Meta Forced to Scale Back Gemini AI Usage After Google Intervenes
In March, Meta became a major player in the tokenmaxing trend—a practice where employee performance is tied to the volume of AI tokens consumed. But things took a turn when Google told Meta it could no longer handle the load. As a result, Google imposed restrictions on how much Meta could use its Gemini models, and those limits are still in place today.
These restrictions have already stalled or delayed several internal AI projects at Meta. According to the Financial Times, the caps—combined with a broader push to cut AI spending—have led Meta to urge its employees to use AI tokens more sparingly. Tokens are the units that measure AI usage, and sources say this shift shows Meta is trying to adapt to tighter conditions and reduce costs.
The Computing Power Crunch
In early June, Google signed a deal to lease computing capacity from SpaceX for a staggering $920 million per month. This move highlights just how serious the strain on computing systems has become, as Google scrambles to keep its models running smoothly amid skyrocketing demand.
Both Meta and Google have declined to comment on the situation. Still, it's clear that changes in how these AI technologies are being used are reshaping their strategies and growth. While Meta works to adjust to the new reality, Google keeps searching for ways to sustain its services.
This episode underscores a broader challenge facing tech companies today: the soaring demand for computing power needed to run artificial intelligence. To stay competitive, firms must strike a delicate balance between pushing innovation and keeping costs under control. With resources stretched thin, rethinking how AI is used has become essential for long-term stability and growth.
As companies like Meta adapt to new limitations on AI usage, other tech giants are also cutting back on AI expenditures due to escalating token consumption. This trend reflects a broader industry shift as firms navigate rising operational costs while striving to maintain their competitive edge in the rapidly evolving AI landscape.