The Applied Intuition Playbook

COMMANDER · ISSUE 16 · October 5, 2026 · Istanbul

In 2013, two Google product managers sat down to build a self-driving car and talked themselves out of it. Qasar Younis — a former GM engineer who had sold a startup to Google and would go on to run Y Combinator — and Peter Ludwig, an automotive-software engineer from a family of GM men, ran the numbers and concluded that building a Level-4 autonomous vehicle would take the capital and scale of an automaker or a tech giant, and they were neither. When GM bought Cruise in 2016, it confirmed the math. So in July 2017 they founded Applied Intuition to build the opposite of a car company: the simulation and validation software that every team trying to build an autonomous vehicle would need, whether that team ever shipped a car or not. Eight years later the company runs at roughly $830 million in annual recurring revenue — double the prior year — at about 85% gross margins, counts eighteen of the world's twenty largest automakers as customers, and was valued at $15 billion in June 2025. It has never sold a vehicle.

The last three issues were about builders — the ship, the weapon, and the factory that makes the parts. Applied Intuition is the one that barely built anything at all. While hundreds of companies poured billions into the hardware of autonomy, Applied Intuition sold them the one thing none of them could ship without and none of them wanted to build themselves: the software to prove the hardware actually works. Everyone treated validation as an afterthought. Applied Intuition treated it as the business — and the gap between what a company builds and how it built it is the entire reason Commander exists. This is that gap in its purest form: a company you can't understand by looking at a product, because it doesn't really have one you can hold.

The business behind the hardware. Not what they build, but how they built it.

Younis and Ludwig started from a decision most founders in a hot market never make: they decided not to compete in it. Autonomy in 2017 was a capital bonfire — Waymo, Cruise, and a dozen well-funded startups spending billions to put a robot on the road. Applied Intuition sold them shovels instead. The product was the infrastructure every one of those teams needed regardless of who won: scenario generation, sensor simulation, the ability to test an autonomy stack against millions of virtual miles before a single real one. You didn't have to pick the winner of the AV race. You sold to everyone racing.

The strategic point is that the tool layer is larger and safer than any single robot. A company betting on one autonomous vehicle lives or dies on that vehicle; a company selling the tools lives on the size of the field. When the field is every automaker, every trucking firm, and eventually every defense program trying to make a machine move on its own, the picks-and-shovels position is not the humble one. It is the biggest seat at the table.

Takeaway: When a market is a capital bonfire, don't bring more fuel — sell the tools everyone burning it needs. The shovel layer outlives any single robot, and it doesn't require you to pick the winner.

The glamorous part of autonomy is the demo — the car driving itself down the street for the cameras. The unglamorous, mandatory part is proving it is safe enough to ship, which means testing it against millions of edge cases no road can produce on schedule. Applied Intuition planted itself on the mandatory part. Its early products — Simian for simulation, Spectral for sensor modeling — addressed the validation problem every serious autonomy program is forced to solve and none can hand-wave past.

This is a better place to stand than the demo. Owning the step a customer wants to do makes you a nice-to-have; owning the step a customer is operationally and legally required to do makes you infrastructure. No automaker ships a driver-assistance system it hasn't validated, and validation at scale is only tractable in simulation. By owning the checkpoint everyone must pass through, Applied Intuition made itself unavoidable — not by being the most exciting vendor, but by being the one you can't route around.

Takeaway: Own the step your customer is forced to take, not the one they show investors. The mandatory checkpoint is more defensible than the flashy feature — you become infrastructure instead of an option.

Everyone else in autonomy was building capital-heavy physical things — vehicles, sensors, compute. Applied Intuition sold software, and the economics show it: roughly 85% gross margins and an asset-light model where the core product is code, not steel. That ARR ran to about $830 million in 2025, doubling from $415 million the year before. Those are software-company numbers earned inside a hardware industry that mostly can't produce them.

The lesson is about where the margin lives. In any hardware sector — cars, ships, drones, defense — the physical layer is expensive, slow, and thin-margined, while the software and validation layer on top is cheap to replicate and priced like a subscription. Applied Intuition chose the layer with the best economics and the best scaling, then sold it to the people stuck with the expensive layer. The highest-margin seat at a hardware table is almost always the software one.

Takeaway: In any hardware market, find the software layer and own it. The metal is expensive and thin-margined; the code on top scales at 85% — be the layer, not the machine.

Applied Intuition doesn't sell a monolith. It licenses software by engineering seats and modules, so a customer can start small — one team buying Simian for basic simulation — and grow into the rest of the platform as the relationship deepens: sensor modeling, data tooling, a full vehicle operating system. That land-and-expand motion is how the company came to serve eighteen of the top twenty global automakers, names like Toyota, GM, Hyundai, Porsche, Audi, and the TRATON truck group among them.

The mechanics matter for anyone selling into large, cautious buyers. A small first module is easy to approve, hard to regret, and impossible to fully rip out once a team builds its workflow around it. From that beachhead, every new module is an upsell into an account that already trusts you, at a fraction of the cost of winning a new logo. The company grows inside its customers, not just across them — which is why revenue can double without the sales cost doubling with it.

Takeaway: Get in cheap on one workflow, then colonize the account. A small, easy-to-approve first module becomes the beachhead every future upsell launches from — grow inside customers, not just across them.

Applied Intuition built its technology, its customer base, and its cash in automotive — then walked the same autonomy and simulation stack into defense. It won a $49 million US Army and Defense Innovation Unit contract for the Robotic Combat Vehicle program back in 2022, and in February 2025 acquired EpiSci, a tactical-autonomy software company working across air, land, sea, and space, stapling a defense-native capability onto a commercially proven platform. A UK subsidiary followed with £50 million in planned investment aimed at sovereign programs.

The direction of travel is the whole point. Most companies try to go defense-first and then find a commercial market; Applied Intuition did the reverse, and the reverse is easier. The commercial business funded the R&D, hardened the product against eighteen demanding automakers, and generated the margins that make a defense expansion optional rather than existential. When you arrive at the Pentagon with a stack already validated by industry at scale, you are not a science project asking for a development contract — you are a mature vendor offering a proven tool.

Takeaway: Build in the big commercial market first, then walk the mature product into defense — not the reverse. Commercial scale funds the R&D, hardens the product, and lets you meet the government as a proven vendor rather than a science project.

Applied Intuition's acquisitions read as a map of domains, not a list of businesses. Embark ($71 million) extended it into trucking. EpiSci carried the platform into air, sea, and space autonomy. Reblika added real-time AI-agent simulation; exclusive rights to Ghost Autonomy's patent portfolio locked in intellectual property. Each deal bolted a new capability or domain onto the same core platform rather than buying a company to run as a separate line.

The discipline is buying reach for the thing you already own, not entering businesses you can't operate. A platform company's acquisitions should make the platform work in more places — new domains, new sensors, new customers — while the core engine stays single. Done that way, every acquisition compounds the same product instead of fragmenting the company into a conglomerate of half-integrated parts. Applied Intuition buys to extend the surface area of one system, which is the only kind of M&A that makes a platform stronger instead of heavier.

Takeaway: Acquire to extend the platform's reach, not to enter businesses you can't run. Each deal should make the one core product work in more places — buy surface area, not a portfolio.

Applied Intuition's $600 million Series F in June 2025 had a detail worth reading twice: only $200 million was primary capital going into the company. The other $400 million was secondary — buying shares from existing holders and employees. A business raising mostly secondary is not a business scrambling for runway. It is one profitable and cash-efficient enough that the round exists to set a $15 billion price and reward the people who got there early, not to keep the lights on.

That is the strongest position a founder can raise from. When your operating margins mean you don't need the money, capital stops being oxygen and becomes a tool — you take it to establish valuation, to buy liquidity for your team, to signal strength, and on your own terms rather than the market's. The companies that raise ahead of catastrophe take whatever price they can get; the ones that raise ahead of nothing in particular set the price themselves. Applied Intuition's discipline in the operating model is what bought it discipline in the cap table.

Takeaway: The strongest raise is the one you don't need. When your margins cover you, capital becomes a tool for setting price and rewarding your team — not oxygen you take on someone else's terms.

Refuse to compete in the capital bonfire — sell the tools everyone burning it needs. Plant yourself on the mandatory step, the validation nobody is allowed to skip, so you become infrastructure instead of an option. Bring software's 85% margins into a hardware industry that can't produce them. Land on one cheap module and expand into the whole account until revenue doubles without the sales cost following. Earn your technology, customers, and cash in the commercial market, then walk the proven stack into defense rather than starting there. Acquire to extend one platform's reach across domains, never to run a portfolio. And raise from strength — when the money is a tool, not oxygen. The last three archetypes won by building the hardware; this one won by selling everyone else the software to make theirs work.

Four issues, one thread: the ship, the weapon, the factory, and the software that proves the machines work — three builders and the one that refused to build. The next archetype goes back to hardware, but the hardest kind: an aircraft that sixty years of programs tried and failed to build. This company's answer is to refuse to build the final product at all — and climb toward it one cheap, flyable prototype at a time. No company names yet. Collect the set: each issue is one operating archetype, and this one — own the software, skip the hardware — sets up the archetype that sequences the impossible.

— COMMANDER

———————————————————————————

A new Playbook every Monday, the Weekly Transmission every Friday. If it earned its place in your inbox, forward it to one operator who's building something — that's how Commander spreads.