by Jean-Louis Gassée
Today, we take a look at Intel. Can CEO Pat Gelsinger, perform miracles?
I’m a great admirer of Intel CEO Pat Gelsinger. He embodies a rare combination of personal virtue and technological savvy. If you have the time and inclination, see the Monday Notes Intel 2.0: The History and Culture Gelsinger Inherits and Intel 2.0 Reboot.
After becoming Intel’s first Chief Technology Officer, Gelsinger left the company in September 2009 to become President and Chief Operating Officer of EMC. In 2012, he moved to cloud computing company VMware as CEO…and then rejoined Intel, as CEO this time, two years ago in February 2021.
Upon taking the helm of one of our most venerated tech companies, the new boss set a resolute tone:
“We’re bringing back the execution discipline of Intel. I call it the Grovian culture that we do what we say we will do. That we have that confidence in our execution. That our teams are fired up. That we said we’re going to do x, we’re going to 1.1x, every time that we make a commitment. That’s the Intel culture that we are bringing back.”
We’re two years into the Gelsinger regime and, unfortunately, the latest numbers don’t paint a pretty picture:
(You’l find more mind-numbing details in Intel’s 10-K and the slightly more digestible quarterly earnings presentation.)
In a downdraft environment that has led to tens of thousands of layoffs across the tech sector, Gelsinger, to his credit, is concentrating on talent where it’s needed, pushing forward with a 5 nanometer chip. As it says in the quarterly presentation “5 nodes in 4 years”.
But is it small enough, soon enough? 5 nanometer chips are already available elsewhere; TSMC is on its way to 3 nanometers. (See Yahoo’s Intel’s Earnings Call transcript.)
Intel lags behind in the technology game.
Intel runs many businesses of which three are notable: CCG (chips for PCs, $6.6B in Q4 2022, down 36% compared to Q4 2021), DCAI (high-power processors, mostly for Cloud applications, $4.3B, down 33%), NEX (Networking products, $2.1B, down 1%). A few other businesses contribute smaller, almost immaterial amounts of revenue and occasional losses. (We’ll look at one of them later.)
Once upon a time, Intel generated fat margins by milking its commodity business for PCs (CCG) and high-end server chips (DCAI). Today, in the midst of a promising Cloud Computing and AI market, CCG and DCAI barely make money, only $0.4B in Q4 22. This is a grave problem that casts doubt on Intel’s ability to produce revenue in a product segment that they used to own.
The reason they no longer own the segment is obvious: Amazon Web Services is the acknowledged Cloud Computing leader and trendsetter. AWS has led the migration away from Intel’s DCAI products, and has made big bets on the Graviton line of high-end chips based on the ARM-64 architecture.
There’s more bad news. Intel’s PC chips segment, CCG, offers little room for growth. When the desktop/laptop market was expanding, Intel’s OEM clients, ASUS, Dell and others, had to accept Intel’s higher prices. Now, Gelsinger faces a combination of lower volumes and lower prices. In this context, Apple’s Mac + iPad business ($18B last quarter, 0% growth year-to-year in a declining market) raises questions about the future of Wintel devices. A likely answer is that Apple will gradually expand its share of the market without necessarily disrupting it. For its part, Microsoft, the Win in Wintel, in its last December quarter, saw a 39% decrease in its Devices category, mostly Surface laptops. Last year, Microsoft announced the availability of an ARM-64 native version of its Windows 11 Office 365 products. What this means in practice, on a sizable scale, is currently unclear.
Of the “other businesses” that don’t add much to Intel’s bottom line, Intel Foundry Services (IFS) has yet to offer a strategic alternative to Intel’s traditional IDM (Integrated Devices Manufacturing) activities for PCs and servers. IFS is a big bet on a culture change, bowing to customer desires as opposed to telling them when the next x86 generation will ship and how much it will cost. Competition, customs chip makers, are well-trained, ingrained and demonstrably ruthless. To add to the IFS challenge, existing competitors can offer more advanced technology.
Let’s summarize Pat Gelsinger’s Struggles.
First, the PC business is and, in all likelihood will continue to be, in decline. This will add pressure on revenue, profit, and possibly cash. Notice that Intel just changed an internal accounting rule: “Effective January 2023, Intel increased the estimated useful life of certain production machinery and equipment from five years to eight years.” (Page 3 of the Earnings Release). This bit of accounting legerdemain, seemingly innocent but telling, strives to give an appearance of improved profitability. Not a great sign.
Second, the high-end high-margin DCAI server chips business is not in a declining market — au contraire, but it is, nonetheless, threatened by better devices that are built on more advanced technology non-Intel architectures manufactured by TSMC and others.
Third, keeping in mind the providers with better technology just mentioned, the Intel Foundry Services business is limited to customers such as the auto industry. This is a highly competitive, price-sensitive sector that’s dominated by lower-cost lower technology vendors. If Intel is to have a dog in this race, it has a lot of catching up to do. The “5 nodes in 4 years” effort promised by the CEO is good start, but will TSMC, Samsung, and others stop making progress in the meantime?
Fourth, other businesses such a Networking, Graphics, and Mobileye look like distractions. Last year, Intel floated a poorly received IPO for Mobileye. Perhaps it can cut the cord, and do the same for the other two product groups.
My professed high regard for Pat Gelsinger will grow even higher as he successfully deals with these challenges. No one wants to see Intel slide into another HP.
— jlg@gassee.com