OKX app offers multilingual support, copy trading, and API solutions for global traders. New users enjoy exclusive fee discounts!
Download for Android Download for IOS
OpenAI has significantly expanded its Stargate AI infrastructure project through a new partnership with Oracle, adding 4.5 gigawatts (GW) of computing capacity. This development pushes the initiative's total pipeline beyond 5 GW — enough energy to power over two million advanced AI chips. The expansion builds upon the existing Stargate I facility in Abilene, Texas, with OpenAI CEO Sam Altman calling it "a gigantic infrastructure project" in social media posts.
——This move comes as Elon Musk's xAI announced plans to deploy computing power equivalent to 50 million Nvidia H100 units within five years——. Industry analysts note this would represent 【500 times】 the capacity of what was considered cutting-edge supercomputing just twelve months prior. Musk's firm is preparing to launch its Colossus 2 supercomputer, utilizing 550,000 next-generation GB200 chips.
Originally announced as a $500 billion initiative under the Trump administration, the Stargate project has faced implementation hurdles despite its ambitious job creation targets. Recent reports indicate scaled-back near-term objectives, with partners now focusing on completing a single data center by year-end. Interestingly, the project's energy requirements now exceed those of several small nations, raising questions about sustainable AI development.
The parallel developments highlight the intensifying competition in AI infrastructure. OpenAI's expansion with Oracle demonstrates the growing importance of energy partnerships in large-scale AI deployment. Meanwhile, xAI's bold compute targets suggest alternative approaches to achieving similar capabilities through hardware innovation rather than sheer scale.
Market observers point to the contrasting strategies — while OpenAI expands through established partnerships, xAI appears focused on vertical integration. Both approaches carry significant risks and potential rewards in the race for AI supremacy. As of press time, neither company has disclosed detailed timelines for achieving their respective capacity goals.