Intel's Gaudi AI chips are far behind Nvidia and AMD and won't even reach the $500 million target

“We will not achieve our $500 million revenue target for Gaudi in 2024,” CEO Pat Gelsinger said on the company's third-quarter 2024 earnings call just today.

Although Intel launched its latest Gaudi 3 accelerator just last quarter, Gelsinger said, “overall adoption of Gaudi has been slower than expected, with adoption rates impacted by the Gaudi 2 to Gaudi 3 product transition and the software's usability.” became.”

Despite the missed target, Gelsinger says, “We remain encouraged by the market available to us.” There is a clear need for superior solutions [total cost of ownership] based on open standards, and we continue to improve Gaudi’s value proposition.”

Later in the call, Gelsinger appeared to have a few things to complain about, pointing out that the industry's huge spending on AI chips so far has been focused on training AI models in the cloud. “Training is about building the weather model, not using it,” he says, suggesting once again that in the long run it may be more important to integrate AI into all chips, not just those in the cloud.

Intel today reported quarterly profit of $13.3 billion, down 6 percent from a year ago but up from last quarter – and losses of a whopping $16.6 billion. However, those losses were due to impairments and $18.5 billion in restructuring costs, the cost of Intel's decision to retool for greater profitability in the future.

Last quarter, the company announced a $10 billion cost-cutting plan and over 15,000 layoffs. Now it also details some of the structural changes within the company – including moving its edge computing business into the Client Computing Group, which generally handles its desktop and laptop chips, and integrating its software teams into the company's core business areas.

Gelsinger says Intel will “focus on fewer projects, with the top priority being to maximize the value of our x86 franchise in the client, edge and data center markets.”