Meta’s $60bn AMD AI Chip Deal: Diversifying Beyond Nvidia
- Meta Platforms has struck a deal with Advanced Micro Devices (AMD) to purchase up to $60 billion worth of artificial intelligence chips over the next five years, a...
- The deal comes as US tech companies are collectively projected to spend $660 billion on AI assets this year, according to analysts.
- “OpenAI had to go multi-vendor because they got to a size where being locked in with just Nvidia limits their growth,” explained Alvin Nguyen, an analyst at Forrester.
Meta Platforms has struck a deal with Advanced Micro Devices (AMD) to purchase up to $60 billion worth of artificial intelligence chips over the next five years, a move signaling the social media giant’s commitment to diversifying its AI supply chain and building out substantial AI infrastructure. The agreement, announced on , also includes Meta acquiring approximately 10% of AMD, mirroring a similar investment AMD made with OpenAI last year.
The deal comes as US tech companies are collectively projected to spend $660 billion on AI assets this year, according to analysts. While Meta has already secured a chip supply agreement with Nvidia, the move to partner with AMD reflects a growing desire among major AI players to mitigate risks associated with relying on a single supplier, particularly given reported supply chain bottlenecks at Nvidia, the current market leader.
“OpenAI had to go multi-vendor because they got to a size where being locked in with just Nvidia limits their growth,” explained Alvin Nguyen, an analyst at Forrester. “Meta are already big enough where they need multiple options.”
AMD will supply Meta with 6 gigawatts (GW) worth of chips, beginning with 1 GW of its forthcoming MI450 hardware in the second half of . The agreement encompasses both central processors (CPUs) and AMD’s flagship graphics processing units (GPUs), with some CPUs being customized specifically for Meta’s needs, designed to maximize performance while minimizing energy consumption, according to AMD CEO Lisa Su.
The investment by Meta is described by Su as a “big bet on AMD,” and underscores a potential shift in Meta’s overall AI strategy. Nguyen suggests the company is pivoting away from directly competing with AI leaders like OpenAI and Anthropic, and instead focusing on becoming a major provider of AI infrastructure. “They seem to have shifted to, ‘We want to host AI infrastructure for you’… Not that they’re not a player, but they’re not the major player [in the AI race],” Nguyen said.
This strategic direction is supported by Meta’s ongoing construction of a massive $27 billion data center campus in Louisiana, a project designed to support its expanding AI capabilities. The AMD deal complements this investment, providing the necessary processing power to fuel the infrastructure.
The timing of this agreement is notable, coming amid broader market concerns about a potential AI bubble and recent volatility in software stocks triggered by the emergence of new “agentic” AI tools. On , Anthropic announced a new plugin for its Claude Cowork tool, designed to integrate seamlessly with popular productivity applications like Google Drive, Gmail, and DocuSign, further intensifying competition in the enterprise AI market.
Despite these concerns, Meta’s commitment to AI remains strong, as evidenced by its substantial investments in talent acquisition over the past year, including attempts to poach top employees from rivals with signing bonuses reaching as high as $100 million. While that initial “talent spending spree” reportedly slowed due to fears of an AI bubble, the AMD deal suggests a renewed focus on building a robust and diversified AI foundation.
While AMD currently holds a smaller market share than Nvidia, analysts believe it presents a viable alternative. Nguyen noted that AMD’s chips perform comparably to Nvidia’s, and even offer advantages in certain areas, while also providing technologies that facilitate the conversion of workloads from Nvidia systems. He anticipates that other AI companies may also diversify their chip suppliers, potentially including Intel in the future.
Meta’s infrastructure head, Santosh Janardhan, emphasized that the company intends to continue sourcing chips from multiple vendors and developing its own in-house processors concurrently. “The scale at which Meta was building datacentres and infrastructure required multiple chip vendors and approaches,” Janardhan stated. “All of the chipmakers end up having sort of a seat at the table.”
The $60 billion agreement with AMD represents a significant financial commitment for Meta, but one that appears strategically aligned with its long-term vision of becoming a leading provider of AI infrastructure, and services. The deal not only secures a critical supply of AI chips but also positions Meta to capitalize on the growing demand for AI-powered solutions across various industries.