Google has placed restrictions on Meta's use of its Gemini AI models after Meta requested more processing power than the competing tech group could offer, according to sources who spoke on condition of anonymity. The move highlights ongoing capacity constraints in the AI industry despite massive investments in chips and data centers.
Capacity Shortfall Disrupts Meta's AI Initiatives
Google, owned by Alphabet, informed Meta in March that it could not provide the full Gemini capacity Meta sought. The shortfall disrupted and delayed several of Meta's internal AI projects. The source indicated that other Google clients have also been affected, though to a lesser degree, with Meta hit hardest due to its extraordinarily high demand for Google's models.
Meta has responded by urging employees to use AI tokens—units that measure AI usage—more efficiently to cope with limitations.
Industry-Wide Processing Power Strain
Corporations continue to invest billions in chips and data centers but still struggle to secure enough processing capacity for surging AI demand. Despite Google Cloud's first-quarter revenue reaching USD 20 billion, CEO Sundar Pichai said processing power constraints limited further growth and contributed to the cloud unit's backlog nearly doubling quarter over quarter.
According to Pichai, the capacity crunch is a significant bottleneck for scaling AI services. The situation underscores the intense competition for computational resources among tech giants racing to deploy advanced AI models.
Impact on Cloud Revenue and Backlog
Google Cloud's strong revenue performance was tempered by the inability to meet all customer demand. The backlog, which represents committed but unfulfilled contracts, surged as customers sought more capacity than Google could deliver. This trend is expected to persist as AI adoption accelerates across industries.
Analysts note that capacity constraints could reshape cloud market dynamics, with customers potentially diversifying providers or investing in on-premises infrastructure. Google and its rivals are racing to expand data center capacity, but supply chain issues and energy demands remain hurdles.
