SoftBank’s latest play: a robotics company that builds data centers, with a $100B IPO in sight

SoftBank is at it again — spinning off another ambitious bet. This time it’s a robotics company that builds data centers. And yes, they’re already eyeing a $100 billion IPO.

Let that sink in for a second. A company that doesn’t exist yet, targeting a valuation that would put it in the same league as Meta or Tesla. That’s classic SoftBank: swing big, worry about the details later.

But here’s the twist that caught my attention. The pitch is that you need infrastructure to build AI and robots, but you also need AI and robots to build that infrastructure. It’s a circular logic that actually makes sense when you think about it.

Data centers are getting monstrous. We’re talking facilities that consume as much power as a small city. Building them at scale is a logistical nightmare — labor shortages, supply chain bottlenecks, safety regulations. SoftBank’s bet is that robotics can automate large chunks of the construction process, from welding server racks to laying cables and even assembling modular components.

This isn’t a new idea, by the way. Construction robotics has been a thing for years. Companies like Built Robotics and Dusty Robotics have been trying to automate job sites. But no one has attempted this at the scale SoftBank is thinking about. They’re not just selling robots to contractors — they’re creating a vertically integrated company that both builds data centers and manufactures the robots that build them.

That vertical integration is the key differentiator. Most robotics startups struggle with adoption because construction firms are conservative and slow to change. SoftBank can bypass that by being its own customer. Build the robots, use them to build your own data centers, prove the economics, then sell the whole package to third parties. It’s the same playbook Tesla used with its Gigafactories — automate your own production first, then sell the automation.

The $100 billion IPO number feels like a stretch, but let’s be honest: SoftBank has a track record of pulling off absurd valuations. Arm’s IPO wasn’t that long ago, and it’s now worth over $100 billion. If this robotics company can show even a fraction of that momentum, the market will pay attention.

What I’m less sure about is the execution risk. Building a robotics company is hard. Building a data center construction company is hard. Doing both at the same time, while preparing for a massive IPO, is a recipe for chaos. SoftBank has the capital to absorb mistakes, but talent is another matter. Finding engineers who understand both robotics and large-scale construction is not easy.

Still, I can’t shake the feeling that this is the kind of moonshot the industry needs. AI infrastructure is becoming a bottleneck. If SoftBank can figure out how to build data centers faster and cheaper using robots, that’s a real contribution — not just another financial engineering trick.

The IPO target is ambitious, borderline delusional. But that’s SoftBank. They swing for the fences, and sometimes they connect.