Intel has struggled for more than two decades to develop a successful GPU/accelerated computing strategy, going all the way back to the aughts and the ill-fated Larrabee effort. Its most recent efforts centered around Ponte Vecchio and Gaudi chips, neither of which have gained any traction.
Still, CEO Lip-Bu Tan has made AI a top priority, reorganizing the corporate structure around a new unit, the AI Group, but that was dealt a setback last November when Sachin Katti, the head of the group, left Intel for a job at OpenAI.
Still, this shows Intel is firmly behind GPU/accelerator, says Jon Peddie, president of Jon Peddie Research. “The company has been inconsistent in its messaging and its layoff explanations, causing needless speculation and confusion,” he said. “His biggest challenge will be the Intel management and marketing machine. But that bureaucracy is not unknown to him.”
The results of his involvement will probably not be notable for a year, says Peddie. “He can’t do much about the long-term projects already underway, but he can help them avoid a few land mines, maybe,” he said.
