Intel’s Gaudi AI chips lag far behind Nvidia and AMD, won’t even reach $500M goal

“We will not achieve our target of $500 million in revenue for Gaudi in 2024,” CEO Pat Gelsinger said today on the company’s Q3 2024 earnings call.

Although Intel has just launched its latest Gaudi 3 accelerator Last quarter, Gelsinger said, “Overall adoption of Gaudi has been slower than we anticipated as adoption rates were impacted by the product change from Gaudi 2 to Gaudi 3 and the ease of use of the software.”

Despite missing the target, Gelsinger says, “We are encouraged by the market we have. There is a clear need for resolution with superiors [total cost of ownership] Based on open standards, and we are continuing to enhance the Gaudi value proposition.”

Later on the call, Gelsinger had some sour experiences to share, pointing out that until now, the bulk of the industry’s spending on AI chips has been focused on training AI models in the cloud. “Training is building a weather model, not using it,” he says. suggestion once again Putting AI into all chips, not just in the cloud, may be more important in the long run.

Intel today reported quarterly revenue of $13.3 billion, down 6 percent year over year but up from the previous quarter – and a massive loss of $16.6 billion. But those losses were based on a $18.5 billion loss and restructuring charges, the cost of Intel’s decision to rework itself for greater profitability in the future.

Last quarter it had announced a $10 billion cost-cutting plan More than 15,000 layoffs, And it’s also now detailing some structural changes inside the company — including moving its edge computing business to the client computing group that typically handles its desktop and laptop chips, and moving its software teams to the company’s Involves integrating into core business units.

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.”