Google said on the 9th (local time) that it has extended its artificial intelligence (AI) chip partnership with U.S. semiconductor corporations Intel.
Google has used Intel's central processing units (CPUs) in its cloud servers for 30 years, and under the new multiyear strategic collaboration, it will deploy Intel's next-generation semiconductor capabilities across its infrastructure, including AI data centers.
Under the partnership, Google will adopt Intel's latest "Xeon 6" CPUs and other key semiconductors and will co-develop a "custom ASIC-based infrastructure processing unit (IPU)" to maximize efficiency and performance in large-scale AI environments. The size of the deal and specific term were not disclosed.
Amin Vahdat, Google Cloud vice president for AI infrastructure, said in a statement that "CPU and infrastructure acceleration technologies are the core foundation of AI systems," adding, "We are confident that Intel's Xeon roadmap will meet the performance and efficiency demands of ever-growing AI workloads."
The market assessed that as Intel's "Xeon 6" CPUs are used for AI training and inference tasks, Intel's chances of securing a stronger foothold in the AI market—so far led by Nvidia—have increased. Intel earlier said it would participate in the large-scale semiconductor manufacturing facility "Terafab" being pushed by Tesla CEO Elon Musk.
Lip-Bu Tan, Intel's chief executive officer (CEO), said, "Scaling AI requires more than just accelerators; it needs a balanced system," adding, "CPUs and IPUs play a key role in delivering the performance, efficiency, and flexibility demanded by today's AI workloads."
On the news of the companies' deal, Intel shares jumped 5% intraday, and shares of Alphabet, Google's parent, also closed slightly higher.