SoftBank to acquire DigitalBridge for $4B to scale AI data centers

SoftBank is turning its rhetoric about Artificial Super Intelligence into hard infrastructure bets. DigitalBridge isn’t a model lab or an apps company; it’s a sprawling portfolio of data centers, fiber, towers and edge sites that you need when you want to train and serve frontier models at global scale. By owning a major digital‑infrastructure manager outright, SoftBank is trying to internalize more of the AI compute stack instead of just renting from hyperscalers.

This changes the competitive map in two ways. First, it strengthens SoftBank’s hand as a capital allocator in AI: if you’re pouring tens of billions into OpenAI and Stargate‑style mega‑data‑centers, having an in‑house platform that can originate, finance and operate capacity worldwide is a big advantage. Second, it nudges AI infrastructure further toward vertical integration. Nvidia is moving down into networking and systems; cloud providers are moving into custom silicon; now SoftBank is pulling infrastructure up into an AI‑centric portfolio. That’s a recipe for more concentrated control over compute, which is the real scarce resource in the AGI race.

For other players, the signal is blunt: if you want to be taken seriously as an AGI contender, you can’t just talk about models. You need a credible plan for thousands of megawatts of power, land, fiber and cooling — or very tight partnerships with those who do.