Yandex Was the Best Company You Weren't Watching. Its Unraveling Is a Masterclass in Fragility.

There's a version of history where Yandex is remembered as one of the great engineering companies of its era. It built a search engine that competed with Google on home turf — the only one anywhere to do so at scale. It shipped neural-machine translation before it was fashionable, built self-driving cars that roamed real streets, and ran a ride-hailing and delivery empire whose ML stack was, by most accounts, genuinely ahead of Uber's in several dimensions.

Then, between 2022 and 2024, the whole thing was pulled apart. The Dutch holding entity was severed from its Russian assets in a deal valuing the international remnant at a fraction of its former self. The Russian operations were restructured under new ownership. Thousands of engineers scattered.

American founders tend to file this under "geopolitics, not my problem." That's a mistake. Yandex's rise and retreat contains at least three lessons that apply directly to companies that will never operate outside the US.

Lesson One: Vertical Integration Is a Superpower Until It's a Trap

Yandex famously built everything in-house: its own data centers, its own map data collection fleet, its own ML framework, its own hardware for autonomous vehicles. This gave it independence from Western cloud providers and a cost structure competitors envied.

But the same depth that made it resilient to vendor risk made it brittle to political risk. When the shock came, there was no clean way to separate the layers. The lesson isn't "don't build deep" — it's "know which seams you can cut along." Ask yourself: if a regulator, a sanctions regime, or a single court ruling demanded I sever a market, a supplier, or a technology tomorrow, where would the cut go? If the answer is "nowhere," you don't have a company. You have a monolith.

Lesson Two: Regulatory Risk Is a Product Requirement

Yandex spent years navigating a government that increasingly viewed its data and its algorithmic power as strategic assets. Every American founder reads about TikTok and thinks "that's a China thing." It's not. Data localization rules, AI regulation, antitrust pressure on platform companies, and state-level privacy laws are all converging on US tech companies right now.

The Yandex pattern is instructive: the pressure didn't arrive as one big law. It arrived as a drip — data rules here, content rules there, quiet pressure on leadership — until the cumulative weight made the structure unsustainable. Founders should map their drip now: which three regulatory changes, if they landed in the same quarter, would force a restructuring? What's the plan for that quarter?

Lesson Three: Talent Is the Asset That Walks

Perhaps the most underrated part of the Yandex story is what happened to its people. Many of its best engineers — the ones who built CatBoost, who scaled the search infrastructure, who ran the ML platform — are now at OpenAI, NVIDIA, Google, and a wave of startups. The knowledge left; the buildings stayed.

For founders, this reframes "infrastructure independence." Your true redundancy isn't your multi-region Kubernetes setup. It's whether the institutional knowledge in your team's heads is documented, distributed, and transferable. Yandex's knowledge survived its collapse because it lived in people who scattered and kept building. Your company's knowledge will survive whatever hits it only if you've invested in making it portable.

The Uncomfortable Meta-Lesson

Here's the part that should sting: at its peak, Yandex did almost everything right by conventional startup wisdom. It out-executed a global giant at home. It diversified into multiple successful business lines. It attracted elite talent. It built its own infrastructure.

And it still couldn't survive the environment changing underneath it. That's not an argument for paranoia. It's an argument for humility about what you actually control. Founders spend enormous energy on product-market fit and almost none on "jurisdiction-market fit" — whether the political and regulatory environment they've built on top of is durable.

You don't need Yandex's problems to learn from them. You just need to assume that, at some point in your company's life, the ground will shift in a way no roadmap predicted. The companies that make it are the ones that did the thought experiment in advance.

Yandex built brilliant machines. The machines couldn't protect the company from the world around them. Build accordingly.