Intel struggles with Gaudi 3 roll out
Cette publication existe aussi en Français
Intel is struggling with the roll out of its third generation AI accelerator chip, Gaudi 3.
The latest results show that the roll out of the 3nm chip has been hit by a lack of software portability, despite a deal to use it in IBM’s cloud. The chip is intended to compete with Nvidia in the fast-growing AI data centre chip market.
“The overall uptake of Gaudi has been slower than we anticipated as adoption rates were impacted by the product transition from Gaudi 2 to Gaudi 3 and software ease-of-use. As a result, we will not achieve our target of $500 million in revenue for Gaudi in 2024,” said Pat Gelsinger, CEO of Intel. “That said, taking a longer-term view, we remain encouraged by the market available to us.”
Intel aims at Nvidia with Gaudi 3 AI chip
Two co-founders from Habana Labs, which developed the Gaudi architecture, left Intel back in September to set up their own AI chip company called Touch.
Intel also launched its Xeon 6 processor, code-named Granite Rapids, for data centre applications and is working with Amazon Web Services on a custom version of the chiplet-based processor.
“This solidifies our position as the head node of choice in AI servers,” said Gelsinger. “Greater than 70% of GPU accelerated servers are already using Intel Xeon as the host CPU, and we have a significant opportunity to build on this as we continue reestablishing Xeon’s competitive strength and market leadership.
The company is also looking to sell off a stake in its Altera FPGA business in 2025. Revenue increased 14% sequentially and operating profit turned positive in Q3. “We remain focused on selling a stake in Altera on a path to its IPO in the coming years. To that end, we have begun discussions with potential investors and expect to conclude in early 2025.
If you enjoyed this article, you will like the following ones: don't miss them by subscribing to :
