AI Companies Keep Buying Their Way Into New Industries

Something interesting happened this week that didn't make headlines, but should have. While everyone was busy analyzing Google's Gemma 4 release and OpenAI's latest funding round, three separate acquisition announcements flew under the radar—and together, they tell a story about where the AI industry is actually heading.

OpenAI acquired TBPN to "accelerate global conversations around AI." Sony Interactive Entertainment bought Cinemersive Labs, a UK startup that converts 2D images into 3D volumes. And Shift Up, makers of Stellar Blade, acquired Shinji Mikami's Unbound Inc. At first glance, these deals seem unrelated. Look closer, and you'll see the same strategy playing out across the industry: AI companies have decided it's faster to buy expertise than to develop it internally.

This isn't necessarily a bad thing. Acquisitions have always been part of tech's playbook. But what's notable is the speed and breadth of these moves. We're watching AI companies rapidly expand into domains that have nothing to do with their core business. Sony's gaming division isn't just improving graphics—they're fundamentally changing how 3D content gets created. OpenAI isn't content being an AI research lab—they want to own the platforms where AI gets discussed.

The acquisition spree reveals something deeper about the current state of AI development. Despite all the talk about frontier models and breakthrough capabilities, the real competition isn't happening in model architecture anymore. It's happening in application layers. The companies winning aren't the ones training the biggest models—they're the ones figuring out what to do with them.

Consider what Sony is really buying with Cinemersive Labs. It's not the technology itself—Sony has plenty of talented engineers who could probably replicate the 2D-to-3D conversion tech given enough time. What they're buying is time to market and domain expertise. The Cinemersive team already knows the pain points of 3D content creation. They've already made the mistakes and learned the lessons. That knowledge is worth more than the code.

This strategy makes particular sense right now because we're entering what I'd call the "integration phase" of AI development. The foundational models exist. The APIs work. The remaining question is: how do we actually make this useful? That's not a research problem—it's a product problem. And product problems get solved by people who understand specific industries deeply.

The risk, of course, is that we end up with a handful of mega-companies controlling every layer of the AI stack, from the models themselves to the specialized applications built on top of them. When OpenAI acquires a media platform, when Sony absorbs a 3D imaging startup, when gaming conglomerates buy up independent studios—each deal individually seems reasonable. Collectively, they're concentrating power in ways that should make us pause.

What's missing from this acquisition frenzy is the middle ground. Where are the independent AI application companies that can survive and thrive without getting absorbed? Where's the ecosystem of specialized AI tools built by domain experts who aren't racing to an exit?

The next few months will likely bring more of these deals. AI companies have raised record amounts of capital—OpenAI just secured $122 billion—and that money needs to go somewhere. Building from scratch is slow. Acquiring is fast. The math is simple.

But speed isn't everything. The most interesting AI applications will come from people who deeply understand a problem domain and use AI as a tool, not as the entire solution. Those people are increasingly finding themselves with two options: get acquired, or compete against companies with functionally unlimited resources.

We should be rooting for more of them to choose the latter.