When President Donald Trump stood alongside Sam Altman, Masayoshi Son, and Larry Ellison in the White House on January 21, 2025, to announce a $500 billion artificial intelligence infrastructure initiative, skeptics abounded. Half a trillion dollars? For data centers? In four years?

Nearly one year later, the Stargate project has silenced many of those doubters. With $400 billion in committed investments across six announced U.S. sites, the venture is not just on track—it's ahead of schedule.

From Announcement to Execution

The scope of what Stargate has accomplished in twelve months is staggering by any measure of infrastructure development. The flagship site in Abilene, Texas is already under construction, with groundbreaking completed in spring 2025. Five additional sites have been announced, bringing the combined planned capacity to nearly 7 gigawatts of computing power.

To put that in perspective: 7 gigawatts could power approximately 5 million average American homes. Dedicating that much electricity to computing represents a bet that artificial intelligence will become as fundamental to the economy as electricity itself.

"This is the largest single infrastructure investment in American history, surpassing even the Interstate Highway System in inflation-adjusted terms. We're building the foundation for the next century of American technological leadership."

— Stargate project spokesperson

The Power Challenge

Perhaps the most impressive aspect of Stargate's progress has been solving the power problem. AI data centers are voracious electricity consumers, and finding locations with sufficient grid capacity has been the binding constraint on AI infrastructure expansion.

Stargate's solution involved a January 2026 partnership with Oracle that secured 4.5 gigawatts of power capacity, along with a separate $1 billion investment in SB Energy to develop dedicated power generation. The combination of grid access and proprietary power projects has allowed Stargate to plan at a scale that would have been impossible relying on existing infrastructure alone.

The Financial Structure

Understanding how Stargate is financing this unprecedented buildout reveals the bet that each partner is making. According to filings reported by The Information, SoftBank and OpenAI have each committed $19 billion of direct capital and hold 40% ownership stakes. Oracle and MGX (the UAE sovereign wealth fund vehicle) contributed $7 billion each, with remaining funds coming from limited partners and debt financing.

SoftBank CEO Masayoshi Son serves as chairman and bears "financial responsibility" for the venture, while OpenAI holds "operational responsibility." This structure means that if Stargate succeeds, OpenAI gains access to computing infrastructure that would cost competitors years and tens of billions to replicate. If it fails, SoftBank absorbs most of the financial pain.

The Geopolitical Dimension

Stargate isn't just a business venture—it's a geopolitical statement. The explicit goal is to ensure that the most advanced AI infrastructure resides on American soil, developed by American companies, using chips designed in America (if not always manufactured there).

This aligns with the broader technology competition with China that has intensified under the Trump administration. By concentrating AI computing capacity domestically, Stargate creates a strategic asset that would be difficult for any adversary to replicate or disrupt.

The project has also expanded internationally, with announced data centers in Norway, the United Arab Emirates, and the United Kingdom. These sites extend Stargate's reach while keeping the core infrastructure—and the most sensitive AI development work—in the United States.

Critics and Concerns

Not everyone is convinced that Stargate represents wise allocation of capital. Critics point to several concerns:

  • Execution risk: Infrastructure projects of this scale routinely face delays, cost overruns, and technical challenges. Stargate has avoided major setbacks so far, but the hardest construction work lies ahead.
  • Demand uncertainty: The entire project is premised on AI compute demand growing exponentially. If the AI boom plateaus or a technological breakthrough reduces compute requirements, Stargate could become an expensive white elephant.
  • Environmental impact: Data centers at this scale have massive carbon footprints. While Stargate has committed to renewable energy procurement, critics argue the project will still contribute significantly to climate change.
  • Concentration risk: Centralizing so much AI infrastructure in a single corporate venture raises antitrust and national security questions that regulators haven't fully addressed.

What's Next

The coming year will prove critical for Stargate's ambitions. Construction must accelerate at all six announced sites to meet the 2029 completion target. Power infrastructure must be built out in parallel. And customers—primarily OpenAI itself, but potentially others—must demonstrate the demand that justifies such massive investment.

OpenAI's own trajectory is intertwined with Stargate's success. The company is reportedly preparing for an IPO in 2026, and demonstrating access to virtually unlimited computing infrastructure could be a key selling point to public market investors.

The Comparison Point

Historical comparisons for Stargate are hard to find because nothing quite like it has been attempted before. The Manhattan Project, often cited by Stargate proponents, cost roughly $30 billion in today's dollars. The Interstate Highway System, built over decades, totaled about $600 billion adjusted for inflation.

Stargate is attempting to deploy similar resources in just four years, for a single purpose: ensuring that the United States leads the artificial intelligence revolution.

Whether that bet pays off will determine not just the fortunes of the companies involved, but potentially the technological trajectory of the 21st century economy. As the project approaches its one-year anniversary, the early returns suggest the gamble is at least being executed competently—even if the ultimate outcome remains far from certain.