Intel is actually killing Falcon Shores, his next generation GPU for work calculations and high -performance artificial intelligence work.
The move arrives while Intel tries to correct the course after a series of disappointing launches of the product and historical losses, while competitors such as AMD and Nvidia earn ground.
Intel Michelle Johnston Holthaus co-centered declared that during the call of profits of the fourth quarter of the company on Thursday, the company “will exploit Falcon Shores as an internal test chip, without taking it to the market”.
“The IA Data Center (…) is an attractive market for us,” Holthaus said during a gain call on Thursday. “(B) UT I’m not happy where we are today. We are not yet participating in the Data Center market based on cloud significantly (…) one of the immediate actions I have taken is to simplify our roadmap and concentrate our resources. “
The goal, on the other hand, will be on the Jaguar banks, which Holthaus has defined Intel’s opportunity to “develop a system solution on a rack scale (…) to deal with the data center for in a wider way”.
Holthaus tempered expectations for Falcon Shores last month, when he insinuated that it was an “iterative” step compared to the previous chip dedicated to the company dedicated data center, Gaudi 3.
“We really have to think about how we go from Gaudi to our first generation of Falcon Shores, which is a GPU,” Holthaus said in an apparition to the global technological conference in Barclays. “And I’ll tell you now, will it be wonderful? No, but it’s a good first step. “
Gaudi 3 has been widely seen as a lack for Intel. In November, the company said that it would not be able to achieve its $ 500 million in the sales of Gaudi 3 due to software problems. Today, few important service providers beyond IBM have committed themselves to using the chip.
Intel faces an uphill battle in the space of the AI data centers. The Rivale AMD plans to earn about $ 7 billion with chips AI in 2025, while Nvidia, the AI chip company to beat, could reach $ 195 billion in revenue in the tax year 2026, according to some analysts.
“While I think of our artificial intelligence opportunity, my goal is on the problems that our customers are trying to solve, in particular the need to reduce costs and increase the efficiency of the calculation,” said Holthaus. “As such, a unique approach for everyone will not work and I can see clear opportunities to exploit our fundamental resources in new ways to guide the most convincing total cost of ownership through the continuum.”
