[Image: TechCrunch article header showing an AI-related image.]
Running an AI product demands substantial computing power, and as the tech industry rushes to harness AI models, there’s a concurrent push to develop the infrastructure required by them. Nvidia CEO Jensen Huang predicted on a recent earnings call that AI infrastructure spending could reach $3-$4 trillion by the decade’s end, primarily funded by AI companies. This growth is taxing power grids and the industry’s building capacity.
Major AI infrastructure projects are underway, with significant spending from Meta, Oracle, Microsoft, Google, and OpenAI. We will update developments as the sector expands.
**Microsoft’s 2019 Investment in OpenAI**
Microsoft’s $1 billion investment in OpenAI in 2019 significantly propelled the current AI boom. This partnership established Microsoft as OpenAI’s exclusive cloud provider. As model training demands increased, Microsoft shifted some investments to Azure cloud credits. The deal benefited both companies, boosting Azure sales for Microsoft and providing OpenAI with essential resources. Microsoft’s investment has since grown to nearly $14 billion, anticipating future gains as OpenAI transitions to a for-profit company.
Though the partnership has changed, with OpenAI opting to explore other cloud services, the practice of AI companies aligning with specific cloud providers is now commonplace. Competitors like Anthropic, funded by Amazon, and smaller AI companies are partnering with Google Cloud, while OpenAI secured a $100 billion investment from Nvidia for further GPU acquisition.
**The Rise of Oracle**
Oracle’s prominence increased with a disclosed $30 billion cloud deal on June 30, 2025, with a partner later identified as OpenAI, propelling Oracle to a key position among OpenAI’s post-Microsoft hosts. A further $300 billion, five-year contract in September 2025 with OpenAI highlighted Oracle’s growing influence, momentarily making Larry Ellison the wealthiest global individual. This predicted growth signals Oracle’s emergence as a leader in AI infrastructure.
**Nvidia’s Investment Strategy**
Nvidia’s GPUs are central to AI labs building infrastructure, allowing the company to reinvest in innovative ways. Nvidia acquired a 4% stake in Intel for $5 billion in September 2025, continuing with a $100 billion deal with OpenAI using GPUs for data center projects. Subsequently, Nvidia has formed further partnerships, including Elon Musk’s xAI and a GPU-for-stock arrangement with AMD. This strategy ensures Nvidia retains the value of its scarce GPUs and aligns its future with OpenAI’s potential gains.
**Building Tomorrow’s Hyperscale Data Centers**
For companies like Meta with existing infrastructure, expansion is costly. Meta CEO Mark Zuckerberg announced a $600 billion U.S. investment through 2028, with $30 billion spent in the first half of 2025 alone to drive AI goals, including a $10 billion deal with Google Cloud. Meta’s investments also involve significant new data centers, including Hyperion in Louisiana and Prometheus in Ohio, utilizing local energy resources.
This expansion, however, raises environmental concerns, as evidenced by xAI’s data center in Memphis impacting local air quality due to natural gas turbines, contrary to the Clean Air Act.
**The Stargate Moonshot**
President Trump announced “Stargate,” a $500 billion U.S. AI infrastructure project with SoftBank, OpenAI, and Oracle, framing it as historically significant. Skepticism has slowed momentum, but progress includes eight data centers in Texas set for 2026.
**The Capex Crunch**
Tech companies projected significant capital expenditures in 2026, led by Amazon at $200 billion and Google at $175-$185 billion. Meta’s spending estimate, partially off-books, suggests a heavy focus on data centers in 2026. The reported investments raise investor concerns, compounded by debt financing these expansions. The commitment to AI infrastructure indicates its perceived criticality to future growth, despite market apprehension.
*Originally published on September 22.*