SoftBank to Acquire DigitalBridge for $4B in Strategic Move Toward AI Infrastructure

The acquisition of DigitalBridge by SoftBank reflects converging trends in technology investment: the rising demand for digital infrastructure to support AI, and SoftBank’s accelerating strategy to build capacity along the full stack—compute, connectivity, power, and real estate. [1][2] DigitalBridge brings a rich portfolio of data centers, fiber networks, satellite towers, edge infrastructure and more, managed across multiple global geographies. [2][4]

While US$4.0 billion represents SoftBank’s valuation of DigitalBridge, this figure includes both equity and debt; the “enterprise value” over the company’s existing capital structure. [1][4] Paying US$16.00 per share reflects both market optimism driven by acquisition speculation, as well as a premium for control. The 50% premium vs. DigitalBridge’s unaffected 52-week average suggests large potential upside baked into speculative/bullish AI infrastructure narratives. [4]

Strategically, SoftBank gains immediate scale in AI-adjacent infrastructure—an increasingly scarce asset class globally. This deal accelerates its ability to deploy hyperscale AI applications, enhances its participation in Project Stargate (a multibillion-dollar AI infrastructure initiative), and could serve as a hedge against hardware shortages or supply chain constraints. [5][2]

Key risks include regulatory headwinds (expected closing in H2 2026 pending approvals), integration challenges—though DigitalBridge will remain independently managed—and macro-factors such as energy costs, land-resource constraints, and the potential for overcapacity if AI infrastructure demand slows. Also, while the premium is high, the share price behavior prior to the announcement had already reflected takeover speculation, possibly leading to overpayment risk. [4][3]

The deal’s timing is notable: with recent SoftBank divestitures (e.g. its NVIDIA stake) and realignments, this move deepens its physical AI infrastructure capacity just as compute hardware and data bandwidth are emerging as binding constraints in the AI ecosystem. [1][2][5]