Google taps Intel for 3M AI chips in 2028

Alphabet’s Google has reportedly entered into a supply agreement with Intel to manufacture over three million specialised artificial intelligence chips starting in 2028. The partnership marks a significant shift in the tech giant’s procurement strategy as it seeks to diversify its hardware manufacturing partners. The deal involves the production of custom silicon designed to power the next generation of data centres and machine learning workloads.

The move represents a strategic pivot for Google, which has historically relied heavily on Taiwan Semiconductor Manufacturing Co. (TSMC) for its high-end processor requirements. By engaging Intel’s foundry services, the search giant aims to mitigate supply chain risks associated with geographic concentration in the semiconductor industry. The massive volume of the order highlights the scale of infrastructure investment required to maintain a competitive edge in the evolving AI landscape.

Intel stands to benefit significantly from this collaboration as it attempts to revitalise its foundry business and challenge the dominance of Asian manufacturers. The agreement serves as a major validation of the American firm’s advanced node technology and its ability to handle high-volume orders for complex AI components. Industry analysts suggest that securing a client of Google’s stature is a crucial milestone for Intel’s long-term turnaround plan.

The shift in manufacturing also reflects broader industry trends where major cloud providers are designing their own proprietary hardware. These custom chips are often optimised for specific neural network tasks, offering better performance and energy efficiency than general-purpose processors. As demand for generative AI services continues to climb, the pressure on global manufacturing capacity has intensified, forcing companies to seek alternative production facilities.

Taiwanese manufacturers currently lead the market in advanced logic processes, but geopolitical tensions and logistical bottlenecks have prompted a search for more diverse production hubs. Google’s decision to tap Intel suggests a growing appetite for domestic manufacturing capabilities in the United States and Europe. The logistics of producing three million units will require significant coordination over the next several years to ensure production targets are met.

The specific technical details of the chips have not been disclosed, though they are expected to be integrated into Google’s vast cloud infrastructure. These processors will likely play a central role in training large language models and serving real-time AI applications to millions of users worldwide. The long lead time for the 2028 rollout indicates the complexity involved in securing foundry capacity and finalising silicon designs.

Looking ahead, the success of this partnership could influence how other technology majors distribute their manufacturing contracts between established leaders and emerging foundry competitors. Google is expected to continue its multi-sourcing strategy to ensure a steady flow of hardware for its expanding digital services. Market observers will be monitoring Intel’s ability to meet the rigorous quality and timeline requirements stipulated in this high-profile manufacturing agreement.