Alphabet's $80 Billion AI Bet: What It Means for the Factory Floor

But the models have gotten better. Dramatically better. And $80 billion buys you a lot of compute, a lot of training data, and a lot of very smart people working on exactly the kind of generalisation problems that kept AI out of serious industrial applications. Google's already shown what they can do with warehouse robotics through their Everyday Robots work (before they folded that team into DeepMind, which, by the way, tells you something about their priorities).

The question I keep asking myself is whether this money flows into competing with the Fanuc and ABB robots of the world, or whether it flows into the software layer that sits on top of everyone's hardware. I suspect it's the latter, at least initially. Google's never been great at manufacturing physical things at scale. They're better at being the intelligence layer. The Android of robotics, if you want to think of it that way.

That said, I could be wrong. I called my old colleague at Siemens about this, and he thinks I'm underestimating how much Alphabet wants to own the full stack. "They've got the money to buy their way into hardware," he said. Fair point. $80 billion is acquisition money as much as it's R&D money.

What I do know is that the traditional automation vendors, the ones I spent my career working with and competing against, are watching this very carefully. When you've got a company that can raise more capital in a single equity offering than your entire industry generates in revenue, the competitive dynamics change. It's not that Alphabet is going to start selling palletising robots next year. But they might make the software that makes everyone else's robots dramatically more capable. Or they might buy someone who does.

The Berkshire involvement is the part that actually surprised me most. Buffett's historically avoided tech investments he doesn't understand. If his people are comfortable putting serious money into Alphabet's AI plans, that suggests they see this as infrastructure, not speculation. Railroads and utilities, not dot-com dreams.

I don't have a neat conclusion here. It's too early to say exactly how this reshapes industrial automation. But when this much money moves this decisively in one direction, you pay attention. I've seen enough technology transitions in my career to know that the winners are usually the ones who saw the wave coming, not the ones who insisted the old ways would hold.

James Chen · 25 Jun · 5 min