Intel launched its next-generation Xeon server processors on Tuesday in an effort to reclaim market share in the data centre space.
TakeAway Points:
- Intel debuted its next generation Xeon server processors on Tuesday in a bid to regain data center market share
- The company added that the cost of its Gaudi 3 artificial intelligence accelerator chips will be far less than that of its competitors’ offerings.
Intel’s Next Generation Xeon
The company also said that the cost of its Gaudi 3 artificial intelligence accelerator chips will be far less than that of its competitors’ offerings.
For Intel, which has been gradually losing market share in the data centre segment to Advanced Micro Devices (AMD), the sixth generation Xeon chips are essential.
According to research from Mercury Research, Intel’s share of the x86 data centre processor market has decreased by 5.6 percentage points to 76.4% over the past year, with AMD now having a 23.6% share.
Because AMD uses Taiwan Semiconductor Production Co. (TSMC) to build its chips, the company has been able to gain market share due to errors in Intel’s production process.
Xeon Specifications
There are two primary variants of the Xeon 6 server processors: an enhanced, more potent model and a “efficiency” model marketed by Intel as a substitute for chips from previous generations.
They now need roughly 67% fewer server racks to get the same level of processing power as their second generation chips, thanks to the efficiency model, which is made to provide video, websites, and carry out database computations.
“Simply put, performance up, power down,” Intel Chief Executive Pat Gelsinger said at the Computex trade fair in Taipei, where he gave a presentation of the server.
The more powerful performance model is intended to handle jobs requiring greater processing power, such as processing the calculations required to produce replies from intricate AI models.
The “performance” Xeon processor will be released in the third quarter, while the “efficiency” model is accessible starting on Tuesday. Intel intends to release more variants in the upcoming year. The corporation decided to switch to a different production technique, which resulted in a one-year delay for the sixth generation chip.
Gaudi 3 Accelerator Kit
During a press meeting, Intel revealed that the Gaudi 3 accelerator kit, which comes with eight AI chips, costs roughly $125,000, while the Gaudi 2 model, which is older, retails for $65,000.
Speaking in Taipei, Gelsinger stated that the costs appeared “very appealing,” particularly when compared to rival companies.
“In other words, it crushes the competition.”
However, AMD and Nvidia do not talk about how much their chips cost. Thinkmate, a bespoke server manufacturer, states that a comparable HGX server system with eight Nvidia H100 AI chips can cost more than $300,000.
With the Gaudi 3 AI chip, which Intel unveiled in April, the company is positioning it as a far more affordable and competitive option to Nvidia’s H100 CPUs.
Additionally, Intel said on Tuesday that Lunar Lake, its upcoming laptop chip, features a more potent AI processor and consumes 40% less power. According to Intel, the chip will be shipped in the third quarter.
Gelsinger expressed his excitement at witnessing a large number of new PCs running Lunar Lake.
“It will power the largest number of next-generation AI PCs in the industry; we are committed to the AI PC.” he added.
Additionally, he expressed gratitude to TSMC, citing their assistance with numerous essential technologies for Lunar Lake as crucial.
With an eye towards future iterations, Gelsinger stated that Panther Lake will be released in 2025 and Arrow Lake the following year.
