Meta Secures $100B AMD Chip Deal Amid Massive AI Infrastructure Push

Meta Platforms Inc. has entered into a landmark $100 billion multiyear agreement with Advanced Micro Devices (AMD) to secure AI accelerator chips, marking one of the largest semiconductor supply deals in tech history. The announcement, made on February 24, 2026, comes just days after the company revealed its plan to deploy millions of Nvidia GPUs to fuel its artificial intelligence ambitions. This dual-sourcing strategy signals a seismic shift in the AI hardware landscape, as Meta seeks to diversify its supply chain, reduce dependency on a single vendor, and scale its generative AI models at unprecedented levels.

According to Reuters, this deal represents AMD’s second mega-scale AI chip contract within months, following a previous agreement with a major cloud provider. The partnership with Meta is expected to drive significant revenue growth for AMD’s Data Center division, which has been steadily gaining market share against Nvidia’s dominant position in AI accelerators. While financial terms were not fully disclosed, analysts estimate the agreement spans at least five years and includes custom-designed MI300X and future-generation AI chips optimized for Meta’s internal training clusters and AI inference workloads.

Meta’s decision to partner with AMD follows a broader industry trend of tech giants hedging their bets amid global chip shortages and geopolitical tensions affecting semiconductor supply chains. The company’s AI infrastructure needs are staggering: it plans to train next-generation Llama models with trillions of parameters, requiring tens of thousands of high-performance accelerators. By sourcing from both AMD and Nvidia, Meta gains flexibility in procurement, pricing negotiations, and technological innovation cycles. CNBC reports that Meta will deploy approximately 6 gigawatts of AMD-powered GPU capacity across its global data centers, complementing its existing Nvidia installations.

The $100 billion figure, as reported by MSN, reflects not only the volume of chips but also includes long-term service agreements, co-development of custom silicon, and infrastructure integration support from AMD. This goes beyond a simple vendor-buyer relationship; it is a strategic alliance aimed at accelerating the development of next-generation AI hardware tailored for Meta’s unique workloads, including large-scale language modeling, multimodal AI, and real-time content recommendation systems.

Industry experts view this move as a direct challenge to Nvidia’s near-monopoly in AI chips. While Nvidia controls over 80% of the AI accelerator market, Meta’s decision to allocate half its AI hardware budget to AMD could incentivize other hyperscalers—such as Google, Microsoft, and Amazon—to follow suit. AMD’s ability to offer competitive performance-per-watt and open software ecosystems like ROCm further enhances its appeal to environmentally conscious and cost-sensitive tech giants.

Meta’s Chief Technology Officer, Andrew Bosworth, stated in an internal memo: “Diversification is not just a risk mitigation strategy—it’s a catalyst for innovation.” The company is also reportedly working with AMD on co-designing next-gen chiplets and memory architectures to reduce latency in AI training pipelines. This level of collaboration is rare in the semiconductor industry and underscores the strategic importance Meta places on controlling its AI destiny.

Investors reacted positively to the news, with AMD’s stock rising over 8% in after-hours trading. Meanwhile, Nvidia shares remained stable, suggesting markets view Meta’s diversification as a sign of healthy competition rather than a threat to Nvidia’s dominance. As AI demands continue to explode, the battle for AI chip supremacy is no longer just between vendors—it’s between the tech giants who control the infrastructure that powers the next wave of artificial intelligence.