The Cloud Czars’ Move into Chips Doesn’t Bode Well for Inte…
Intel (NASDAQ:INTC) has been a top chip maker since forever, but there could be trouble looming for Intel stock in the form of the Cloud Czars. It is no longer news that the five of them, Apple (NASDAQ:AAPL), Amazon.com (NASDAQ:AMZN), Alphabet (NASDAQ:GOOGL), Microsoft (NASDAQ:MSFT) and Facebook (NASDAQ:FB), are all in the chip business.
It’s a way they can differentiate their clouds from one another.
They all have plenty of capital for designing chips, and the vast majority of chip companies, like Nvidia (NASDAQ:NVDA), Qualcomm (NASDAQ:QCOM) and Advanced Micro Devices (NASDAQ:AMD), are now just design houses.
That’s because the capital costs of making chips, based on “Rock’s Law” continues to increase, doubling every four years.
Thus there are only four microprocessor manufacturers left – Samsung Electronic (OTCMKTS:SSNLF), Taiwan Semiconductor Manufacturing (NYSE:TSM), privately-owned Global Foundries, and, of course, Intel.
Intel shares remain well below their dot-com era peak because it must compete with its customers. Should it?
Differentiating clouds, in other words, is the new AI arms race. And the best way to differentiate is to design your own silicon.
What started on the server side of clouds-and-devices is naturally going to migrate to the client side. Microsoft’s Project Brainwave envisions custom silicon in clients, and Apple is moving away from Intel for modems and other iPhone chips.
A Two-Company Solution for Intel Stock
While the Cloud Czars can design chips, they can’t make them.
It’s why I continue to believe, as I did in 2016, that the best thing for Intel stock would be for the company to break itself in two. Have one company that designs chips, and one company that makes chips.
Have one company that is focused on finding competitive niches within the semiconductor market and serve them, have another that is focused on what chip designers want in terms of manufacturing capability.
These are, increasingly, two different markets, and Intel’s continued refusal to see them that way is hampering it in both areas. Intel has been very, very late to supporting 10 nanometer designs in its manufacturing, and it is getting killed by old design flaws, which are destroying its reputation.
The general good health of the semiconductor industry, with shipments expected to grow over 12% this year, has continued to cover up this weakness. But Intel’s new CEO search offers it an opportunity to re-start the company from a blank sheet of paper.