According to informed sources, Alphabet's Google has placed an order with Intel for the production of over 3 million Tensor Processing Units (TPUs) by 2028.
Reports indicate that NVIDIA is also evaluating Intel's technology for manufacturing processors that combine four graphics chips into a single unit, although the company has not yet placed an order.
Driven by signs of steady business improvement under CEO Pat Gelsinger, Intel's stock surged over 9% in early trading, extending its year-to-date gain to nearly 169%.
The potential order for Google's custom AI chips is expected to boost Intel's foundry business, as the company seeks to reclaim the chip manufacturing leadership it lost to TSMC due to years of management missteps.
A surge in chip demand fueled by the AI boom has made it difficult for TSMC to provide sufficient supply. This capacity shortage has prompted several major AI chip design firms to turn to Intel.
Jacob Bourne, a technology analyst at eMarketer, stated that this news "proves the giants in the AI field are scrambling to diversify a supply chain that remains highly concentrated with TSMC."
Since Gelsinger took the helm, Intel has secured billions in investments from the Trump administration, NVIDIA, and SoftBank.
Intel has also successfully secured Tesla as the first major customer for its next-generation 14A process, producing chips for the advanced AI chip complex "Terafab" that Elon Musk is planning in Austin.
An official last month stated that the Trump administration has also been working to secure business for Intel.
"Beyond the usual diversification needs, Google and NVIDIA have more incentive than ever to work with Intel. Supporting Intel means supporting domestic U.S. manufacturing, which is crucial for relationships with the U.S. government," said Gil Luria, an analyst at D.A. Davidson.
Last month, reports indicated that after more than a year of intensive negotiations, Intel had reached a preliminary agreement with Apple to produce some chips for Apple's devices.
Meanwhile, Google has been working to establish its custom AI chips as a viable alternative to the GPU dominance of NVIDIA, with sales of its Tensor Processing Units (TPUs) becoming a significant driver of revenue growth for its cloud business.
Comments