Intel 7nm announces an extension: it does not rule out the transfer of over-computing GPU chips to a tripartite contract.

Despite a good second-quarter results for fiscal 2020, revenue rose 20 percent year-on-year and net profit of $5.1 billion increased 22 percent year-on-year. The disclosure comes as Intel confirms the less encouraging news that the 7nm will be delayed by six months, or a year later than originally planned, due to a process defect, and that the first consumer-grade CPU products are not expected to be available before the end of 2022.

On the 7nm issue, Intel talked only about the CPU, not the GPU, and The TPU’s view is that GPU products are decoupling from Intel’s own process roadmap and are being built on a pipeline based on Samsung or TSMC.

Intel 7nm announces an extension: it does not rule out the transfer of over-computing GPU chips to a tripartite contract.

Anand noted that Intel CEO Bob Swan stressed the need for Intel to be more pragmatic, and it was clear that the transfer to a three-way foundry was in line with that positioning.

The most solid 7nm product in Intel’s product plan is Ponte Vecchio (“Vickio Bridge”), built on Intel’s top Xe_HPC superGPU core, which will be deployed to the U.S. Department of Energy’s 10 billion-times super-counting Aurora in 2021, although Intel’s current calibre is still not giving up hope that it is working on 7nm Xe_HPC, sly borre admits, and is evaluating the outsourcing of some of its manufacturing links to the three-way fab. “Back”.

Once due to the 10nm delay, Intel vowed to catch up with wasted time by 7nm and make Moore’s Law shine again. However, it now seems to backfire, once to “short-lived” 10nm will dominate for at least the next 2 to 3 years, yes, next year there will be 14nm Rocket Lake desktop processor!

Intel 7nm announces an extension: it does not rule out the transfer of over-computing GPU chips to a tripartite contract.