Intel shares rise on report that Google could place a large TPU manufacturing order

Intel’s stock climbed after multiple outlets, citing industry reporting, said Google is looking at Intel as a manufacturing partner for a large run of its in-house AI accelerators. The reports, which were not accompanied by any statement from Intel or Alphabet, point to a potential order spanning millions of tensor processing units, or TPUs, with deliveries targeted for 2028.

The central claim is that Google has tapped Intel to manufacture more than 3 million TPUs through 2028, according to reporting referenced by Yahoo Finance and other technology industry coverage. Intel’s foundry business, which focuses on producing chips for external customers, has been one of the company’s strategic bets as it tries to win share in advanced semiconductor manufacturing.

A separate track in the same reporting ecosystem suggests that Nvidia is also looking at Intel more broadly, including evaluation of Intel’s next-generation manufacturing process and packaging capabilities. Packaging matters for AI accelerators because modern systems often combine multiple chip dies and specialized interconnects to deliver more compute power in a single product.

One reason these reports matter for Intel is that they would announcement additional traction in advanced AI-chip production. Unlike a traditional CPU vendor that sells completed chips, a foundry model depends on persuading customers that the manufacturing process, yield, and packaging ecosystem can meet demanding performance and reliability requirements for large-scale AI deployments.

For Google, TPUs are the core compute building blocks for training and running many of its machine-learning workloads. Google has been steadily expanding internal silicon to reduce reliance on third-party GPUs, improve performance-per-watt, and control supply for ever-larger AI systems. If Google’s next TPU runs include Intel manufacturing, it would also reduce single-supplier risk for a company whose AI spending depends on uninterrupted chip availability.

The potential partnership also plays into a wider industry shift. Multiple reports describe AI-chip capacity constraints and a growing willingness among major buyers to diversify manufacturing and packaging across more than one supplier. In that context, Intel’s appeal would be less about replacing the leading foundry outright and more about adding a credible second source for high-volume accelerator production.

Still, key details remain unclear because the companies have not published confirmations or contract terms. The reporting does not make public the size of the full program beyond the headline million-unit figure, the exact TPU generation referenced, the percentage of Google’s total future TPU demand that Intel would supply, or whether the arrangement is binding versus contingent on further validation.

What to watch next is whether Intel or Alphabet addresses the reports, either through an investor communication or a formal procurement update. Investors may also look for any technical milestones tied to advanced packaging and process readiness, since manufacturing partnerships at this scale typically hinge on validated yields and system-level performance before large volumes are finalized.