NVIDIAAnnounceWill invest up to $1000 billion to help OpenAI buildAI data center with a capacity of at least 10GW, and fully adopts NVIDIA GPU acceleration system architecture. The scale of this cooperation will provide computing power support for OpenAI's next generation of large-scale language models.
According to the plan, NVIDIA's investment will be gradually released as each 1GW of capacity comes online. The first batch of data centers is expected to be put into operation in the second half of 2026 and will be based onNVIDIA's next-generation Vera Rubin architecture platformNVIDIA CEO Jensen Huang said the Vera Rubin architecture platform will be "another big step forward" compared to the existing Blackwell chip architecture, with a more significant leap in performance and efficiency.
In his statement, Jensen Huang reviewed the decade-long collaboration between NVIDIA and OpenAI, from the initial DGX supercomputer to the creation of ChatGPT, with both parties continuously pushing the boundaries of technology. OpenAI CEO Sam Altman stated that computing infrastructure will be central to the future economy. Leveraging NVIDIA's contributions, OpenAI will be able to continuously drive breakthroughs in AI technology and empower businesses and users at scale.
This is not the only big investment made by NVIDIA recently.Invest $50 billion in Intel, followed by the U.S. governmentAcquired a 10% stake in IntelThe partnership was further strengthened. NVIDIA also invested over $9 million to license technology from AI startup Enfabrica and recruit its core team.
For OpenAI, this investment echoes previous collaborations, such as OpenAI and Oracle.Reaching an Agreement, plans to build a 4.5GW data center capacity, using more than 200 million Oracle chips asStargate ProjectThe project, jointly initiated by SoftBank, OpenAI, NVIDIA, Oracle, Arm, and Microsoft, aims to invest $5000 billion in building AI infrastructure in the United States.
This move not only demonstrates the deep integration of NVIDIA and OpenAI within the AI ecosystem, but also signals a new round of computing power competition in the global AI training market. Upon completion of the 10GW AI data center, OpenAI will boast unprecedented model training scale, placing significant pressure on competitors such as Google DeepMind, Anthropic, and xAI.
