Oracle to Spend $40 Billion on Nvidia Chips for OpenAI’s Massive Texas Data Center

Oracle is set to invest $40 billion in Nvidia’s latest AI chips to power a massive new data centre in Abilene, Texas, as part of the Stargate initiative led by OpenAI and SoftBank. The facility will deliver 1.2 gigawatts of computing power, making it one of the largest AI infrastructure sites globally. Oracle plans to lease the site for 15 years and provide computing services to OpenAI.This marks a major shift for OpenAI, which is moving away from exclusive reliance on Microsoft Azure due to rising compute demands. The Stargate project aims to raise up to $500 billion over the next four years to build AI superclusters worldwide. OpenAI and SoftBank have each pledged $18 billion, with Oracle and Abu Dhabi’s MGX contributing $7 billion apiece.Construction in Texas began in 2024, backed by $15 billion in financing from JPMorgan, Crusoe Energy, and Blue Owl Capital. OpenAI also announced an international expansion with a 5GW data centre in Abu Dhabi.Oracle’s move not only secures its position in the AI cloud market but also strengthens Nvidia’s dominance in AI hardware, signalling the start of a new global AI infrastructure race.

Oracle is making a bold $40 billion investment in Nvidia’s most advanced AI chips to power a next-generation data center in Texas. The project, part of the Stargate initiative led by OpenAI and SoftBank, represents one of the largest infrastructure efforts ever seen in the field of artificial intelligence.

With global tech giants racing to secure processing power for AI training, Oracle’s bet on Nvidia chips signals a new era of AI-driven cloud infrastructure. This move could reshape the power dynamics in big tech.

A Record-Breaking AI Infrastructure Project in Texas

Oracle’s massive chip purchase will support a 1.2 gigawatt data center in Abilene, Texas, making it one of the largest AI computing facilities in the world. The site is part of Project Stargate, a $500 billion vision to transform the U.S. into a dominant force in artificial intelligence infrastructure.

Once completed in mid-2026, the data center will house 400,000 Nvidia GB200 “superchips”, purpose-built for large-scale AI workloads. Oracle will lease the facility for 15 years, offering computing services to OpenAI and other clients.

The scale of the site rivals similar ambitions from Elon Musk’s “Colossus” data center in Memphis and Amazon’s upcoming Virginia facility, both aiming to break the 1GW power barrier.

Oracle Bets Big on Nvidia’s Next-Gen GB200 Chips

The Nvidia GB200 chip is a cutting-edge processor designed specifically for AI model training and inference. It is significantly more powerful than its predecessors, such as the H100 and H200, which have been the workhorses of many existing data centers.

Oracle’s purchase of 400,000 GB200 chips suggests that the company is not just scaling but also future-proofing its AI infrastructure capabilities. These chips will fuel OpenAI’s demand for computational power as it continues to train and operate increasingly complex models like GPT-5 and beyond.

The Stargate Project: A $500 Billion Bet on AI

Project Stargate is a long-term, multi-national plan to invest up to $500 billion over the next four years. The initiative was first unveiled in January 2025 by U.S. President Donald Trump and is backed by a coalition of tech and investment powerhouses:

• OpenAI and SoftBank have each committed $18 billion
• Oracle and Abu Dhabi’s MGX sovereign wealth fund have pledged $7 billion each

While Stargate itself has not committed capital to any single site yet, the Texas facility is widely seen as its flagship project. Once operational, Stargate will help reduce OpenAI’s reliance on Microsoft Azure and create a more diversified AI infrastructure ecosystem.

Breaking Ties: OpenAI Moves Away from Microsoft

For years, OpenAI depended almost entirely on Microsoft for its compute needs, supported by nearly $14 billion in Azure cloud credits. But as OpenAI’s models became more sophisticated and power-hungry, Microsoft’s infrastructure could no longer keep up.

In early 2025, OpenAI and Microsoft ended their exclusivity agreement, opening the door for new infrastructure partners. This shift marks a strategic turning point in OpenAI’s growth and sets the stage for independent operations at an unprecedented scale.

The Texas data center, backed by Oracle, is the clearest sign yet that OpenAI is building its own compute empire.

Financing the Future: $15 Billion Raised for Construction

Building a 1.2GW data center is a massive undertaking. The site’s co-owners, Crusoe Energy and Blue Owl Capital, have already raised $15 billion in debt and equity to finance the Abilene project.

The financing includes:

• $9.6 billion in loans from JPMorgan, with $7.1 billion announced this week
• $5 billion in direct equity investments from Crusoe and Blue Owl

This capital will fund the eight-building campus in Texas. Construction began in June 2024, and work is well underway.

A New AI Arms Race: Global Superclusters in the Making

While the Abilene facility is historic in scale, it is not alone. Across the globe, tech titans are scrambling to build “AI superclusters.”

• Elon Musk plans to upgrade his Colossus data center in Memphis to include 1 million Nvidia chips, becoming what he calls the “first gigawatt AI training supercluster”
• Amazon is building a data center in northern Virginia that will surpass 1GW in power
• OpenAI is working on an international Stargate site in Abu Dhabi, covering 10 square miles with a power capacity of 5GW. This is enough to support 2 million Nvidia chips

The Abilene site is the first chapter in a global story of AI infrastructure expansion. The companies involved are investing at a scale previously only seen in national energy projects.

Why Oracle’s Move Matters

Oracle’s $40 billion commitment is more than just a tech purchase. It is a strategic declaration.

• It positions Oracle as a major cloud AI infrastructure provider alongside Amazon AWS and Microsoft Azure
• It cements Nvidia’s dominance in the AI chip space while competitors like AMD and Intel struggle to keep pace
• It gives OpenAI the freedom to scale beyond Microsoft, paving the way for more innovation and independence

The long-term lease of the Texas data center also allows Oracle to generate predictable cloud revenue from leasing compute power, rather than building or operating everything in-house.

Stargate Expansion: Abu Dhabi Enters the Game

As part of Stargate’s global ambitions, a new mega data center is coming to Abu Dhabi. This UAE-US AI campus will:

• Cover 10 square miles
• Deliver 5GW of compute power
• Support more than 2 million GB200 chips
• Be built by G42, an Emirati AI company

The campus was announced during President Trump’s Gulf tour in May 2025. It demonstrates the growing geopolitical significance of AI infrastructure. The U.S., UAE, and private tech leaders are now partnering in a new kind of technology diplomacy based on silicon, power, and algorithms.

Final Thoughts: The Future of AI Is Being Built Right Now

The Abilene data center is more than a construction site. It is a symbol of where artificial intelligence is headed. With Oracle’s $40 billion chip order, OpenAI’s growing independence, and Stargate’s multi-national funding, we are witnessing the industrialization of artificial intelligence.

Just as railroads and power grids transformed economies in previous centuries, AI superclusters are laying the foundation for the digital economies of the future.