During Intel’s Computex 2021 press conference, it was revealed that the 4th Gen Xeon Scalable processors, otherwise known as Sapphire Rapids will now be launched in the first half of 2022 instead of late 2021. This is despite the fact that multiple company execs had stated in the past that the next-gen server processors would launch by the end of 2021. This is likely the result of limited 10nm production which will become Intel’s primary process node in the second half of the year.
At the moment, only the mobile-grade Ice Lake and Tiger Lake (U and H) are based on the 10nm node, with the Ice Lake-SP lineup using the older iteration of the same process. Towards the beginning of 2022, everything from Intel’s desktop to notebook to server lineups will be using the 10nm node in one form or another. First, the octa-core Tiger Lake-H lineup along with the Tiger Lake-U refresh (which was demoed at today’s conference), then Alder Lake-S for desktop platforms, followed by Alder Lake-P for notebook PCs and convertibles.
Intel has already stated in the past that its overall PC shipments will drop as a result of this transition to the 10nm node. While Tiger Lake is based on the 10nm SuperFin node with the Willow Cove core architecture, both Alder Lake and Sapphire rapids will leverage the 10nm Enhanced SuperFin node along with the Golden Cove core architecture (+Gracemont for the former). The latter is expected to be a major upgrade for Intel’s server platform with up to 80 PCIe 5.0 lanes, CXL, DDR5 memory, and on-die HBM2 memory.
At the same time, the company was quick to state that it has doubled its chip production over the last four years, and external factors are the primary bottleneck on its manufacturing capabilities. The chipmaker expanded the production of substrates for chipsets and processors assembled at Intel’s Vietnam facility, with an overall investment of $4.5 billion towards the expansion of non-CPU products.