Running AI products requires a lot of computing power. As the technology industry races to harness the power of AI models, there is a parallel race underway to build the infrastructure to power them. On a recent earnings call, Nvidia CEO Jensen Huang estimated that $3 trillion to $4 trillion will be spent on AI infrastructure by the end of 2020, with much of that money coming from AI companies. In the process, they are putting enormous strain on the power grid and pushing the industry’s construction capacity to its limits.
Below, we list everything we know about the largest AI infrastructure projects, including significant spending from Meta, Oracle, Microsoft, Google, and OpenAI. We’ll keep updating it as the boom continues and the numbers rise further.
Microsoft’s $1 billion investment in OpenAI
This is arguably the deal that sparked the entire modern AI boom. In 2019, Microsoft made a $1 billion investment in a high-profile nonprofit called OpenAI, primarily known for its association with Elon Musk. Importantly, the deal makes Microsoft the exclusive cloud provider for OpenAI. As demand for model training grows, many of Microsoft’s investments are now in the form of Azure cloud credits rather than cash.
This was a great deal for both parties. Microsoft was able to capture more Azure sales, and OpenAI was able to get more funding for its largest single expense. Over the next few years, Microsoft would grow its investments to nearly $14 billion. This move will provide huge benefits for OpenAI as it transforms into a commercial company.
The partnership between the two companies was recently terminated. OpenAI announced in January that it would no longer use Microsoft’s cloud exclusively, giving the company a right of first refusal for future infrastructure needs, but would pursue others if Azure could not meet its needs. More recently, Microsoft has started considering other foundational models to power its AI products, giving it more independence from the AI giant.
The agreement between OpenAI and Microsoft has been so successful that it has become common for AI services to sign on to a specific cloud provider. Anthropic has received an $8 billion investment from Amazon while making kernel-level changes to its hardware to make it suitable for AI training. Google Cloud has also signed deals with smaller AI companies like Lovable and Windsurf to be its “principal computing partner,” but those deals did not include any investments. And even OpenAI went back to the well, receiving a $100 billion investment from Nvidia in September, giving it the ability to buy more of the company’s GPUs.
Rise of the Oracle
On June 30, 2025, Oracle disclosed in an SEC filing that it had entered into a $30 billion cloud services agreement with an unnamed partner. This is more than the company’s cloud revenue for the entire previous fiscal year. OpenAI was eventually revealed as a partner, securing Oracle a spot alongside Google as one of OpenAI’s series of post-Microsoft hosting partners. Not surprisingly, the company’s stock price soared.
tech crunch event
san francisco
|
October 27-29, 2025
A few months later, the same thing happened again. On September 10, Oracle unveiled a five-year, $300 billion deal for computing power starting in 2027. Oracle’s stock price rose further, and founder Larry Ellison briefly became the world’s richest man. The scale of the contract is surprising. Since OpenAI doesn’t need to spend $300 billion, this number estimates tremendous growth and considerable trust for both companies.
But before a single dollar is spent, this deal has already firmly established Oracle as one of the leading AI infrastructure providers, and it has economic power to be reckoned with.
Nvidia’s huge investment
As AI labs rush to build out their infrastructure, nearly every company is buying GPUs from one company: Nvidia. The deal leaves Nvidia flush with cash, which it is investing in the industry in increasingly unconventional ways. In September 2025, the company bought a 4% stake in rival Intel for $5 billion, but even more surprising was the deal it made with its own customers. A week after the deal with Intel was revealed, the company announced a $100 billion investment in OpenAI, paid for by GPUs used in OpenAI’s ongoing data center projects. Nvidia then announced a similar deal with Elon Musk’s xAI, and OpenAI started a separate stock GPU deal with AMD.
If it looks like it’s cyclical, it should be. Nvidia’s GPUs are valuable because they’re so rare, and trading them directly into ever-expanding data center plans allows Nvidia to keep it that way. The same is true for OpenAI’s private equity, but it’s not available on the public market, making it even more valuable. For now, OpenAI and Nvidia are doing great and no one seems too worried, but if momentum starts to wane, this type of deal will come under even more scrutiny.
Build tomorrow’s hyperscale data center
For companies like Meta that already have significant legacy infrastructure, the story is more complicated, albeit just as expensive. Mark Zuckerberg said Meta plans to spend $600 billion on U.S. infrastructure by the end of 2028.
In the first half of 2025 alone, the company spent $30 billion more than the previous year. This is primarily driven by the company’s expanding AI ambitions. Some of that spending is going toward big-ticket cloud deals, such as the recent $10 billion deal with Google Cloud, but even more resources are being poured into two massive new data centers.
Construction of the new 2,250-acre site in Louisiana, called Hyperion, will cost an estimated $10 billion and provide an estimated 5 gigawatts of computing power. Notably, the site includes an arrangement with a local nuclear power plant to handle the growing energy load. A smaller site in Ohio called Prometheus is scheduled to be operational in 2026, powered by natural gas.
This type of augmentation has real environmental costs. Elon Musk’s xAI has built its own hybrid data center and power plant in South Memphis, Tennessee. The plant quickly became one of the county’s largest emitters of smog-causing chemicals, thanks to a series of natural gas turbines that experts say violate the Clean Air Act.
stargate moonshot
Just two days into his second term in office, President Trump announced a $500 billion joint venture between SoftBank, OpenAI, and Oracle to build AI infrastructure in the United States. The project, named “Stargate” after the 1994 film, was announced with incredible hype, with President Trump calling it “the largest AI infrastructure project in history.” Sam Altman seemed to agree, saying, “I think this will be the most important project of our time.”
Broadly speaking, the plan was for SoftBank to provide the funding and Oracle to build it with input from OpenAI. President Trump oversaw everything, promising to remove any regulatory hurdles that might slow construction. But there were doubts from the beginning, with some, including Altman’s business rival Elon Musk, claiming there was no available funding for the project.
The project lost some momentum as the hype died down. In August, Bloomberg reported that the partners had not reached an agreement. Nevertheless, the project is moving forward with construction of eight data centers in Abilene, Texas, with construction on the last building expected to be completed by the end of 2026.
This article was first published on September 22nd.
