In 2005, Intel CEO Paul Otellini surprised the company’s board of directors. The New York Times reported that he suggested Intel could buy Nvidia for “up to” $20 billion.
Even some Intel executives believe that Nvidia’s designs could eventually play a key role in the data center, according to Times sources (“two people familiar with board discussions”). He says he was thinking about it. This idea would become a reality in the modern AI boom, but the board opposed it. There were concerns about integrating the company, as it would be Intel’s most expensive acquisition. Otellini retreated and that was it.
Instead, Intel’s board supported Larabee, an in-house graphics project now helmed by CEO Pat Gelsinger. It used Intel’s x86 technology, and the GPU was a kind of hybrid between a CPU and a GPU. Intel eventually canceled the project, but later returned to graphics with the Xe and Arc projects.
On the AI front, Intel has made several acquisitions, including Nervana Systems and Movidius in 2016 and Habana Labs in 2019. But none could match Nvidia’s current status as a behemoth with a market capitalization of more than $3 trillion. Intel’s Gaudi 3 AI chip is positioned as a cheaper alternative to Nvidia products, but the company is largely considered to have missed the boat when it comes to AI.
Intel is now a much smaller company than Nvidia, at less than $100 billion, due to other struggles building customers in manufacturing and foundry businesses. Intel is working on NPUs for consumer technology, including laptops and now desktop CPUs.
This isn’t the only time Intel has given up on getting into AI early. In 2017 and 2018, Intel had the opportunity to purchase a stake in OpenAI when it was still a small nonprofit research company. But then-CEO Bob Swan balked at the deal, believing the AI model was far from reaching a widespread market.
