SoftBank, DigitalBridge, and Stargate: The Next Phase of OpenAI’s Infrastructure Strategy

OpenAI framed Stargate as an AI infrastructure platform; a mechanism to secure long-duration, frontier-scale compute across both training and inference by coordinating capital, land, power, and supply chain with major partners. When OpenAI announced Stargate in January 2025, the headline commitment was explicit: an intention to invest up to $500 billion over four to five years to build new AI infrastructure in the U.S., with $100 billion targeted for near-term deployment.

The strategic backdrop in 2025 was straightforward. OpenAI’s model roadmap—larger models, more agents, expanded multimodality, and rising enterprise workloads—was driving a compute curve increasingly difficult to satisfy through conventional cloud procurement alone. Stargate emerged as a form of “control plane” for:

• Capacity ownership and priority access, rather than simply renting GPUs.

• Power-first site selection, encompassing grid interconnects, generation, water access, and permitting.

• A broader partner ecosystem beyond Microsoft, while still maintaining a working relationship with Microsoft for cloud capacity where appropriate.

2025 Progress: From Launch to Portfolio Buildout

January 2025: Stargate Launches as a National-Scale Initiative

OpenAI publicly launched Project Stargate on Jan. 21, 2025, positioning it as a national-scale AI infrastructure initiative. At this early stage, the work was less about construction and more about establishing governance, aligning partners, and shaping a public narrative in which compute was framed as “industrial policy meets real estate meets energy,” rather than simply an exercise in buying more GPUs.

July 2025: Oracle Partnership Anchors a 4.5-GW Capacity Step

On July 22, 2025, OpenAI announced that Stargate had advanced through a partnership with Oracle to develop 4.5 gigawatts of additional U.S. data center capacity. The scale of the commitment marked a clear transition from conceptual ambition to site- and megawatt-level planning.

A figure of this magnitude reshaped the narrative. At 4.5 GW, Stargate forced alignment across transformers, transmission upgrades, switchgear, long-lead cooling equipment, and—critically—financing structures capable of absorbing multi-year construction timelines and staged ramp schedules.

September 2025: Five New Sites Define a Clearer U.S. Footprint

On Sept. 23, 2025, OpenAI, Oracle, and SoftBank Group announced five new U.S. Stargate data center sites, describing a program nearing 7 GW of planned capacity and more than $400 billion in projected investment over the next three years.

Subsequent reporting clarified the site mix: three Oracle-led developments, including Shackelford County, Texas, and Doña Ana County, New Mexico, alongside two additional sites tied to SoftBank and its affiliates, including Lordstown, Ohio, and Milam County, Texas.

OpenAI later confirmed that the initially referenced “Midwest” site would be located in Wisconsin, developed by Oracle in partnership with Vantage Data Centers.

By this point, Stargate was clearly evolving from a single flagship build into a portfolio strategy: multiple campuses, multiple utility territories, diversified construction schedules, and, implicitly, diversified political and regulatory risk.

October 2025: Supply Chain Alignment and Campus Expansion

Samsung and SK join (Oct. 1, 2025)
OpenAI announced that Samsung and SK hynix had joined Stargate-related efforts, framed around scaling advanced memory production and broader AI infrastructure enablement.

The announcement underscored a defining 2025 constraint: high-bandwidth memory had become one of the primary governors of effective GPU deployment. Participation in Stargate was as much about long-term capacity assurance across the HBM supply chain as it was about signaling alignment with OpenAI’s infrastructure ambitions.

Wisconsin campus (Oct. 22, 2025)
OpenAI, Oracle, and Vantage confirmed plans for a Stargate campus in Wisconsin, explicitly tying the site to the earlier 4.5-GW expansion announced in July.

Michigan campus (Oct. 30, 2025)
OpenAI announced a Stargate campus in Saline Township, Michigan, stating that the overall program now exceeded 8 GW in planned capacity and $450 billion in projected investment across previously announced U.S. sites.

The Michigan project quickly became a case study in “AI infrastructure meets local process.” Coverage noted that regulators approved power delivery—reported at roughly 1.4 GW—in a manner that triggered public backlash over transparency and community impact, highlighting the role of local governance in shaping AI infrastructure timelines.

December 2025: SoftBank Strengthens Its Role Through DigitalBridge

On Dec. 28, 2025, SoftBank announced its acquisition of DigitalBridge, bringing approximately $108 billion in digital infrastructure assets; including data centers, fiber networks, and edge platforms—under its control.

The move marked a shift in SoftBank’s role within Project Stargate, from primarily a financial participant to a stakeholder with direct exposure to infrastructure ownership and operations, broadening and deepening its alignment with OpenAI.

International Expansion: UAE First, Argentina Explored

In May 2025, OpenAI announced Stargate UAE, describing a 1-gigawatt compute cluster to be built by G42 and operated by OpenAI and Oracle, with Cisco and NVIDIA named among key participants.

In October 2025, Reuters reported that OpenAI had signed a letter of intent with Sur Energy to explore “Stargate Argentina,” a proposed $25 billion project targeting up to 500 MW. The announcement framed Argentina as Stargate’s first potential foothold in Latin America.

The Less Glamorous—but Decisive—Work of 2025

Financing: Why Bankability, Not GPUs, Became the Constraint

At Stargate’s scale, progress rarely takes the form of ribbon cuttings. More often, it appears as debt facilities, equity partnerships, and contract structures (such as take-or-pay power agreements and long-term leases) that determine whether projects can move from paper to steel.

By late 2025, reporting pointed to financing volatility around specific campuses. Coverage by the Financial Times and Reuters, for example, described uncertainty surrounding Oracle’s roughly $10 billion Michigan project after talks with a key backer reportedly stalled. The episode illustrated a core reality of Stargate’s approach: even with massive top-line commitments, each campus must independently clear underwriting and risk-allocation hurdles.

Grid Interconnection and Permitting: The Quiet Schedule Drivers

The Michigan experience also underscored a broader 2025 reality. Grid interconnection timelines, utility approvals, and local buy-in increasingly functioned as gating items for AI infrastructure delivery. “AI urgency” alone does not compress public utility oversight, regulatory review, or community process: factors that can materially shape schedules regardless of capital availability or demand signals.

What 2025 Progress Meant for OpenAI—Beyond the Stargate Program Itself

Compute optionality beyond Microsoft.
Stargate gave OpenAI meaningful leverage. The company could add capacity through Oracle- and SoftBank-led developments while continuing to rely on incumbent cloud relationships, most notably Microsoft, where they remained strategically or operationally advantageous.

A shift from cloud customer to infrastructure orchestrator.
Across 2025 announcements, OpenAI was consistently positioned not as a tenant, but as the operational lead and architectural coordinator of a multi-partner industrial buildout spanning sites, power procurement, supply chain alignment, and complex financing structures.

A playbook for the AI factory era.
The cadence of 2025 vis a vis gigawatt-scale capacity commitments, multi-site portfolios, and early international replication made Stargate resemble a standardized rollout model rather than a one-off mega-campus. The project increasingly looked like a template for how frontier AI infrastructure will be deployed at scale.

The 2025 Bottom Line

By the end of 2025, OpenAI had moved Stargate from announcement to mapped portfolio: multiple U.S. sites across Texas, New Mexico, Ohio, Wisconsin, and Michigan, alongside early international nodes in the UAE and exploratory efforts in Argentina.

The clearest indicators of progress were not completed facilities, but gigawatt-level capacity commitments, named partners, identified sites, and the start of the complex utility, financing, and public-process work that will ultimately determine whether deliveries between 2026 and 2028 arrive on schedule.

How SoftBank’s DigitalBridge Deal Reframes Stargate’s Infrastructure Stack

In late December 2025, SoftBank Group materially reshaped the infrastructure context around Project Stargate with its aforementioned acquisition of DigitalBridge, bringing roughly $108 billion in digital infrastructure assets spanning data centers, fiber networks, and edge platforms under SoftBank’s control.

For Stargate, the significance of the move extended beyond balance-sheet strength. Until then, SoftBank’s role in the initiative had been understood primarily as financial sponsorship. With DigitalBridge, SoftBank gained direct exposure to infrastructure ownership and operations, shifting its position from capital provider to embedded infrastructure stakeholder.

That shift matters as Stargate transitions from capacity announcements to delivery.

From Financial Backing to Infrastructure Alignment

Stargate’s execution depends on aligning three difficult variables at once: long-duration capital, builders capable of delivering at hyperscale speed, and operating platforms familiar to lenders, utilities, and regulators. DigitalBridge’s portfolio gives SoftBank experience across all three, strengthening its ability to support Stargate as projects move deeper into construction, interconnection, and commissioning phases.

This also sharpens the contrast between Stargate’s infrastructure ambitions and the more fragmented funding environment that characterized parts of 2025.

Builders and Capital Partners: Crusoe and Blue Owl in Context

Several Stargate-adjacent developments over the year illustrated how sensitive AI infrastructure delivery remains to capital structure and risk tolerance.

Crusoe, for example, emerged as a notable builder and operator aligned with power-constrained, speed-to-market deployments, particularly where on-site generation and unconventional energy strategies were required. Crusoe’s model - combining development, construction, and operational execution - fit naturally within Stargate’s broader emphasis on power-first site selection and vertically coordinated delivery.

By contrast, Blue Owl Capital illustrated the other side of the equation. Blue Owl had been a prominent financier of large-scale Oracle-linked data center developments, but reporting late in 2025 pointed to a pullback from at least one major Oracle project, underscoring how underwriting assumptions can shift even amid strong AI demand signals.

That pullback did not invalidate Stargate’s strategy, but it did reinforce a key lesson: at gigawatt scale, AI infrastructure is only as durable as the alignment between builders, operators, lenders, and long-term asset owners.

Why DigitalBridge Changes the Risk Profile

Against that backdrop, SoftBank’s acquisition of DigitalBridge reduces one layer of uncertainty. DigitalBridge brings with it a portfolio of operating assets, established development pipelines, and institutional familiarity with the kinds of long-duration cash flows lenders expect from infrastructure-class investments.

For OpenAI, this strengthens Stargate in several ways:

• Greater execution resilience, as infrastructure ownership and operational experience sit closer to the capital stack.
• Reduced dependence on single funding channels, particularly as individual financiers reassess risk exposure.
• Improved global optionality, as Stargate expands beyond the U.S. into regions where DigitalBridge already understands regulatory and operating environments.

Importantly, this does not make SoftBank the owner of Stargate, nor does it eliminate execution risk. OpenAI remains the orchestrator, responsible for defining architectural requirements, coordinating partners, and pacing deployments against model demand.

What the DigitalBridge deal does is narrow the gap between AI ambition and infrastructure reality. In an environment where even well-capitalized projects can stall due to financing friction, grid constraints, or shifting risk appetites, that alignment increasingly becomes a competitive advantage.

Oct. 6, 2025 — In a live CNBC broadcast, Jensen Huang, CEO of NVIDIA, joined Sam Altman, CEO of OpenAI, and OpenAI President Greg Brockman to outline a new strategic partnership aimed at enabling AI at unprecedented scale. Under the agreement, OpenAI plans to deploy at least 10 gigawatts of NVIDIA systems for next-generation AI infrastructure, including the NVIDIA Vera Rubin platform, while NVIDIA intends to invest up to $100 billion in OpenAI as capacity is brought online.

At Data Center Frontier, we talk the industry talk and walk the industry walk. In that spirit, DCF Staff members may occasionally use AI tools to assist with content. Elements of this article were created with help from OpenAI's GPT5.

Keep pace with the fast-moving world of data centers and cloud computing by connecting with Data Center Frontier on LinkedIn, following us on X/Twitter and Facebook, as well as on BlueSky, and signing up for our weekly newsletters using the form below.