Amazon Strikes Up to $60B Multi-Year Data Center Chip Deal with Qualcomm as Saudi Compute Ties Deepen
SEATTLE / SAN DIEGO — In a move set to reshape the global computing landscape, Amazon has entered into a landmark multi-year agreement with Qualcomm allowing the cloud and e-commerce giant to purchase up to $60 billion in artificial intelligence chips and data center infrastructure products. The commercial pact comes as hyperscale cloud providers aggressively move to diversify silicon supply chains and curb long-standing dependence on Nvidia.
The strategic collaboration spans multiple silicon generations, focusing specifically on AI inference—the computational phase where trained foundation models process user prompts, reason, and execute live enterprise workloads. In parallel, both companies are co-engineering high-speed optical connectivity solutions delivering data transfer rates of up to 1.6 Terabits per second (Tbps) to address severe internal networking bottlenecks inside modern hyperscale facilities.
While the $60 billion figure represents the potential commercial ceiling tied to deployment volumes rather than an immediate upfront payout, the financial structuring underscores deep mutual alignment. Qualcomm granted Amazon stock warrants to acquire approximately $4 billion in shares at an exercise strike price of $161.26 per share, with vesting milestones tied directly to Amazon's volume purchases of Qualcomm data center silicon. The equity incentive effectively grants Amazon a vested financial stake in Qualcomm’s enterprise success, sending Qualcomm’s shares up more than 3 percent following the announcement.
For Qualcomm, the transaction marks a critical strategic pivot. Historically dominant in smartphone processors and cellular basebands, the company has faced softening consumer handset cycles and Apple’s phased transition toward in-house modems. In response, Qualcomm unveiled a dedicated data center silicon roadmap featuring the Dragonfly C1000 CPU and the AI200, AI250, and AI300 inference accelerators, targeting $15 billion in annual data center chip revenue by 2029.
The agreement simultaneously reinforces Amazon Web Services (AWS)' dual-track silicon strategy. While AWS' internal custom silicon division—led by Trainium and Inferentia—surpassed an annual revenue run-rate of $25 billion by the end of June, teaming with Qualcomm provides essential multi-vendor redundancy against global semiconductor supply chain bottlenecks.
Direct Implications for Saudi Arabia & the GCC: The U.S. mega-deal carries immediate operational significance for Saudi Arabia, where Qualcomm's data center roadmap is already being deployed at national scale:
200 MW Sovereign Cluster with HUMAIN: Qualcomm and Saudi Public Investment Fund (PIF)-backed HUMAIN have agreed to deploy 200 megawatts of advanced AI compute infrastructure across the Kingdom beginning in 2026, powered by Qualcomm's AI200 and AI250 accelerator suites.
Riyadh AI Engineering Center: Qualcomm is establishing a dedicated AI engineering facility in Riyadh to supervise local cluster optimization and expand the domestic data center ecosystem.
Live Enterprise Workloads: In August, creative software giant Adobe transitioned regional AI inference workloads onto HUMAIN's domestic infrastructure in Saudi Arabia, running directly on Qualcomm hardware.
With hyperscalers confronting global energy, cooling, and transmission constraints, Qualcomm’s low-power architecture shifts competition from raw compute speed toward power efficiency and cost-per-inference. The deal positions the Kingdom—alongside global cloud hyperscalers—at the forefront of an industry-wide transition toward diversified, energy-optimized AI infrastructure.
