Convergence Effects: Five Patterns That Mint New Giants

Remember when you’d buy a GPS device for your car? A little Garmin suction-cupped to your windshield. Standalone product. Whole industry. Billions in sales.

Then the smartphone showed up and said:

“Hey, that entire industry? That’s a button on my home screen now.”

Poof. Same thing happened to:

• Point-and-shoot cameras → a button
• MP3 players → a button
• Flashlights → a button
• Calculators, alarm clocks, voice recorders, pocket dictionaries → button, button, button, button

This is Absorption. A whole industry gets eaten and turned into a feature of a bigger platform.

The tricky part is that Absorption doesn’t look scary from the inside. When the iPhone launched, Garmin’s sales didn’t drop the next day. The GPS was still better than the phone’s map. For a while! But “for a while” is exactly the trap.

Here’s an example that’s happening right now:

Think about all the software you (or your company) pays for that does ONE thing. Scheduling. Note-taking. Transcription. Grammar checking. Translation. Each of these was a real company with a real stock price.

Now watch what AI models are doing to them. Grammar checking? A prompt. Translation? A prompt. Meeting notes? A feature of the video call. The AI platform isn’t competing with these companies. It’s absorbing them, the way the phone absorbed the flashlight.

Digital camera shipments, from the ≈121M peak (2010) to ≈8M (2023). Note the peak came three years after the iPhone launched—absorption looks survivable right up until it isn’t.

Ask one question about any company you own: “Could this entire product be a menu item inside something bigger?”

If the answer is yes, you have two choices. Either you’re betting the company gets acquired (sometimes the absorber just buys you, which is a fine outcome), or you’re holding the Garmin of 2008.

And on the flip side: the absorber gets to charge for the platform while getting the features for free. That’s where value flows.

This one is great because it’s so weird when you actually look at it.

Nvidia spent decades building chips that make video game explosions look pretty. That’s it. That was the business. Make the dragon look shinier.

Then some researchers noticed: “Huh. The math for drawing a shiny dragon is the same math for training a neural network.”

Nvidia didn’t invent AI. But they had already solved AI’s hardest problem, for a completely different reason. Twenty years of gaming R&D turned into the foundation of the most valuable company on earth.

That’s Cross-Pollination. A problem that’s been fully solved in Industry A turns out to be the bottleneck in Industry B.

More examples:

• Lithium-ion batteries were perfected for laptops and camcorders. Then the car industry realized: “wait, if we stack 7,000 of these...” → Tesla.
• CRISPR started as an obscure discovery about how bacteria fight off viruses in yogurt cultures. Now it’s rewriting human DNA in clinical trials.
• The transformer (the “T” in ChatGPT) was built for translating languages. Someone pointed it at protein structures instead of sentences, and it cracked a 50-year-old biology problem.
• Consumer electronics supply chains (cheap sensors, cheap cameras, cheap radios) made it possible to build satellites and rockets for a fraction of the old price. SpaceX and the whole “new space” industry ride on parts that were originally made for your phone.

Notice the direction of value here. In Cross-Pollination, the company that already solved the problem gets a second business handed to them for free. Nvidia didn’t need to rebuild anything. They needed to write some software and change their sales pitch.

Reduction in drilling time Fervo reported between its first enhanced-geothermal wells (2022) and its Cape Station wells two years later. Shale’s learning curve, transplanted whole.

Look at what the frontier industry is struggling with. Then ask: “Who already solved this somewhere boring?”

AI is struggling with power delivery, cooling, and memory bandwidth. Robotics is struggling with cheap, precise actuators. Biotech is struggling with data. Every one of those problems has been solved, or nearly solved, in some unglamorous corner of the market.

The companies sitting on those solutions often trade like boring old industrials. Until someone notices.

Sometimes industries don’t eat each other or borrow from each other. They merge and produce something that didn’t exist before.

• A GLP-1 telehealth company isn’t healthcare and it isn’t tech. It’s a doctor’s office that’s also a pharmacy that’s also a subscription app.
• Starlink isn’t a rocket company or a phone company. It’s a phone company whose cell towers happen to be in orbit.
• A humanoid robot company isn’t a machine company or a software company. It’s a labor company.
• Waymo isn’t Uber and it isn’t GM. It’s a fleet of robots that sells rides. Nobody has a clean valuation model for that yet.

These are New Species. The offspring of two industries, with a business model neither parent could run.

Here’s why New Species are mispriced in the beginning:

They have no comps.

When a company is truly new, nobody knows how to value it. Is Tesla worth 10x earnings like Ford, or 40x like a software company? Analysts argued about this for a decade. When nobody knows what something is worth, the price swings wildly, and the people who understand what the thing actually is have an edge.

Some New Species being born right now:

When you see a New Species, ignore the industry labels on the stock screener. Ask instead: “What does this thing actually sell, and who does it actually compete with?”

A humanoid robot doesn’t compete with other machines. It competes with a $40,000/year warehouse salary. Depending on how you frame it, the math changes completely.

The risk, of course, is that New Species have a high infant mortality rate. Most hybrids don’t survive. But the ones that do become the giants everyone writes books about later.

Here’s how Wall Street works:

Big investment banks have research departments. Each analyst covers a sector. There’s an autos analyst. A semiconductors analyst. A media analyst. A retail analyst. Each one has 15–25 companies they follow, and they write reports that thousands of fund managers read.

Now think about what happens when a company doesn’t fit cleanly into one sector.

• Tesla got covered by auto analysts for years. Auto analysts think in terms of units sold, factory capacity, and profit per car. They compared Tesla to Ford. Meanwhile, the software and energy parts of the business were overlooked by the people writing the reports.
• Amazon was covered by retail analysts. Retail analysts think about same-store sales and margins on shipped goods. AWS, which turned out to be the most profitable part of the company, grew inside a retail company for years while getting valued like a warehouse.
• Nvidia was a “gaming chip company” in the analyst notes long after the AI business was obviously the real story.

This is the Blind Spot. When a company sits between two sectors, it gets covered by the wrong analyst, or by nobody.

Analysts are paid to be right within their sector. The autos analyst gets no bonus for correctly predicting that a car company’s software business will be worth $200 billion. That’s not his job. His job is Ford versus GM versus Toyota.

Plus, when analysts don’t have a model, the big funds have a harder time buying. Many institutional investors need a research report to justify a position. No report → no position → the stock trades on retail and a handful of specialist funds → volatile, mispriced, ignored.

Revenue AWS booked in 2015—the year Amazon first broke it out as a line item. A cloud giant had been hiding inside “a retailer” for nearly a decade, and the stock repriced the week the disclosure landed.

When you find a company that’s a Cross-Pollination story or a New Species (patterns #2 and #3 above), check who’s covering it.

Look at the analyst list. What sector are they from? Are the questions on the earnings call about the old business or the new one? If a company is doing 40% of its growth in something the analysts keep skipping past, you’ve found a Blind Spot.

That said, a Blind Spot works both ways. Sometimes Wall Street ignores a new business that’s real but misunderstood. Other times, they ignore it because it’s a pipe dream and the CEO is full of crap. Your job is to tell the difference.

When a frontier industry grows fast, it needs stuff. Not just its own stuff. Other industries’ stuff. And it needs so much of it, so fast, that it eats every available unit and drives the price of that stuff through the roof.

The classic example is the 2017 crypto boom. Bitcoin and Ethereum miners needed GPUs. They bought every GPU in existence. Gamers couldn’t buy a graphics card for a year. Nvidia’s stock ripped, not because gaming got better, but because a totally separate industry ate their capacity.

Now look at AI:

• AI needs electricity. So much that big tech companies are signing deals to restart shuttered nuclear plants and buying up natural gas turbines. Turbine makers, who were boring utilities suppliers a few years ago, suddenly have multi-year backlogs.
• AI needs memory chips, specifically a type called HBM. A few memory companies that were cyclical, margin-squeezed businesses now can’t make enough.
• AI needs copper, transformers, cooling equipment, data center real estate, and even water rights.

None of these companies “do AI.” They just happened to make the thing AI needs, and AI ate their entire production capacity.

That’s Collateral Boom. A frontier industry’s growth spills over into an adjacent industry that simply can’t scale fast enough to meet it.

The good news: the input company trades at a low multiple for longer than it deserves because analysts still think of it as a boring cyclical.

The bad news: Collateral Booms can bust hard. Remember those 2017 GPUs? When crypto crashed in 2018, miners dumped millions of used graphics cards onto eBay. Nvidia’s stock fell by half. The boom came from outside, so the bust did too.

Data centers’ share of US electricity, 2023 vs 2030 (EPRI projection). Fifteen years of flat demand trained a generation of utility planners for a world that just ended.

When you’re excited about a frontier industry, don’t stop at “who makes the frontier thing?” Those stocks are usually priced for perfection.

Ask: “What does the frontier thing need that nobody’s paying attention to?” Then: “How long does it take to build a new factory for that?”

If the answer is “years,” you’ve found a bottleneck where profits pool.

Here’s the strange advantage you have over a $50 billion fund:

The analyst covering semiconductors is not allowed to spend his week studying electric utilities. The autos guy can’t write a report on labor markets. They’re brilliant people locked in sector-shaped cages.

You have no cage. You can follow a convergence across three sectors in one afternoon, just because you’re curious.

That’s not a small edge. In a market organized by boxes, the person who ignores the boxes sees the crashes first.

So don’t ask, “which sector’s hot?” Ask, “what’s about to hit what?”

The “next big thing” is already in motion.

Exo/Signals tracks frontier technologies before the breakout. Get the signals that move markets, early.