That is unless the US government chooses to ramp subsidies for R&D and capital expenditure for chip making to see Intel through what are set to be a difficult few years.
Intel has confessed its 7nm manufacturing yield is a year behind its internal roadmap (see Intel goes foundry for 7nm due to yield issues) and said the solution to this problem is the use of foundry manufacturing services. The stock market reacted badly to the revelation and marked Intel down from about $60 to about $50 per share.
Such major manufacturing transitions have happened many times before and can take many years, but Intel is identifiably on the slope to fab-lite or specialized fab status. The more it uses foundries, the more it supports their case and diminishes its own reason to stay in leading-edge manufacturing. The list of companies that have undergone the journey is long including: AMD, Globalfoundries, IBM Microelectonics, STMicroelectronics, Infineon, NXP Semiconductors and Texas Instruments, to name but a few.
Intel – despite being the largest semiconductor company by revenue – is not immune. But if Intel is no longer a leading-edge chip manufacturer it is not clear that there are any More-than-Moore strings to its bow. If Intel goes fab-lite it would also come under technical and economic pressure to go fabless.
In a conference call with financial analysts to discuss 2Q20 financial results CEO Bob Swan said Intel would continue to invest in its manufacturing process technology roadmap but that the company would be pragmatic about using the best process available – internally or externally. Given that Intel is no longer best-in-class in semiconductor manufacturing, and is unlikely to get there any time in the next five years, it would appear that Intel will be investing large sums of money to little purpose. Something that investors are unlikely to tolerate for long.
Next: Delays at 10nm, 7nm