He ranks it as neutral.

Here’s how it did the analysis. I highlighted the most important parts in bold.

Adobe is a $23.8 billion revenue software incumbent occupying a structurally privileged position across the content lifecycle — from creative production through marketing orchestration — with 96% subscription revenue, $25.2 billion in ARR, and 40% FCF margins that reflect genuine pricing power and low marginal delivery costs. Its core strength against AI disruption lies in entrenchment: the Creative Cloud and Adobe Experience Platform function as workflow backbones across enterprise content operations, with years of accumulated assets, configurations, partner ecosystems, and multi-year contracts creating migration friction that is costly and operationally risky to overcome. Adobe’s second genuine strength is its early positioning as agentic infrastructure — the Adobe Experience Platform Agent Orchestrator, Firefly Services APIs, and Brand Concierge represent deliberate architecture for machine-to-machine consumption, not retrofitted marketing. Every autonomous marketing agent that an enterprise deploys is a plausible Firefly Services or AEP API consumer. However, Adobe faces two significant vulnerabilities. First, its pricing model remains primarily seat-based, and while the shift to consumption-based Firefly credits and API services is underway, the transition is incomplete and not yet quantified in revenue terms — the company discloses no AI-specific ARR, making the monetisation thesis harder to verify. Second, the competitive threat from AI-native entrants is not theoretical: Adobe’s own filing explicitly names AI-native companies attacking creative tools, document productivity, and marketing automation, and acknowledges ‘limited barriers to entry’ in these markets. The commercially safe Firefly training corpus and IP indemnification are genuine defensive assets, but they are not impenetrable.

Overall verdict: Adobe is a well-positioned incumbent with real structural advantages in entrenchment and agentic infrastructure, but one that faces credible AI-native competition and has not yet demonstrated the quantified AI revenue impact that would make its offense fully antifragile. It is defensible, not dominant.

What do you think of this analysis?

Leave a comment

I personally think it’s a good assessment of Adobe’s current situation.

But the power of Hunter is not to run it on a single company.

The power is to run it on hundreds of companies, and then rank them.

So we looked at 95 software companies that had significant drawdowns over the past year and ran our Hunter through them.

Comparing the companies

That means you now have a list you can use to compare every company. If we select 1 company from each category, it looks like this:

Not all financial metrics are shown here, but you can see, for each company, the rule of 40, and even that rule is corrected for Stock-Based Compensation.

Rubrik, for example, is a fast-growing company classified as Fortress-level, but its SBC weighs on its market pricing.

Duolingo, on the other hand, is at the bottom of the list of the 95 companies. It is considered completely exposed.

And I think the model is right:

Does that mean Duolingo is a bad investment?

Not necessarily.

How to use this data?

Our goal was to build a tool that leverages an LLM to analyze companies’ financial and textual data and score software companies.

And as you’ve seen in our framework, a SaaS enterprise company will always score higher than a B2C company like Duolingo.

But the model does not account for Daily Active Users growth, ARR, or other typical SaaS metrics.

Hunter provides an additional starting point for a company analysis. It provides a more informed list than your typical stock screener.

It is not the end result.

In future articles, we’ll go deeper into some of these companies.

The sweet spot is to look for resilient companies that have experienced significant drawdowns and trade at lower multiples, yet still offer ample growth prospects.

Those are the companies that the market might have thrown out and shouldn’t have.

We’ll write about some of them in the future.

Oh, and before I forget, you can download the Excel file:

Hunter Ai Resilience 2026 03 27 Sonnet Thinking

469KB ∙ XLSX file

Download

Download

Next steps

Hunter is only the first iteration of what we’re building. You can imagine using different industry frameworks and running the same analysis on entire industries. Our final vision is that we select an industry, and Hunter will run that industry's framework across all companies in that industry and score them. (we do hope token cost will go down though 😉)

Let me know what you think.

May the markets be with you, always!

Kevin

In just a few months, the narrative that AI is making software companies even stronger has been called into question by the release of Claude’s latest capabilities, which have reshaped expectations across the Software landscape.

We’re seeing market disruption happening in real time.

And I’ve been using Claude Code for a couple of months now, also for stock analysis, and I do believe the disruption is real.

It’s no longer: “Can I build it?” It’s should I build it?

But the breakdown of software companies won’t come overnight. Multiples will have a hard time rerating. Most of them will die slowly.

While this industry drawdown is playing out in real time, every investor should be asking the same two questions:

1. How do I know if a company is likely to be disrupted by AI?

2. Which names are being wrongfully sold off in this industry drawdown?

To answer both questions, we developed the following checklist.

Checklist

The Checklist consists of nine lenses across four tiers, with three exposure types (Exposed, Neutral, Antifragile).

We assigned a weight to each of them

Tier weights:

The scoring system is as follows:

• Exposed = -1 point

• Neutral = 0 points

• Antifragile = +1 point

Score range: -13.5 to +13.5 (weighted)

Let’s dive into the framework…

Subscribe now

Tier 1: Survival - Do you still have a business in 5 years?

1. Pricing Model (The “Cash Flow” Pillar)

Question: Does the company charge for an outcome or per user seat?

This is the most immediate threat to cash flow. If you charge per human seat and AI makes humans 10x more productive, your revenue will mathematically collapse by 90%. The shift from paying for work to paying for outcomes is the difference between a thriving business and a dying one.

• Exposed: The company charges its customers solely per seat. Revenue is directly tied to headcount, which AI will compress.

• Neutral: The company blends seat-based and usage-based charges. Some revenue still scales with headcount, but AI-driven efficiency can be offset by fees tied to data volume, API calls, or feature usage.

• Antifragile: The company charges based on consumption, outcomes, or value delivered (compute, transactions, storage, data volume). AI adoption amplifies revenue as usage grows.

2. Failure Sensitivity (The “Liability” Pillar)

Question: How accurate does AI need to be to replace this software?

This defines the “Trust Gap”. AI is probabilistic; if a guess results in a typo in a blog post, it’s fine. If it results in a plane crash or a cybersecurity breach, it’s a disaster.

• Exposed: 90% accuracy is good enough for the company’s core use case (e.g., marketing copy, basic graphic design). The software is likely to be replaced by AI.

• Neutral: AI needs to be consistently reliable (97-99% accurate), but occasional errors are tolerable with human review. The software is partially defensible.

• Antifragile: Failure or hallucination is catastrophic (e.g., medical diagnostics, cybersecurity, power grid management, financial compliance). The software is harder to replace with a probabilistic AI model.

Tier 2: Defence - Preventing competitors from eating your lunch

3. Workflow Integration (The “Stickiness” Pillar)

Question: How many system intersections does it control?

Even if an AI tool is better, a company won’t switch if that software is the central nervous system connected to 50 other apps. Replacing a core workflow is like performing heart surgery; most companies would rather just upgrade their current heart with AI.

• Exposed: The software operates as a standalone tool with minimal integrations, few dependencies, and low workflow entanglement. It is easier to replace.

• Neutral: The software connects to multiple systems but does not serve as the central coordination layer. Replacing it creates friction, but not systemic breakdown.

• Antifragile: The software sits at the intersection of many systems, acting as a coordination hub or workflow backbone. Replacement becomes structurally difficult.

4. Switching Cost (The “Friction” Pillar)

Question: How much accumulated value would a customer lose by leaving?

This is distinct from workflow integration. A product can have few integrations but enormous switching friction due to years of accumulated data, configurations, trained employees, and organisational muscle memory. The question isn’t how connected it is? It’s how painful it is to walk away.

• Exposed: Customer data is easily exportable, minimal customisation is built on top, and retraining cost is low. A competitor can replicate the experience with minimal migration effort.

• Neutral: Customers have a moderate level of accumulated investment, such as historical data, custom configurations, or trained workflows, but migration is feasible within a quarter with dedicated effort.

• Antifragile: Customers have years of accumulated data, configurations, custom automations, compliance audit trails, and deeply trained teams. Migration would require months of effort, retraining thousands of employees, and risking operational disruption. The cost of switching far exceeds the cost of staying.

5. Customer Base (The “Inertia” Pillar)

Question: Who does it primarily serve?

Enterprise clients aren’t a moat on their own. However, the cost of software is often negligible compared to retraining 10,000 employees and reconfiguring 100 internal processes. There is a reason why so many people still use Microsoft solutions for use cases where better alternatives exist.

• Exposed: The software mainly serves individuals or small businesses. Switching costs are lower, and experimentation is higher.

• Neutral: The company serves both SMB and enterprise clients. Defensibility depends on contract structure, integration depth, and switching friction.

• Antifragile: The software primarily serves large enterprises. Displacement is significantly harder. Enterprise environments involve procurement cycles, compliance requirements, multi-year contracts, security audits, and cross-department integration.

6. Network Effects (The “Ecosystem” Pillar)

Question: Does the product become more valuable as more participants join the ecosystem?

A product with true network effects creates a self-reinforcing flywheel that is difficult to replicate. It’s not enough to have integrations (that’s Workflow Integration) or switching costs (that’s Friction). This lens measures whether the ecosystem itself compounds in value as scale increases. A marketplace with 10,000 apps is not just 10x better than one with 1,000. There’s a quality difference.

Exposed: The product is a standalone tool with no meaningful ecosystem dynamics. Value is linear; each customer gets the same experience regardless of how many others use the platform. No integration marketplace, no third-party developer community, no cross-customer data benefits.

Neutral: The product has some ecosystem elements, a modest integration marketplace, partner channel, or community, but these are additive rather than compounding. Competitors could replicate the ecosystem with sufficient investment. The product would still function without the ecosystem.

Antifragile: The product sits at the centre of a self-reinforcing ecosystem where value compounds with scale.

Tier 3: Offence - Can AI make this company stronger?

7. AI Adoption Capability (The “Offence” Pillar)

Question: How effectively is the company integrating AI to strengthen its product and expand its market?

Defence alone isn’t enough. The companies that will win the AI era are the ones that use AI to widen their moat by embedding it into their product to deliver more value, expanding into adjacent markets, or creating entirely new revenue streams. A company that is antifragile on defence but slow on AI adoption can still lose to an incumbent that moves faster.

• Exposed: The company has no meaningful AI strategy, no AI features shipped, and no public roadmap for AI integration. It is passively waiting for disruption rather than responding to it.

• Neutral: The company has announced or shipped AI features (e.g., copilots, assistants, AI-generated suggestions, AI-powered automation), but these are incremental enhancements to existing products. AI is a marketing message and a feature, not a proven revenue driver. Most software companies in 2025-2026 fall into this category.

• Antifragile: The company has demonstrated measurable AI impact on its business by not just shipping features, but also proving results. Only a few companies to this day have this.

Tier 4: Alpha - Secondary moats that compound over time

8. Data Ownership (The “Intelligence” Pillar)

Question: How much proprietary data is involved?

Data is only useful if it’s proprietary and clean. If your data is just a mirror of what’s on the public web, it’s not a moat. It only matters if it’s “under-the-hood” data that public models can’t see.

• Exposed: The software relies primarily on public, user-exportable, or easily replicable data.

• Neutral: The company holds some proprietary data, but it is customer-specific and not meaningfully aggregated across the network.

• Antifragile: The company controls proprietary, network-aggregated data that improves as more users join. It benefits from strong data gravity. The system becomes smarter at scale in ways public AI models cannot replicate.

9. Regulatory Exposure (The “Red Tape” Pillar)

Question: Do legal or compliance barriers protect the business?

Regulation is a speed bump moat. It keeps small startups out because they can’t afford the legal fees, but it doesn’t stop a determined incumbent. It’s a great secondary defence, but a weak primary one.

• Exposed: The software operates in a lightly regulated environment with minimal legal constraints.

• Neutral: The company operates in a regulated environment, but compliance is standardised and replicable. Regulation slows entrants but does not prevent disruption.

• Antifragile: The software operates in a highly regulated domain where legal approval, certifications, audits, and liability structures are complex and time-consuming. Regulation becomes a structural moat.

Stress testing HubSpot, SYS-DAT, and Rubrik

Now that we have established the framework, we are going to test it on three companies:

• HubSpot is a CRM platform that helps small to medium-sized businesses attract customers, automate marketing and sales, and manage relationships all in one integrated system.

• SYS-DAT is an Italian IT services company that supports manufacturing businesses in Italy through ERP solutions, cybersecurity, cloud infrastructure, and data management.

• Rubrik is a cybersecurity company that protects and manages enterprise data by providing backup, recovery, and ransomware resilience across cloud and on-premise environments.

Walk through the checklist

HubSpot is down 35% since the beginning of the year. Let’s see how well their moat holds up against AI:

1. Pricing Model

This is the Achilles’ Heel. HubSpot’s revenue is historically tied to Core Seats. If AI allows a small team to do the work of a department, the customer will naturally reduce their seat count, directly eroding HubSpot’s top-line revenue. They are attempting to shift toward platform-wide fees, but the seat-based legacy model remains high risk.

Verdict: Exposed

2. Failure Sensitivity

HubSpot’s primary use cases, like marketing emails, blog posts, and sales outreach, are probabilistic tasks. If an AI generates a draft that is 90% correct or the customer information has minor errors, it’s not the end of the world. However, there will still be human interaction needed, since customer information and blog posts vary and are customer-facing.

Verdict: Neutral

3. Workflow Integration

This is HubSpot’s greatest strength. It acts as the System of Record for the mid-market, connecting marketing, sales, and service. With thousands of App Marketplace integrations, pulling HubSpot out of a business often causes a systemic breakdown in operations.

Verdict: Antifragile

4. Switching Cost

HubSpot primarily serves the 20–500-employee segment, where years of CRM data, sequences, and workflows create meaningful switching friction. Switching is painful but not prohibitive, especially for the smaller end of HubSpot’s base, where customisation depth is lower and new tools could accelerate the process.

Verdict: Neutral

5. Customer Base

HubSpot is the king of the mid-market. While these businesses are stickier than individuals, they are also the first to experiment with lean, AI-only marketing teams to save on overhead during industry shifts. HubSpot primarily utilises annual contracts for its professional and enterprise tiers, while offering more flexible monthly options for starter plans.

Verdict: Neutral

6. Network Effects (The “Ecosystem” Pillar)

HubSpot’s 1,900+ connector marketplace and growing Breeze agent ecosystem are real, but they are additive rather than compounding. The integrations are overwhelmingly connectors to third-party tools like Slack, Google Workspace, Salesforce, and Zapier that push and pull data in and out of HubSpot, rather than ISVs building differentiated, standalone products on top of HubSpot that customers depend on in their own right.

Verdict: Neutral

7. AI Adoption Capability

Breeze Customer Agent resolves over 50% of support tickets during early HubSpot deployments. However, no AI-specific revenue has been disclosed, and the features (copilots, assistants, summarisation) are broadly similar to those offered by every major CRM competitor. Strong execution, but not yet a proven revenue driver or new-market unlock.

Verdict: Neutral

8. Data Ownership

HubSpot holds a massive amount of customer-specific data, but it is siloed. While they have launched “Breeze” to help aggregate insights, the data is ultimately user-owned and exportable. They lack a “network-aggregated” moat where one user’s data makes the tool significantly smarter for everyone else (unlike a cybersecurity firm).

Verdict: Neutral

9. Regulatory Exposure

HubSpot primarily operates in marketing, sales, and customer service - lightly regulated activities with no meaningful certification barriers. Any well-capitalised AI-native startup can compete in this space.

Verdict: Exposed

Weighted Score = (Tier 1 sum x 2) + (Tier 2 sum x 1.5) + (Tier 3 sum x 1.5) + (Tier 4 sum x 1)

HubSpot gets a final score of -1,5 (Score range: -13.5 to +13.5)

Now let’s walk through the checklist with SYS-DAT, which has been sold off by 20% ytd. You can read our deep dive here.

1. Pricing Model

Much of their revenue comes from long-term maintenance contracts, project-based implementations, and modular licenses. These are often calculated as a fixed percentage of the initial software value, regardless of how many people log in. Some revenue still scales with headcount, however.

Verdict: Neutral

2. Failure Sensitivity

SYS-Dat acts as the single source of truth for a business: warehouse management, finance, and supply chain. In the Italian regulatory environment, an ERP error in a VAT return or on a manufacturing line can lead to heavy fines or production shutdowns. You cannot hallucinate an inventory count.

Verdict: Antifragile

3. Workflow Integration

System integration is in their DNA. Their software acts as the central hub for manufacturing execution systems (MES), CRM, and Retail. It is the software glue that holds an Italian SME together.

Verdict: Antifragile

4. Switching Cost

Multi-year implementation projects, custom integrations, and trained internal teams create real friction during migration for SYS-DAT clients. But because SYS-DAT largely relies on third-party platforms, a client could theoretically switch to SYS-DAT as the integrator while retaining the underlying software.

Verdict: Neutral

5. Customer Base

SYS-DAT serves businesses ranging from micro-enterprises to large multinationals across the Italian manufacturing industry. The mix is broad rather than enterprise-focused, which provides revenue diversification but limits the structural inertia that comes from predominantly serving large enterprise clients with complex procurement and compliance requirements.

Verdict: Neutral

6. Network effects

SYS-DAT has no app marketplace, no developer ecosystem, and no cross-customer data flywheel. Value delivered to one client is independent of the number of other clients on the platform. It is a professional services and software reseller business with linear, not compounding, scale dynamics.

Verdict: Exposed

7. AI adoption

SYS-DAT references AI across its offering and has acquired cybersecurity and digital commerce capabilities. However, there is no evidence of disclosed AI-specific revenue, a proprietary AI model, or AI enabling entry into genuinely new markets. The strategy appears to be keeping pace with vendor roadmaps (SAP, Cegid) rather than driving a proprietary AI wedge.

Verdict: Neutral

8. Data Ownership

They own deep, industry-specific data silos (e.g., textile production cycles in Tuscany). While this data is clean and valuable, much of it is specific to individual clients rather than to a network-aggregated “mega-brain” that gets smarter across all users.

Verdict: Neutral

9. Regulatory Exposure

Italy has some of the most complex tax and labour laws in Europe. SYS-Dat builds these local legal nuances directly into its code. A Silicon Valley AI agent doesn’t know the specificities of Italian digital product passports or local labour audits. SYS-Dat is building its AI suite (ArIA) to be compliant by design within this specific Italian framework.

Verdict: Antifragile

SYS-DAT gets a final score of 3 (Score range: -13.5 to +13.5). Looking a little bit better than HubSpot, but a bit stuck in the middle.

Subscribe now

Last but not least, let’s look at Rubrik, which is down 30% ytd:

1. Pricing Model

Rubrik charges based on Subscription ARR (linked to data volume and capacity). As AI models generate massive amounts of new data and “Shadow Data” stores, Rubrik’s taxable surface area grows. They don’t care if a company has 10 employees or 10,000; they care about the Petabytes of data being protected.

Verdict: Antifragile

2. Failure Sensitivity

In the 2026 threat landscape, mostly recovered isn’t good enough. If Rubrik fails to restore an immutable backup after a ransomware attack, the client faces total operational collapse or even loss of life (in healthcare). AI hallucinations in recovery are unacceptable; the system must be deterministic.

Verdict: Antifragile

3. Workflow Integration

Rubrik is the central nervous system of a company’s Disaster Recovery (DR) plan. With the launch of Rubrik Agent Cloud in late 2025, they have moved to monitor and audit the actions of AI agents themselves, making them the coordination hub for AI safety and data integrity.

Verdict: Antifragile

4. Switching Cost

Rubrik customers accumulate years' worth of immutable backup snapshots, SLA policy configurations, compliance audit trails, and recovery runbooks within the platform. The switching cost is not primarily financial. It is the risk of losing validated recovery capabilities and audit continuity during any migration window. In a domain where downtime is measured in millions per hour, that risk is prohibitive.

Verdict: Antifragile

5. Customer Base

As of early 2026, 85% of their revenue comes from customers spending over $100k annually. These are Fortune 500 giants with zero trust mandates. The psychological and technical friction of moving an entire enterprise backup and security architecture to a new vendor is monumental.

Verdict: Antifragile

6. Network effects

Rubrik does not operate a marketplace that others build upon, and its value does not compound significantly across customers the way a platform with third-party developers would. There is a partial cross-customer benefit through shared threat intelligence (the Data Threat Engine learns from signals across its installed base), but this is not yet a primary, publicly quantified moat.

Verdict: Neutral

7. AI adoption

Rubrik’s Ruby AI companion handles backup failure diagnosis, root cause correlation, and guided cyber recovery workflows, and machine learning models in the Data Threat Engine continuously learn to reduce false positives. This is meaningfully integrated AI, not just a copilot wrapper, but Rubrik has not yet disclosed AI-specific revenue contribution or demonstrated that AI has opened genuinely new market segments. Strong execution, not yet proven commercial uplift.

Verdict: Neutral

8. Data Ownership

Rubrik sits on the most valuable data a company has: its backups. With Rubrik Annapurna, enterprises can now create secure embeddings from this data to feed into internal RAG (Retrieval-Augmented Generation) models. They are the gatekeepers of the company's proprietary brain.

Verdict: Antifragile

9. Regulatory Exposure

Rubrik is a certification fortress. With FedRAMP High, GovRAMP, and their new Security Cloud Sovereign (Jan 2026), they meet national security standards that generic software startups cannot touch. In 2026, data sovereignty is a legal requirement that Rubrik has productized into a moat.

Verdict: Antifragile

Rubrik gets a final score of 10,5 (Score range: -13.5 to +13.5).

Conclusion

HubSpot (-1.5) is vulnerable.

HubSpot falls into the Vulnerable classification with meaningful AI risk, and a valuation that reflects this discount. I expected it to be in this category; however, I am still not certain if the harshness of the 70% sellof is justified for a company that could still grow modestly with smaller margins at 4x sales.

SYS-DAT (3) is neutral.

SYS-DAT is in the neutral classification, where only some parts of the business are affected, and only some parts benefit from AI. If they successfully integrate AI into their processes, they might actually benefit from the AI-eats-software narrative because of depressed valuations of future M&A targets.

Rubrik (10,5) is a fortress.

Rubrik is part of the fortress classification thanks to its cloud-native architecture and the implementation of AI in its core products. Those are the companies that will not only survive but also flourish, which is why Investors should pay extra attention to this category.

We’ll use this framework to analyze about 150 software companies, like Adobe, Duolingo, and many others.

It will give you an AI-resilience quality score that you can use to include in your overall analysis and valuation of the company.

Big Excel incoming…

May the markets be with you, always!

Kevin

Sometimes, there are weeks without finding any opportunities in the markets, and then suddenly they all come at once.

The window to get in at a good price is closing, though. I expect the next quarter to have good news.

This company is a small company that has suddenly been backed by a leader in their industry.

A microcap that gets access to the revenue flow of a billion-dollar company.

You can imagine that once that revenue starts flowing, it creates leverage for this company.

For certain companies, I need days or weeks of due diligence before acting. But after a couple of hours on this one, I was sold.

That doesn’t happen often.

My bet is on a double or triple in the coming years from the current price.

Let’s dive in.

Read more

Use the link below to watch and review our update.

Use this passcode: 9=55q+UF

WATCH VIDEO

Questions or comments, send me a message:

Have a great weekend!

Kevin

Read more

As I mentioned in our previous article.

Time is of the essence.

And now I have obtained a recording with the CEO and COO being interviewed by an institutional player.

In the call, they ask individual investors to disconnect from the call…

So the general public was not supposed to have this information?

Well, now we do, and it makes the case all the more compelling.

The thesis is getting more real, and the possibility for outsized returns, too.

Here’s the analysis ⬇️

Read more