Most FAANG engineers are useless in a startup environment. Here's why

Why they struggle

I have worked with FAANG engineers, and the pattern is often the same. They are spoiled by tooling and used to being told what to do next. Drop them into a startup and they over-engineer everything. Microservices and Kubernetes for what is basically a todo list. They optimize for scale problems the business will not have for years, if ever.

It is not that they are bad engineers. Big tech selected and trained them for a different environment: enormous scale, narrow scope, a lot of process, and a clear next ticket. A startup is the opposite. Tight budget, tight deadline, and no one to tell you what matters. You do not need perfection there, you need function, and a lot of them freeze when that is the bar.

Meta just cut around 8,000 jobs in a single round, more than 2,000 of them engineers. Ask yourself how many of them go on to start a company. Most will not. But even if 1 in 10 do, that is already a lot of new businesses, and this is just one layoff round.

The bet

So this is my bet. We are about to see a lot of successful startups led by software engineers partnering up with business people in their own communities.

That is the combination I want to see more of. An engineer who can build, paired with someone who knows the business and the customers. Not chasing the next FAANG. Building useful products for actual people, often the ones right around them. The bakery, the clinic, the local distributor.

When that happens across thousands of small circles, the shape of the economy changes. More owners, more independent businesses, fewer platforms capturing everything. That is what a more decentralized economy actually looks like on the ground.

The game, by the numbers

Most startups are not trying to become FAANG. When a startup wins, it almost always wins by being acquired, not by going public. In 2025, Crunchbase counted roughly 2,300 venture-backed acquisitions against just 65 IPOs. If you are raising now, you are far more likely to get bought than to ring the bell at the Nasdaq.

And most startups do not win at all. CB Insights tracked thousands of venture-backed companies and found around 70% end up dead or merely self-sustaining, with the odds of becoming a unicorn hovering near 1%. The FAANG-scale outcome is that 1% tail. So the realistic best case for a startup is a sale, and usually the buyer is a large incumbent.

Then what happens to the ones that get bought? A large chunk get shut down or quietly absorbed. Harvard Business Review puts the M&A failure rate at 70 to 90%, measured against the buyer's own goals. You get eaten by the machine. The product fades, the team scatters, the thing you built stops existing.

This is the part that matters. A Yale study found the share of venture-backed startups that exit by acquisition instead of going public went from about 10% to about 90% over three decades, and the researchers were specifically asking whether that wave of acquisitions is shielding big tech from competition. Instagram is the obvious case. Standalone today, it might have been a real threat to Meta.

That is the same caging I see at the individual level, just moved up to the company level. Keep the talent inside. Buy the threats before they grow. Sometimes shutting the acquired company down is the whole point.

All of this is legal. It is also why I think the engineer who learns to do business, and builds for their own community instead of building to be sold into that machine, is the more interesting bet right now.