News
Meta's New AWS Deal Is a Bet on Millions of Custom AI Chips
Meta Platforms is scaling aggressively into AI infrastructure, committing to tens of millions of AWS Graviton processors to power next-generation agentic AI.
The deal reflects a shift in AI computing, where CPUs are becoming critical alongside GPUs for real-time decision-making and running AI systems at scale. Meta is diversifying its compute strategy and locking in supply as demand for AI chips surges, signaling an intensifying race for infrastructure dominance.
Meta announced it will use "tens of millions" of AWS Graviton chips to advance its AI efforts, making it the most high-profile entity to deploy AWS's custom chips at scale.
The deal, announced on Friday, propels the Mark Zuckerberg-led company to become one of the largest Graviton customers, AWS says.
"AWS has been a trusted cloud partner for years, and expanding to Graviton allows us to run the CPU-intensive workloads behind agentic AI with the performance and efficiency we need at our scale," said Santosh Janardhan, Meta’s head of infrastructure, in a statement.
Meta's investment signals its push toward the next wave of AI, particularly agentic AI. Demand for AI chips is outpacing supply, and companies like Meta are locking in long-term deals to secure compute.
GPUs are still used to train AI models, but CPUs like Graviton are critical for real-time decision-making, orchestrating tasks, and running AI systems at scale, Meta explained.
Graviton was AWS’s first custom cloud processor and is used by companies such as Uber, Pinterest, Airbnb, and Formula 1.
AWS boasts that its in-house Graviton5 chips offer "faster data processing and greater bandwidth critical for AI systems that need to reason through and execute tasks at scale continuously."
This is in line with Meta's current needs and comes on the back of a $10 billion deal the Menlo Park-based company signed with Google Cloud in August 2025.
"As we scale the infrastructure behind Meta's AI ambitions, diversifying our compute sources is a strategic imperative," Janardhan said.