Google has reportedly told Meta it can't supply all the Gemini AI capacity the Facebook parent wants, citing a shortage of compute to go around.
According to reporting summarized by MSN, Google informed Meta it would limit access to its Gemini models because there isn't enough capacity available to serve the demand. The move is forcing Meta to rethink how it uses AI and has even constrained its own employees.
Yahoo Finance reports that Google "rationed" Meta's Gemini access amid a compute shortage, and that the restrictions disrupted some of Meta's internal AI projects. In response, the company told staff to use AI tokens more efficiently.
Gizmodo framed the episode bluntly, saying Meta got "too addicted" to Google's AI tokens and had to be cut off.
The squeeze may carry real costs. According to NewsBytes, Meta's plans to release a new AI model now face delays as Google, citing capacity shortages, limits its access to the powerful Gemini models amid soaring global demand.
The situation also caught Wall Street's attention. Barron's noted that Alphabet stock was up, attributing the rise to a combination of Google's entry into the Dow and the AI limits it placed on Meta.
The details across these reports are described as preliminary and based on reporting rather than official confirmation from either company.
Why it matters: The episode shows that even the world's largest tech companies are now bumping against the physical limits of AI computing power — and that controlling that capacity is becoming a strategic lever one rival can pull against another.