The AI Race Isn’t About Speed Anymore—It’s About Ownership

There’s a strange disconnect happening in AI right now. On one side, researchers are publicly freaking out about the pace of development—models are getting bigger, training runs are getting shorter, and nobody really knows where the ceiling is. On the other, Mark Zuckerberg is framing the whole thing as a question of ownership. Not speed. Not capability. Ownership.

And honestly, he might be onto something.

The fear among researchers isn’t new. Every major lab has people who signed on to build useful tools and are now watching those tools get deployed faster than anyone can properly evaluate them. I’ve talked to people inside these companies who say the internal mood is less ‘excitement’ and more ‘we’re all just trying to keep our heads above water.’ The public narrative is about breakthroughs, but the private one is about whiplash.

Zuckerberg’s take cuts through some of that. When he talks about who owns AI, he’s not just talking about equity or corporate structure. He’s talking about control over the infrastructure, the data, the distribution channels—the whole stack that makes these models actually reach people. That’s a different kind of race, and it’s one where the biggest players have a massive head start.

Meanwhile, Black Forest Labs is doing something that feels almost quaint by comparison: building robots. The company behind Stable Diffusion has been quietly pushing into physical AI, and it’s a reminder that not everyone is chasing the same goal. Some people want the smartest model. Some people want the biggest platform. Black Forest seems to want the most capable physical system.

That’s not a bad bet, actually. The digital AI race is crowded and increasingly hard to differentiate. But physical AI—robots that can actually do things in the real world—is still wide open. The hardware is harder, the timelines are longer, and the capital requirements are steeper. But the moat is also deeper.

What worries me is that we’re conflating two very different questions. One is about capability: can we build AI that’s smarter than us? The other is about control: who gets to decide what that AI is used for? The researchers are asking the first question, and they’re scared. The executives are answering the second one, and they’re excited. Those two groups are not talking to each other.

And that’s the real problem. Not the race itself, but the fact that we’re racing without a shared destination.