Google is allegedly limiting the use of its Gemini AI models to Meta after the social media firm’d demand exceeding the models’ computing capacity, according to reports from the Financial Times.

Back in March, Google reportedly told Meta that it did not have the capcity to fullfil the amount of Gemini use the social media company sought to buy. This shortfall, according to reports, lead to delays and disruptions for Meta’s own AI projects.

According to the reports other clients of Google have been affected by the tech titan’s inability to meet demand for its Gemini model compute capacity.

The FT report said that Meta has been particuarly affected by the capacity squeeze due to its high demand on Gemini AI.

The capacity limitations have meant that Meta is now encouraging employees to be more economical when using AI tokens, the report claimed.

Recommended reading

• Hackers Dupe Meta AI Bot To Break Into Intagram Accounts
• Google Cloud to Invest $750M to Scale Enterprise AI for Partners
• Meta to Cut 10% of Workforce

The reports have yet to be verfied by other media outlets – FT credits them to sources close to the matter – but DIGIT has reached out to Google and Meta for comment.

Despite the shift in investment towards AI infrastructure such as data centres and chips which should improve compute capacity of models, the report shows that even top AI firms are still struggling to meet growing demand.