Google restricts Meta’s Gemini AI access amid computing crunch – NaturalNews.com
The battle for dominance in artificial intelligence has taken an unexpected turn, as Google has reportedly imposed restrictions on Meta's access to its Gemini AI models. According to a Financial Times report, Google notified Meta in March that it could not fulfill the social media giant's full request for computing capacity, disrupting a number of Meta's internal AI projects. The move has raised fresh questions about the fragility of AI infrastructure at a moment when tech giants are racing to control the future of machine intelligence.
The capacity crunchGoogle's decision to cap Meta's Gemini access spotlights a systemic problem across the AI sector: a deepening shortage of computing power. Despite billions poured into chips and data centers, major players are struggling to keep pace with surging demand.
Google CEO Sundar Pichai acknowledged the problem during a recent earnings call. "Obviously, we are compute-constrained in the near term," he said, adding that Google Cloud revenue "would have been higher if we were able to meet the demand." The cloud unit's backlog of signed but undelivered contracts nearly doubled quarter over quarter to more than $460 billion — a figure that illustrates just how far supply has fallen behind demand.
Ripple effects across the industryThe limitations are not exclusive to Meta. Other Google clients have also faced reduced Gemini access, albeit to a lesser extent, according to the Financial Times. Meta bore the brunt due to its exceptionally high appetite for Google's models. To shore up its own capacity, Google earlier this month signed a deal worth $920 million per month to lease computing power from SpaceX, a move that underscores how even the best-resourced companies in tech are scrambling for infrastructure. AI lab Anthropic, maker of the Claude chatbot, has struck a similar arrangement with SpaceX.
Meta's response and strategic shiftsFaced with the restrictions, Meta has directed employees to use AI tokens — the units that measure AI processing — more efficiently. The directive is a notable one for a company whose CEO Mark Zuckerberg has spoken openly about pursuing what he calls "personal superintelligence."
Internally, Meta had been relying on Gemini to help automate safety work such as identifying scams and removing harmful content, as well as for customer service and advertising tools and some coding workflows. More recently, however, Meta has begun shifting toward its own Muse Spark model, which is seen as more competitive with Gemini and would reduce the company's dependence on a rival's platform. Meta has also committed to investing $600 billion in U.S. infrastructure by 2028 as part of a longer-term push to build out its own data center capacity.
AI explosion is unsustainableThe Google-Meta situation is a revealing snapshot of where the AI industry actually stands beneath the hype. Companies are spending at an extraordinary scale and still cannot keep up. For smaller businesses and startups that depend on access to third-party AI models, the implications are sobering; capacity rationing by the biggest players could leave them with even less.
The Financial Times report notes that Meta initially turned to Gemini because it outperformed the company's own open-source Llama models, a reminder that even a company of Meta's size and resources has had to rely on competitors to fill gaps. The push toward proprietary models like Muse Spark may reduce that vulnerability over time, but it also signals a broader trend: the AI industry is consolidating around a handful of vertically integrated giants who control both the models and the infrastructure to run them. Whether that concentration serves the public interest or mainly the interests of shareholders is a question worth asking.
Sources for this article include: