Running AI products demands enormous computing power, and as companies race to harness the potential of AI models, there’s a simultaneous race to build the infrastructure that will support them. On a recent earnings call, Nvidia CEO Jensen Huang estimated that AI infrastructure spending could reach $3–$4 trillion by the end of the decade, with much of that investment coming from AI-focused companies. This surge is putting significant strain on power grids and pushing the limits of the industry’s building and operational capacity.

Figure 1. Trillion-Dollar AI Infrastructure Deals Powering the Boom.

Below, we summarize the largest AI infrastructure projects to date, highlighting major investments from Microsoft, Oracle, Google, Meta, and OpenAI. This list will be updated as the AI boom continues and investment figures climb even higher. Figure 1 shows Trillion-Dollar AI Infrastructure Deals Powering the Boom.

Microsoft’s $1 Billion Bet on OpenAI

The arrangement benefited both parties: Microsoft boosted its Azure sales, while OpenAI received funding for its largest expenses. Over time, Microsoft increased its investment to nearly $14 billion—a stake set to pay off substantially when OpenAI transitions to a for-profit structure.

The partnership has since evolved. In January, OpenAI announced it would no longer rely exclusively on Microsoft’s cloud, instead granting the company first refusal on future infrastructure needs while exploring alternatives if Azure couldn’t meet demand. Microsoft has also begun developing its own foundation models, reducing reliance on OpenAI.

OpenAI’s successful collaboration with Microsoft has become a model for AI companies partnering with cloud providers. Anthropic, for example, secured $8 billion from Amazon and modified its hardware for AI workloads. Google Cloud has also formed primary computing partnerships with smaller AI firms like Lovable and Windsurf, though without investment. More recently, OpenAI received a $100 billion investment from Nvidia, giving it the capacity to acquire even more GPUs.

Oracle Steps into the AI Infrastructure Race

On June 30, 2025, Oracle disclosed in an SEC filing that it had signed a $30 billion cloud services deal with an unnamed partner—exceeding the company’s entire cloud revenue from the previous fiscal year. The partner was later revealed to be OpenAI, securing Oracle a position alongside Google as one of OpenAI’s key hosting providers following its Microsoft-era partnership. The announcement triggered a significant surge in Oracle’s stock.

A few months later, history repeated itself. On September 10, Oracle announced a staggering five-year, $300 billion deal for compute power, scheduled to start in 2027. The announcement sent Oracle’s stock soaring, briefly making founder Larry Ellison the richest person in the world. The scale of the deal is almost unbelievable: OpenAI does not currently have $300 billion to spend, implying massive anticipated growth for both companies—and a healthy dose of optimism.

Even before a single dollar is spent, the agreement has cemented Oracle’s status as a leading AI infrastructure provider and a formidable financial force.

Nvidia’s Investment Spree

As AI labs rush to expand their infrastructure, most are buying GPUs from a single company: Nvidia. This demand has left Nvidia flush with cash, which it has been reinvesting into the industry in unconventional ways. In September 2025, Nvidia acquired a 4% stake in rival Intel for $5 billion. Even more striking were its deals with customers: just a week later, the company announced a $100 billion GPU-for-investment deal with OpenAI, fueling the AI lab’s ongoing data center projects. Similar arrangements have followed with Elon Musk’s xAI, while OpenAI also executed a separate GPU-for-stock deal with AMD.

It’s a circular setup by design. Nvidia’s GPUs remain highly valuable because they’re scarce, and by directly trading them into these massive infrastructure projects, Nvidia ensures that scarcity persists. Likewise, OpenAI’s privately held stock retains its value precisely because it’s unavailable on public markets. For now, both companies are riding high, but any slowdown in momentum could invite intense scrutiny.

Building Tomorrow’s Hyperscale Data Centers

For companies like Meta, which already operate significant legacy infrastructure, the story is more complex—and equally costly. Mark Zuckerberg has announced plans to spend $600 billion on U.S. infrastructure through 2028.

In just the first half of 2025, Meta spent $30 billion more than the previous year, largely driven by AI ambitions. Part of this goes toward cloud contracts, including a recent $10 billion deal with Google Cloud, but the majority is funding massive new data centers.

A 2,250-acre site in Louisiana, dubbed Hyperion, is projected to cost $10 billion and provide 5 gigawatts of compute power. The project includes an arrangement with a local nuclear power plant to handle the energy load. Another smaller site in Ohio, called Prometheus, is expected to come online in 2026, powered by natural gas.

Such buildouts carry environmental consequences. Elon Musk’s xAI built a hybrid data center and power plant in South Memphis, Tennessee, which quickly became one of the county’s largest emitters of smog-producing chemicals, thanks to a series of natural gas turbines that experts say violate the Clean Air Act.

The Stargate Moonshot

Just two days after his second inauguration, President Trump announced a joint venture between SoftBank, OpenAI, and Oracle to spend $500 billion building AI infrastructure in the U.S. Named “Stargate” after the 1994 film, the project came with enormous hype. Trump described it as “the largest AI infrastructure project in history,” and Sam Altman echoed the sentiment, calling it “the most important project of this era.”

The plan envisioned SoftBank providing funding, Oracle overseeing the buildout, and OpenAI offering technical guidance. Trump promised to eliminate regulatory obstacles that could slow progress. However, doubts emerged immediately, including from Elon Musk—Altman’s business rival—who questioned whether the project had the necessary funds to succeed.

Source: TC

Cite this article:

Priyadharshini S (2025), AI Boom Fueled by Billion-Dollar Infrastructure Deals, AnaTechMaz, pp.848