Meta secures massive AMD chip supply to scale AI infrastructure

• Meta’s AMD chip deal shows hyperscalers securing long-term AI compute.
• Goal of supporting AI demand.

The race to secure computing power for artificial intelligence is moving to long-term supply commitments. Meta’s new agreement with AMD shows how large tech firms are trying to lock-in the hardware needed to run future AI systems.

Under the deal, AMD will supply Meta with large volumes of AI chips over five years, in an arrangement that could be worth around US$60 billion, according to Reuters. The chips are expected to support Meta’s growing data centre capacity, with deployments beginning in the second half of 2026.

The agreement gives Meta access to AMD’s next-generation AI hardware, including systems designed for inference workloads, the stage where trained models generate responses for users.

Securing compute for AI

Meta has been expanding its AI infrastructure to support features in its social platforms and research efforts. Such workloads need large-scale compute resources, pushing hyperscalers to secure long-term chip supply.

The Reuters report notes the AMD agreement could provide Meta with up to six gigawatts of computing capacity. That level of compute capacity requires data centre expansion and energy planning.

The partnership also includes a financial component. AMD granted Meta a performance-based warrant for up to 160 million shares, which could give Meta roughly a 10% ownership stake in the chipmaker if milestones are reached. Such arrangements are becoming more common as large AI customers seek deeper ties with suppliers to ensure access to future hardware.

The move follows another large agreement AMD signed with OpenAI, reinforcing a pattern where chip vendors and AI operators form relationships not buyer-seller contracts. For hyperscalers, securing compute supply has become as important as developing the software that runs on it.

Reducing dependence on a single supplier

Meta’s decision to expand its partnership with AMD also signals a change in how cloud operators manage risk in their supply chains. Nvidia still dominates the AI accelerator market, but heavy demand and limited supply have encouraged major tech firms to diversify their hardware sources.

Reports suggest that Meta is continuing to work with Nvidia and also investing in its own custom silicon projects. The AMD deal adds an additional layer to its strategy, helping Meta reduce reliance on a single vendor and maintain flexibility.

Any supply bottleneck can delay product launches or increase operating costs. By securing multi-year commitments, companies aim to stabilise pricing and availability.

The approach also reflects how AI infrastructure planning now resembles long-term energy or telecom planning. Companies are forecasting compute demand years ahead and negotiating supply deals that stretch into the future.

Signals for the wider cloud market

Previous cloud growth focused on virtual machines and software platforms. Current demand is driven by AI, which requires specialised hardware, unique AI data centre design, dense computing clusters that need specialist cooling and networking equipment, and huge quantities of power compared to ‘traditional’ data centres.

For enterprise customers, this trend may affect the availability and pricing of cloud AI services. As hyperscalers commit huge resources to internal AI workloads, they must balance their with demand from customers.

The deal reinforces AMD’s growing role in the AI hardware market. Long dominated by Nvidia, the sector is opening up as hyperscalers seek more options. Large multi-year supply agreements with companies like Meta and OpenAI suggest AMD is positioning itself as an alternative for AI workload hardware.

More broadly, the agreement shows that the next phase of cloud competition is shaped at the hardware level. Software models and AI applications often receive the most public attention, but their existence depends on the chips, power, water, and data centres underneath.

As companies scale AI in products and services, securing reliable compute capacity is becoming a priority. Meta’s move signals that the future of cloud computing may be determined by long-term infrastructure partnerships as well as advances in software.

Want to learn more about AI and big data from industry leaders? Check out AI & Big Data Expo taking place in Amsterdam, California, and London. The comprehensive event is part of TechEx and is co-located with other leading technology events, click here for more information.

AI News is powered by TechForge Media. Explore other upcoming enterprise technology events and webinars here.