OpenAI Stargate project reaches first steel milestone at Texas site
Original: First steel beams went up today at our Stargate site in Milam County, Texas. Exciting to see this project taking shape with @SoftBank and @SBEnergy. View original →
What is confirmed
On February 27, 2026, OpenAI co-founder Greg Brockman posted on X that the first steel beams had been installed at the Stargate site in Milam County, Texas. He also referenced SoftBank and SB Energy as partners involved in the project. In large-scale infrastructure programs, first-steel work is a meaningful transition point from planning and permitting into visible structural construction.
Why this matters for AI infrastructure
OpenAI has framed Stargate as part of a broader compute strategy to support rising demand for both model training and inference capacity. As frontier AI usage expands, competitive advantage increasingly depends on physical infrastructure: power procurement, land development, server deployment, and high-capacity networking. This update indicates that at least one site in that strategy is progressing beyond announcement stage and into material build-out.
Ecosystem implications
Hyperscale AI facilities affect more than one company’s roadmap. They influence regional power planning, construction supply chains, semiconductor and cooling logistics, and cloud capacity timelines. The explicit mention of both a capital partner and an energy partner reinforces a common pattern in next-generation AI build-outs: financing, energy, and compute must be coordinated early to avoid bottlenecks. That coordination will likely shape how quickly new capacity can be brought online.
What remains unknown
The post does not specify total site capacity, commissioning dates, customer allocation policy, or detailed operating parameters. So the currently verifiable facts are the first-steel milestone and named partners. Even with limited detail, the update is still notable: it suggests AI infrastructure expansion is moving from high-level investment narratives to execution checkpoints that can eventually constrain or enable model deployment at scale.
Source: Greg Brockman on X
Related Articles
OpenAI announced on January 15, 2026 that it is opening an RFP to expand US AI infrastructure. The company said it is prioritizing near-term 100MW+ capacity with pathways to 1GW+ scale and deeper domestic manufacturing support.
OpenAI and SoftBank Group said they are investing $1 billion in SB Energy, with OpenAI signing a 1.2 GW lease for initial data center buildout. The announcement ties the deal to Stargate and positions SB Energy as both an infrastructure partner and OpenAI customer.
NVIDIA announced the Rubin platform at CES 2026 in January. Comprising six new chips, the Vera Rubin superchip delivers 5x improved inference performance over GB200. Major AI companies including OpenAI, Meta, and Microsoft plan to adopt it.
Comments (0)
No comments yet. Be the first to comment!