Useful information
Prime News delivers timely, accurate news and insights on global events, politics, business, and technology
Intel is effectively killing Falcon Shores, his next -generation GPU for high performance computing and AI workloads.
The measure occurs when Intel tries to correct the course after a series of disappointing products and historical losses, while competitors like AMD and NVIDIA gain ground.
Intel’s co-zo, Michelle Johnston Holthaus, said during the company’s quarter-quarter’s profit call on Thursday, the company “will take advantage of Falcon’s coasts as an internal test chip, without taking it to the market.”
“The AI data center (…) is an attractive market for us,” Holthaus said during a profit call on Thursday. “(B) ut I am not happy with where we are today. We are still not participating in the AI Data Centers market based on the cloud in a significant way (…), one of the immediate actions I have taken is to simplify our road map and concentrate our resources. “
On the other hand, the approach will be in Jaguar Shores, which Holthaus called Intel’s opportunity to “develop a solution at the system at rack scale (…) to address the most widely AI data center.”
Holthaus tempered the expectations for Falcon Shores last month, when she realized that it was an “iterative” step on the chip of the Data Center of the company’s previous dedicated, Gaudi 3.
“We really have to think about how we pass from Gaudi to our first generation of Falcon shores, which is a GPU,” Holthaus He said in an appearance at the Barclays Global Technology Conference. “And I will tell you right now, will it be wonderful? No, but it’s a good first step. “
Gaudi 3 was widely seen as a failure for Intel. In November, the company said it could not reach its $ 500 million objective in Gaudi 3 sales due to software problems. Today, few important service providers beyond IBM have pledged to use the chip.
Intel faces a uphill battle in the chips space of the AI data center. The AMD rival expects to win around $ 7 billion in AI Chip revenues in 2025, while Nvidia, the firm of AI Chip To Beat, could reach the revenues of $ 195 billion in fiscal year 2026, according to some analysts.
“While I think about ours, my approach is the problems that our clients are trying to solve, especially the need to reduce the cost and increase the efficiency of calculation,” said Holthaus. “As such, a unique approach for all will not work, and I can see clear opportunities to take advantage of our central assets in new ways to boost the most convincing total cost of property in the entire continuum.”
