OpenAI teams up with Oracle and SoftBank to grow Stargate through five additional AI datacenter locations
OpenAI, Oracle, and SoftBank announced five new datacenter sites for their Stargate project, bringing the total to eight US locations. The $500 billion initiative aims for 10 gigawatts of AI computing capacity and is projected to create tens of thousands of jobs. While the expansion addresses the critical compute bottleneck in AI development and accelerates an industry-wide infrastructure race, questions remain about energy demand, water usage, environmental impact, and effects on local electricity prices.
For anyone following the race to build the physical backbone of artificial intelligence, a major update just landed. OpenAI, Oracle, and SoftBank have announced plans to add five new datacenter sites to their Stargate project, pushing the initiative closer to its ambitious goal of 10 gigawatts of computing capacity across the United States. This article breaks down what the Stargate project is, why these new sites matter, and what the expansion means for the AI industry and the broader economy.
What Is the Stargate Project?
Stargate is a large-scale infrastructure partnership between OpenAI, Oracle, and SoftBank, launched earlier this year with a stated investment of $500 billion. The goal is straightforward but enormous: build the datacenters, power systems, and networking capacity needed to train and run next-generation AI models inside the United States.
AI models have grown dramatically in size and capability, and that growth depends on one thing above all — computing power. Training frontier models requires tens of thousands of specialized chips working together in massive facilities that consume enormous amounts of electricity. Stargate exists to create that capacity at a scale no single company has attempted before.
The project got underway with an initial site in Abilene, Texas, and additional locations were announced in the months that followed. The five new sites bring the total to eight planned datacenter locations nationwide.
What the New Sites Add
With the five new locations included, the project is on track to reach roughly 10 gigawatts of planned capacity, the full commitment made when Stargate was unveiled. For context, a gigawatt is enough electricity to power hundreds of thousands of homes — dedicating that much power to AI computing gives a sense of the scale involved.
The buildout is also moving faster than many expected. OpenAI says the project hit its initial capacity targets ahead of schedule, and the new sites are part of an accelerated timeline. The locations of the newly announced sites span several states, chosen in part for access to power, land, and workforce.
Beyond compute capacity, the expansion carries a significant jobs component. The companies project tens of thousands of American jobs connected to construction, facility operations, and related roles. Datacenter projects of this size typically employ thousands of construction workers during the build phase, followed by long-term operational staff once facilities come online.
Why This Expansion Matters
It addresses a real bottleneck. One of the biggest constraints on AI development today is not talent or ideas — it is compute. Access to large-scale computing capacity has become a competitive advantage, and companies that control it can train bigger models and serve more users. By building its own infrastructure, OpenAI reduces its dependence on third-party capacity and gains more control over its roadmap.
It signals long-term commitment. A $500 billion pledge is easy to announce but hard to execute. Hitting early milestones and expanding the number of sites suggests the partnership is moving from press release to reality, which matters for investors, partners, and policymakers watching the project closely.
It raises the stakes for the industry. Competitors including Google, Meta, Microsoft, and Amazon are all pursuing their own massive datacenter expansions. Stargate's acceleration adds momentum to an infrastructure arms race that will shape how quickly AI capabilities advance over the next decade.
Open Questions and Concerns
The announcement is not without unanswered questions. Projects of this scale raise legitimate concerns about energy demand and where the electricity will come from, especially as grids in some regions are already under strain. Water usage for cooling, local environmental impacts, and effects on electricity prices for nearby residents are all topics that local communities and regulators will scrutinize.
There is also the question of financing. While $500 billion is the headline figure, actual spending will unfold over years and depends on continued demand for AI products. If revenue growth slows, the pace of construction could too. For now, though, the companies involved appear committed to the full buildout.
What Comes Next
Construction on the new sites is expected to begin as planning and permitting are completed, with capacity coming online in phases over the next several years. As sites go live, they will progressively expand the computing power available for OpenAI's models, which in turn affects how quickly new capabilities reach everyday users.
For now, the five new Stargate sites mark a concrete step in what may become the largest AI infrastructure project ever attempted — one whose success or shortfall will help determine the pace of AI progress for years to come.
Comments (0)