Intel Corp. shares jumped sharply after a report indicated that Alphabet Inc.’s Google plans to rely on the company to produce more than 3 million specialized artificial intelligence chips by 2028.

According to The Information, Google selected Intel to manufacture a portion of its tensor processing units (TPUs) following extensive testing of the chipmaker’s capabilities. The move highlights growing confidence in Intel’s technology as demand for advanced semiconductor manufacturing continues to outpace supply.

The report also noted that companies are increasingly turning to Intel as Taiwan Semiconductor Manufacturing Co. (TSMC) struggles to meet soaring demand for chip production capacity. Following the news, Intel’s stock climbed as much as 13% in New York trading, marking its biggest gain in a month and capping a year in which the shares have nearly tripled in value.

Neither Intel nor Google provided immediate comment on the report.

The rally adds to a wave of optimism surrounding Intel’s recent performance. The company recently reached a record high after issuing a sales forecast that exceeded Wall Street expectations, suggesting that CEO Lip-Bu Tan is making meaningful progress in reviving the once-struggling chipmaker. After securing significant investments last year to stabilize its finances, Intel is now showing signs of operational improvement.

The Information also reported that Nvidia is evaluating Intel’s manufacturing technology for a future processor design that would combine multiple graphics chips into a single unit. While details remain unclear, such developments could further validate Intel’s efforts to reestablish itself as a competitive force in advanced chip production.

Still, questions remain about how extensively major clients like Google will depend on Intel’s foundry services compared to its chip packaging business. While packaging has traditionally been less critical and lower-margin than fabrication, it has gained importance as companies increasingly integrate multiple chips into a single package to enhance performance—particularly for data center and AI applications.

Intel has already disclosed a multibillion-dollar backlog in packaging orders, reflecting rising demand for these capabilities. However, even a large order such as 3 million chips would not immediately transform its still-unprofitable manufacturing division, representing roughly a month’s output for a major facility.

Nevertheless, the willingness of major technology players to entrust Intel with key AI-related workloads could significantly boost its credibility and improve its chances of securing additional high-profile customers.

Meanwhile, Google has emerged as a leading developer of in-house AI chips, challenging the dominance of Nvidia in certain segments. Its TPUs have gained traction in Silicon Valley, particularly for inference tasks—where trained AI models generate responses—while Nvidia’s GPUs remain the industry standard for training more complex systems.

As competition intensifies across the AI semiconductor landscape, Intel’s potential partnership with Google could mark a pivotal step in its effort to reclaim relevance in one of the fastest-growing sectors in technology.