Meta has secured a multibillion-dollar agreement to rent hundreds of thousands of Amazon's homegrown CPUs for artificial intelligence workloads. The deal marks a significant shift in the chip market as major tech companies diversify beyond GPUs.
Meta Platforms and Amazon Web Services have finalized a major partnership that will see the social media giant deploy millions of Amazon's custom-built processors for AI infrastructure. The chips in question are general-purpose CPUs rather than the GPUs traditionally associated with AI training and inference.
The arrangement signals a strategic pivot in how large technology companies approach chip procurement. Rather than relying exclusively on Nvidia's dominant GPU offerings or developing entirely proprietary solutions, Meta is leveraging Amazon's in-house chip capabilities to scale its AI operations.
Amazon has invested heavily in developing its own silicon, including Trainium chips for training and Inferentia chips for inference workloads. This deal demonstrates that these custom processors have matured enough to attract major customers seeking alternatives to established options.
The CPU-focused approach suggests Meta is expanding its AI infrastructure beyond traditional deep learning tasks. The use of general-purpose processors indicates potential deployment for agentic AI systems—autonomous AI agents that perform complex reasoning and task execution.
This partnership reflects broader industry trends. As AI capabilities become central to business operations, companies are moving beyond single-supplier relationships to diversify their computing resources. The strategy reduces dependency risk and creates competitive pressure in the chip market.
The exact financial terms of the multibillion-dollar deal were not disclosed, but the scale indicates a substantial commitment from Meta to Amazon's chip ecosystem. For Amazon, the arrangement provides validation for its homegrown silicon and opens pathways to similar partnerships with other major technology firms.
The move underscores that the AI chip landscape has evolved beyond the GPU-dominated paradigm. As workloads diversify and companies build more sophisticated AI systems, demand for specialized processors—including CPUs designed for specific AI applications—continues to grow.
A Microsoft-backed AI data center is under scrutiny for allegedly operating dozens of large generators without proper federal permits. The violation marks growing regulatory pressure on tech companies scaling infrastructure for AI workloads.
SoftBank Group Corp. is pursuing a $10 billion loan to refinance debt tied to its investment in OpenAI. The funding move comes as the Japanese conglomerate manages its exposure to the AI sector.
X has launched NFL Gametime, an integrated platform combining live scores, statistics, and creator content for NFL fans. The partnership runs through the 2027 Super Bowl.
A federal judge ruled Thursday that the Trump administration's February designation of AI company Anthropic as a national security supply-chain risk was illegal and baseless. The decision blocks sanctions that could have cost the Claude maker billions in lost business.