It takes a lot of computer science to run an AI product – and as the technological industry is struggling to harness the power of AI models, there is a parallel struggle to build the infrastructure that will supply them. In a recent profit call, NVIDIA CEO Jensen Huang estimates that between 3 trillion and $ 4 trillion will be spent on AI infrastructure by the end of the decade – with many of the money coming from AI companies. Along the road, they place huge pressure on the electrical networks and push the limit of building the industry.
Below, we have created everything we know about the largest AI infrastructure projects, including high expenditure by Meta, Oracle, Microsoft, Google and Openai. We will keep it up to date as the explosion continues and the numbers rise even higher.
Microsoft’s $ 1 billion investment in Openai
This is undoubtedly the agreement that started the entire modern AI explosion: 2019Microsoft made a $ 1 billion investment in a non -profit buzzy named Openai, known mainly for its connection to Elon Musk. Basically, the deal made Microsoft the exclusive cloud provider for Openai – and as the requirements of models became more intense, more than Microsoft investments began to come In the form of Azure Cloud credit instead of cash.
It was very much for both sides: Microsoft was able to claim more Azure sales and Openai got more money for its highest single cost. In the years that followed, Microsoft will build its investment up to nearly $ 14 billion-a move that is going to pay huge when Openai turns into a profit company.
The corporate relationship between the two companies has been unfolding more recently. In January, Openai announced that it would no longer use Microsoft’s cloud exclusively, giving the company the right to refuse to deny future infrastructure requirements, but seeking others if Azure could not meet their needs. More recently, Microsoft has begun to explore other foundation models to supply AI products, creating even greater independence from the AI giant.
The Openai arrangement with Microsoft was so successful that it became a common practice for signing AI services with a specific cloud provider. Anthropic has received investments of 8 billion dollars from Amazon while Make modifications at core level In the company’s material to do it better for AI training. Google Cloud has also signed to smaller AI companies such as Lovable and Windsurf as “prime computers”, although these agreements were not for any investment. And even Openai returned to the well, taking $ 100 billion in investment from Nvidia In Septembergiving it the ability to buy even more of the company’s GPUs.
The rise of Oracle
On June 30, 2025, Oracle was unveiled in a sec that had signed a $ 30 billion cloud agreement with an anonymous partner. This is more than the revenue of the company’s cloud for all the previous use. Openai was finally revealed as a partner, securing Oracle A point next to google As one of the OpenAi strings by Post-Microsoft hosting partners. It was no surprise that the company’s stock went shooting.
TechCrunch event
Francisco
|
27-29 October 2025
A few months later, it happened again. On September 10, Oracle unveiled a five -year $ 300 billion deal on computing power, which will begin in 2027. The Oracle stock went up even higher, making Larry Ellison’s richest man founder. The huge scale of the deal is amazing: Openai does not have $ 300 billion to spend, so the number requires huge growth for both companies and more than a little faith.
But before only one dollar is spent, the deal has already stung Oracle as one of the leading AI infrastructure providers – and an economic power to be taken into account.
Nvidia’s Investment Spree
As Ai Labs it is struggling for the construction of infrastructure, it is mostly everybody buying GPUs from one company: Nvidia. This trade has made Nvidia rinse in cash – and has invested that cash back in the industry in more and more unusual ways. In September 2025, the company bought a share of four percent in the opponent of Intel for $ 5 billion – but even more amazing were the deals with its own customers. A week after the Intel deal was revealed, the company announced A $ 100 billion investment in Openaipaid with GPU to be used in Openai’s current projects. Nvidia has announced a similar agreement with Elon Musk’s XAI and launched Openai A separate GPU layout for storage with AMD.
If this seems circular, it should. Nvidia’s GPUs are valuable because they are so rare-and with their marketing directly in a constantly rinsing data center, Nvidia takes care to remain so. You could say the same thing about Openai’s private stock, which are even more valuable because they cannot be taken through public markets. At present, Openai and Nvidia are driving high and no one seems to be worried too much – but if the momentum begins to signal, this type of adjustment will take much more control.
Create Tomorrow’s Overworked Data Centers
For companies like Meta they already have Important inheritance infrastructureThe story is more complicated – though just as accurate. Mark Zuckerberg said Meta plans to spend $ 600 billion on US infrastructure By the end of 2028.
Just in the first half of 2025, the company spent $ 30 billion more than the previous year, leading to a large extent from the company’s growing ambitions. Some of these costs go to large ticket cloud contracts, such as a recent $ 10 billion agree with Google CloudBut even more resources are poured into two huge new data centers.
A new location of 2,250 acres in Louisiana, Named HyperionIt will cost about $ 10 billion for the construction and supply of about 5 gigawatts computational power. Specifically, the space includes a regulation with a local power plant to handle the increased energy load. A smaller location in Ohio, called Prometheus, is expected to come online in 2026, powered by gas.
This type of construction comes at a real environmental cost. Elon Musk’s XAI created its own hybrid data center and the power plant in South Memphis, Tennessee. The plant quickly became one of the largest broadcasts of clouds producing, thanks to a series of natural gas turbines that Experts say they violate the law on fresh air.
The Stargate Moonshot
Just two days after his second inauguration, President Trump announced a joint venture between Softbank, Openai and Oracle, intended to spend $ 500 billion to build AI infrastructure in the United States. It was named “Stargate” after the 1994 film, the project arrived with incredible amounts of advertising campaign, with Trump calling it “the largest AI infrastructure project in history”. Sam Altman appeared to agree, saying, “I think this will be the most important work of this era.”
In broad strokes, the plan was for Softbank to provide funding, with Oracle handling the buildout with inputs from Openai. Supervising all of this was Trump, who promised to clear up any regulatory obstacles that could slow down the construction. But there were doubts from the beginning, including Elon Musk, Altman’s business opponent, who claimed that the project did not have the available funds.
As the advertising campaign has died, the project has lost some momentum. AugustBloomberg said the partners had failed to reach the consensus. Nevertheless, the project has proceeded with the construction of Eight data centers in Abilene, Texaswith the construction in the final building that will be completed by the end of 2026.
This article was first published on September 22.
