Quick Overview
- Meta has finalized a major multibillion-dollar agreement with Amazon spanning three to five years for AWS Graviton CPU technology.
- The partnership involves deploying tens of millions of Graviton processing cores, predominantly across U.S. data centers.
- Following the announcement, Meta shares climbed 0.6% while Amazon stock jumped 1.4% during premarket hours.
- Amazon joins Meta’s diverse semiconductor ecosystem alongside Nvidia, Broadcom, and AMD.
- Meta now ranks among the top five largest AWS Graviton clients globally.
Meta Platforms has finalized a massive multibillion-dollar partnership with Amazon Web Services that will provide the social networking giant access to millions of AWS Graviton central processing units specifically designed to support its expanding AI agent infrastructure.
According to Nafea Bshara, an Amazon vice president who co-founded Annapurna Labs—AWS’s proprietary semiconductor division—the arrangement spans a three-to-five-year period. The majority of these Graviton processors will be situated within American facilities.
Market reaction was swift: Amazon shares surged 1.4% in early trading, while Meta’s stock advanced 0.6%.
This partnership centers exclusively on AWS Graviton central processing units rather than graphics processing units. While GPUs dominated AI development conversations, CPUs are experiencing renewed relevance.
The emergence of AI agent technology has repositioned CPUs as critical infrastructure components. These processors manage discrete operations and channel workloads to GPUs, creating a synergistic relationship between both chip categories for diverse AI applications.
Additionally, CPUs serve vital functions during the post-training stage of large language model development, when foundational models undergo refinement for specialized applications.
Meta selected Amazon’s Graviton5—a cutting-edge 3-nanometer processor—based on its exceptional cost-effectiveness. “Meta enjoys extensive supplier options. Yet they selected Graviton5,” Bshara emphasized.
Building a Multi-Vendor Semiconductor Ecosystem
This Amazon partnership expands Meta’s existing chip procurement strategy that encompasses Nvidia, Broadcom, AMD, and Arm Holdings. Meta has deliberately pursued a diversified approach to semiconductor sourcing.
“No single chip architecture can efficiently serve every computational task,” Meta stated in its official announcement.
Initial rollout will encompass tens of millions of Graviton cores, with flexibility to expand capacity as Meta’s artificial intelligence requirements evolve.
Funding AI Expansion Through Workforce Reduction
Meta’s artificial intelligence strategy continues accelerating. The company completed its acquisition of AI startup Manus for over $2 billion last December. Manus specializes in developing AI agents capable of executing sophisticated operations, intensifying CPU requirements.
To finance its aggressive AI infrastructure expansion, Meta simultaneously announced Thursday that approximately 10% of its workforce—roughly 8,000 positions—will be eliminated in May.
Earlier this month, Meta unveiled Muse Spark, its first new AI model in twelve months, with additional releases planned for upcoming quarters.
While Meta and AWS have maintained a business relationship since approximately 2016, previous collaboration primarily involved standard cloud services, the Bedrock platform, and GPU cluster access. This agreement represents a substantial evolution toward customized semiconductor solutions.
For Amazon Web Services, securing Meta as a significant Graviton adopter validates its custom silicon strategy. Just days earlier, Amazon revealed a $5 billion commitment to Anthropic, another partnership involving substantial Graviton CPU deployment.
AWS initiated proprietary chip development before 2018, launching its inaugural Graviton processor on Arm architecture.
Bshara verified that this agreement positions Meta within AWS’s top five Graviton customers worldwide.
