Mexico City, January 31, 2026, 07:11 (UTC-6)
Nvidia’s effort to invest as much as $100 billion in OpenAI has hit a roadblock, The Wall Street Journal reported Friday, citing doubts from within the chipmaker. “We have been OpenAI’s preferred partner for the last 10 years. We look forward to continuing to work together,” an Nvidia spokesperson said. OpenAI did not immediately respond to requests for comment. Reuters
The pause comes as OpenAI seeks new funding to cover the costs of its data centers — massive server warehouses that power AI training and operations — while big tech companies jockey to steer those workloads their way. Reuters reported Thursday that Amazon is in early talks to invest up to $50 billion, with CEO Andy Jassy negotiating directly with Sam Altman.
The Information reported separately that Nvidia, Amazon, and Microsoft have been in talks about investments that could reach $60 billion. Nvidia might contribute as much as $30 billion, while Microsoft’s stake is expected to be under $10 billion, according to Reuters. The same report linked Amazon’s potential involvement to separate negotiations over cloud server rentals and selling OpenAI products, including enterprise ChatGPT subscriptions.
The Journal reports that Nvidia and OpenAI are reconsidering their partnership, discussing a reduced equity investment—potentially in the tens of billions of dollars. Nvidia CEO Jensen Huang has privately called the initial agreement non-binding and expressed worries about competition from Google, Anthropic, and others.
Bloomberg News, referencing the Journal, reported the talks have collapsed, highlighting how fast moods shift in AI megadeals. OpenAI is behind ChatGPT, and Nvidia leads the market for processors powering these systems.
Amazon is considering a wider deal to integrate OpenAI’s models into its products and platforms, giving employees access for work purposes, Bloomberg reported, citing an insider. This could build on their current agreement, where Amazon provides computing power to OpenAI, the report added.
Back in September, Nvidia and OpenAI revealed they’d signed a letter of intent — a non-binding framework — to roll out at least 10 gigawatts of “compute,” industry shorthand for the chips and servers driving AI. Gil Luria at D.A. Davidson cautioned that Nvidia might end up as an “investor of last resort.” Meanwhile, Kim Forrest from Bokeh Capital remarked, “This sounds like Nvidia is investing in its largest customer.” Reuters
OpenAI has been on the hunt for extra compute power, reportedly striking a deal worth more than $10 billion over three years with Cerebras, a competitor to Nvidia, for up to 750 megawatts of capacity, according to a Reuters report from mid-January. “Integrating Cerebras into our mix of compute solutions is all about making our AI respond much faster,” OpenAI said, adding the capacity is focused on inference — running trained models to answer questions. Reuters
Key terms remain in flux, and Reuters has yet to verify the Journal’s report. Should Nvidia’s involvement decrease, OpenAI will probably rely more heavily on a mix of chip suppliers and cloud providers to meet capacity demands for its upcoming models.
