Sam Altman’s OpenAI is spearheading a transformative project dubbed Stargate, a $500 billion initiative intended to cement the United States as a global leader in artificial intelligence. Backed by major tech players including Oracle and Microsoft, the effort involves building expansive AI infrastructure and is expected to create over 100,000 American jobs immediately, according to a report from Forbes.
The centerpiece of the project will be a flagship data center in Abilene, Texas, which will house up to 100,000 Nvidia GPUs by the end of 2025. The center will be powered largely by renewable energy from hundreds of towering wind turbines across West Texas.
Crusoe Energy brings sustainable power to AI computing
To meet the enormous energy needs of Stargate, OpenAI has partnered with Crusoe Energy, a company known for converting waste gas from oil wells into electricity.
Founded by Chase Lochmiller and Cully Cavness, Crusoe’s technology offers a cleaner, more sustainable energy source for power-hungry data centers.
Rather than flaring off natural gas—a common and pollutive practice—Crusoe uses it to run engines that generate electricity. Previously, this setup was used to mine Bitcoin. Now, that same infrastructure will support advanced AI computing.
“It will be a ‘factory of factories,’” Lochmiller said, referring to the potential of AI systems to design and develop even more powerful AI technologies in the future.
$100 billion already deployed as construction begins
Though the full $500 billion has not yet been allocated, the initial $100 billion is already being spent to construct two data centers. These will be central to training and deploying OpenAI’s next-generation models.
The hardware and computing systems will be built through a joint effort between OpenAI, Oracle, and Nvidia. SoftBank, Oracle, MGX, and OpenAI are among the first round of investors. Additional technology collaborators include Microsoft, Arm, and Nvidia.
Stargate seen as strategic response to Chinese AI surge
Stargate’s strategic importance is amplified by the rapid ascent of China’s AI sector, particularly the emergence of DeepSeek—a competitor offering AI models at lower costs and with simplified integration.
OpenAI has responded swiftly to this rising competition, recently launching GPT-4o, which includes image-generation capabilities.
Just this week, OpenAI CEO Sam Altman announced the release of GPT-4.1—including new mini and nano versions—through the API.
These models, he said, are optimized for long context understanding (up to 1 million tokens), as well as instruction following and advanced coding tasks.