Announcing the Stargate project, U.S. President Donald Trump speaks at the White House alongside ... [+] Larry Ellison of Oracle, Masayoshi Son of SoftBank and Sam Altman of OpenAI on January 21, 2025. (Photo by Jim WATSON / AFP via Getty Images)AFP via Getty ImagesOn the second day of his new term in office last week, President Donald Trump made a big splash by announcing Stargate, a new private joint venture that could have a major impact on AI innovation in the U.S. and beyond. Although many details of the venture are unclear at this early stage, Stargate intends to build up to 20 large AI datacenters in the U.S., with an initial investment of $100 billion and plans for up to $500 billion by 2029. Even without knowing all the specifics, this is clearly a monumental initiative aimed at reinforcing the U.S.s position as the undisputed global leader in artificial intelligence.Joining Trump for the announcement were Sam Altman, CEO of OpenAI, Larry Ellison, chairman of Oracle, and Masayoshi Son, CEO of SoftBank. Their three companies serve as the principals of the new joint venture. They are joined by a fourth equity investor, the AI-focused Emirati investment fund MGX, which is controlled by the government of Abu Dhabi. Besides OpenAI and Oracle, key technology contributors include Microsoft, Nvidia and Arm.So what else do we know about Stargate, and what could it mean for the U.S. position in AI?How Is Stargate Organized?On its website, OpenAI posted a 300-word announcement about Stargate that mentioned building new AI infrastructure for OpenAI in the United States, and said that it will enable creative people to figure out how to use AI to elevate humanity. In line with OpenAIs company mission, the announcement also (briefly) emphasized the development of artificial general intelligence. In the long run, AGI is intended to perform a wide array of tasks with human-like intelligence, potentially revolutionizing fields including materials science, finance, medicine and environmental science.For the project, SoftBank and OpenAI are categorized as lead partners . . . with SoftBank having financial responsibility and OpenAI having operational responsibility, including AI model development and training. SoftBanks Son will serve as chairman.The initial buildout of 10 datacenters has already begun. As announced nearly a year ago, an AI supercomputer equipped with tens of thousands of A100 GPUs from Nvidia is being built for OpenAI in Abilene, Texas by Microsoft. Once site evaluations are complete, the project will be expanded to other locations. The initial goal is to build 10 datacenters of 500,000 square feet each, then expand the project to another 10 datacenters across the United States.MORE FOR YOUFrom a collaborative perspective, the roster of partners makes sense. Microsoft has been a major backer and partner of OpenAI for years, and OpenAI has collaborated closely with Nvidia since 2016. Microsoft also knows Nvidia very well from using so many of its GPUs in Azure datacenters. The partnership between OpenAI and Oracle is more recent, but AI cant function without massive amounts of high-quality, accessible data, and Oracle will use its data-handling expertise to feed the Stargate datacenters. Having said all that, we dont yet know precisely how this computing environment will be configured.What Will Stargate Actually Do?Heres where the lack of detail leaves us with fewer answers and plenty of questions. Its sensible to think that the Stargate datacenters will be used in the development and operation of multiple AI models but what exactly will that look like?From an infrastructure perspective, will Stargate function as an R&D platform? An AI cloud? A GPU cloud? Or something more than that? Ten separate 500,000-square-foot datacenters racked with AI-specific infrastructure sounds an awful lot like a cloud, but maybe it isnt.If it is a cloud, what does Stargate mean for the cloud market? Anything? While were at it, will Stargate be reserved only for the largest use cases AGI, for instance that would typically require an on-premises cluster? Or will its net be spread wider to enable AI incubation for larger and smaller use cases and organizations?Connecting those questions to a customer perspective, what exactly is the product or service being offered? For that matter, who will have access to Stargate? At the launch event, Ellison spoke in general terms about the potential for personalized vaccines and cancer treatments. That is one of countless use cases that Stargate might support. However, it is fair to say that even a project as ambitious as Stargate cant support every use case for every type of customer. It would be very good to understand more about what Stargate will and wont do, and for whom.From a timeline perspective, the investment of $500 billion over four or five years is incredible. But how soon will the first customer, partner or other user actually get to use it? Regardless of the specific answer, it seems likely that significant ROI on this project will be years further out. Maybe the point is to ramp up to creating 100,000-plus well-paying jobs, secure in the knowledge that a thirsty AI market will find uses for these datacenters in short order.The Geopolitical And Economic Context Of The Global AI RaceBy this point in AIs development, its clear that AI will change the way we live our lives and conduct business, from individual consumers to the largest of organizations. While this observation is not in dispute, a host of questions remain about how AI will be developed, activated and regulated across the private and public sectors.Although the U.S. has led the world in AI innovation and development so far, the AI scene in other countries is also running hot. China, in particular, has aggressively incubated AI innovation both for domestic use and for export to the rest of the world. While the U.S. government has responded to Chinas AI rise with export restrictions and other measures, China has not slowed down its pursuit. For example, Alibabas Qwen model now rivals models from the U.S. On top of that, the past week has been dominated by news of the highly efficient open source Chinese model DeepSeek. Its maker claims to be able to train a model equivalent to GPT (an equivalence OpenAI denies) at a small fraction of the cost. Early news about DeepSeek seemed to spook investors in U.S. tech companies over the prospect that the new model could radically change the calculus of AI infrastructure investment.While there are many other areas of AI competition (and nations wanting to compete) that we could mention, these two examples give a taste of the intensity of the global AI race and its potential long-term impacts.Project Stargate Is About U.S. Dominance In AI But Challenges RemainOpenAIs published announcement addressed global AI competition crisply: This infrastructure will secure American leadership in AI, create hundreds of thousands of American jobs and generate massive economic benefit for the entire world. This project will not only support the re-industrialization of the United States but also provide a strategic capability to protect the national security of America and its allies.Kudos to the Stargate partners and the federal government for recognizing the need for the United States to stay far ahead of its competitors in the hugely important field of AI. This project should create significant economic and security benefits for the U.S., not least because of the impact of advanced AI, and especially AGI, on national and military security. This could enhance U.S. capabilities in data analysis, surveillance and cybersecurity to safeguard against strategic threats, which would be an additional justification for the giant investments in the project beyond simple economic gain.The company leaders who joined Trump at the press conference last week emphasized that his reelection played a key role in moving ahead with Stargate, with Son of SoftBank particularly vocal on that point. For his part, Trump praised the companies ambitious efforts and pledged to ease any concerns about getting adequate electricity to run the new Stargate datacenters.That said, no project is without its challenges, and thats especially true for something with the scale of Stargate. Out of the many potential hurdles that may arise, here are four that attract our attention.Sons enthusiasm notwithstanding, obtaining $500 billion for the project by 2029 may not be easy. Last week Elon Musk publicly disputed whether the group had raised even the initial $100 million, although Altman quickly rebutted him. (Its worth recalling that the two men have a history of disagreements.) If any funding problems do arise, we wonder whether the U.S. government might use economic and national-security justifications to step in and help arrange additional financing.President Trumps assurances about electricity are welcome, but energy is already a persistent issue for large AI datacenters. Although future sites are likely being chosen with available energy as a key criterion, OpenAI and its partners should still be considering how to tap alternative energy sources such as nuclear power or hydrogen if needed. Any energy shortages would likely impose delays on the project.There is already a shortage of AI expertise in the U.S. and beyond, and it is unknown how many new AI professionals will be required to build and manage Stargate.The real impacts of the DeepSeek announcement have yet to be determined. Among other open questions, we are unsure whether the claims about its low cost are accurate. However, we do know that large AI models have important advantages and capabilities presently beyond the reach of smaller models.Moving From More Questions To More AnswersThe Stargate project has clearly come a long way from the initial plan for Microsoft to build a supercomputer with a $100 billion price tag exclusively for OpenAI. To truly evaluate the project, we need more information on how the venture has evolved, how it is being implemented now and how the partners intend to address the infrastructure, market and timeline questions we posed earlier.When the President of the United States shines a spotlight on a big project let alone one as ambitious as this one people take notice. And it seems clear that if Stargate is implemented in line with what we already know, it should have a significant impact on U.S. competitiveness in the global AI race. We hope that as the project moves from plans to reality over the coming months and years, the partners (and maybe the government) will share more information so we can better understand the how, why, what, when and where of Stargate.