According to a report by DigiTimes, Intel is expected to offer competitive prices on its Xeon CPU lineup to tackle AMD’s EPYC chips and retain its competitive lead within the server x86 segment.
Intel To Tackle AMD’s EPYC CPUs With Competitive Prices On Its Current & Upcoming Xeon Chips
The report claims that Intel will be tackling AMD with competitive prices on its Xeon CPU lineup that aims at the server segment. Intel’s main competitive edge at the moment is that it is producing x86 server chips at its own fabs and has also announced plans to outsource fabrication & production to other chip makers if it seems fit. Intel also holds a larger revenue compared to AMD which means it can offer subsidies to large-scale & volume partners to retain its competitive leadership.
According to DigiTimes, Intel has pivoted on its server strategy to fight a supply-constrained AMD, and is now offering more competitive pricing on its server processors. However, it’s noteworthy that Intel hasn’t changed its official recommended pricing, so these deals are likely coming as part of volume purchases with its largest customers.
AMD is currently tightly constrained with TSMC’s 7nm process node and even though they have prioritized server chips over consumer parts, the N7 process is still gobbled up by the rival chip-makers. So Intel with its in-house development and manufacturing can gain an edge on server customers.
The move comes from Intel after AMD has continued to gain market share in the x86 server CPU segment with its EPYC class chips. The company has already breached the double-digit server x86 share and posted a 14-year high. The company (AMD) now plans to extend that lead to reach the historical high of 20-25% which it had achieved with its Opteron CPUs all the back in 2003.
While Intel offers competitive prices on its existing Xeon server lineup, it should be noted that the company has severely delayed its upcoming server processors. The company pushed back upon its Sapphire Rapids-SP plans leading US DOE (Argonne National Laboratory) to sign a contract for an intermediary Supercomputing solution known as Polaris which will be powered by AMD’s EPYC & NVIDIA’s HPC GPUs until the Aurora Supercomputer becomes operational. Polaris will offer up to 44 PFlops of peak computing horsepower which is peanuts compared to the expected 1 ExaFlop of performance that Aurora powered by Intel’s Sapphire Rapids-SP and Ponte Vecchio GPU will offer.