Breaking up Intel could be the end of American dominance

Intel is No. 1 on the list of biggest semiconductor companies of 2020 and has been a pillar of technology growth over the past five decades. Although it has fallen on some hard times, it’s not time to count the company out.

Intel is not just another technology company. It is one of the reasons for the United States’ world dominance, both militarily and economically. Its strength is in manufacturing and it needs to double-down to create an advantage over the competition. While Intel has enjoyed successes and endured failures (even those that were self-inflicted), there is no mistake that the world still runs on Intel.

Breaking up will lead to a situation where American semiconductor manufacturing capability nearly vanishes, and it’s a path of no return for not just Intel but the American industry as a whole. It was 2000 when the clamor for breaking up another American icon started, and just two decades post-breakup hardly anything exists of what we knew as Motorola.

What happened to Intel?

What caused Intel to lose momentum? Simply, the market changed and Intel did not. Intel rode the wave of the electronics industry’s exponential growth. This growth came with evolving needs and Intel found itself creating chips that could do everything. It was acceptable when no one else could do what Intel did. But Intel’s customers have evolved. In mid-2008, 75 percent of Intel’s server chip revenue came from three customers: Dell, HP, and IBM. By late 2012 it had evolved to eight [Wired]. Today those additional customers have become so big they can afford to have their own chips. Intel is a raw material provider and a very expensive one at that. And now, Intel finds itself at an impasse, because customers have figured out how to make the chips themselves to avoid the cost of the merchant—at least some pieces of it.

We need to understand what Intel is before we can diagnose the problem and suggest possible prescriptions. Intel = x86 architecture + manufacturing competency. The rise of ARM - x86 has its limitations against RISC (ARM), power being one. Intel lost the low-power mobility race by selling Xscale to Marvell (a decision that can be classified as one of the biggest blunders in Intel’s history) which in turn opened space from Qualcomm and MediaTek and now it is hurting Intel in the PC (Apple) and possibly server (Amazon Nitro) space too.

Intel’s prowess was in its manufacturing abilities and over time it has lost this advantage to TSMC—and possibly Samsung. The problems began during the launch of 14nm processes. Initially these were small delays, but they were accentuated with 10nm. Even today, Intel’s 10nm is considered equivalent to TSMC’s 7nm performance (various factors go into performance, including density/pitch, quality of transistors, etc.). Intel is behind, but not by a lot. It is possible to catch up with the right execution.

It only takes a quick scan of layoff.com to understand that not all is hunky-dory with employees. We can’t forget that at its core, what Intel does is science, and in order to achieve groundbreaking technologies, you need motivated scientists and engineers. 

What not to do: A lesson learned from Motorola

In any other industry, the “breaking-up” logic may make sense, but not for the semiconductor industry. Intel is already handicapped by a lack of usage data that Microsoft, Amazon, and Google have. Breaking it up further will limit the levers that it has at its disposal. While breaking up Intel might seem like an easy solution, the temptation should be resisted.

Let’s take a page out of history. The last example of a major vertically integrated company that was challenged to “unlock value” was Motorola—a company that can be credited with creating the electronics industry. It was a giant, just like Intel. But it didn’t even last a decade after the “unlocking” started. It looked nice in the beginning. It started off with divesting its semiconductor business in 2004. It was a great success story, or that is what it seemed at that time. Blackstone even ponied up $17 billion to take it private. It was valued at $4 billion in 2011—$13 billion destruction in just seven years. Freescale did not have Motorola, and once divested, the tribal leaders at Motorola had nothing to do with Freescale. It had found new cool friends in Intel, Texas Instruments, and Qualcomm. Except for Qualcomm, none play in mobility anymore. In 2015, Freescale merged with NXP, only to become a shell of itself. Freescale was the new name for the company that previously challenged Intel for dominance in the personal computer space with its PowerPCs. It made semiconductors that ushered the mobile revolution. Alas, the “unlocking” never happened. Motorola mobility died, while Samsung and Apple became the leaders of today by following the same strategy that Motorola was hounded to abandon in early 2000s. Looking back, maybe Motorola was just two decades ahead of the marketplace and its conventional wisdom.

How does Intel move forward?

It’s a new year, with a new beginning. There is no silver bullet, but rather very deliberate steps that begin with accepting that the market has changed.

Here are some key strategies Intel can focus on:

1. Focus on its biggest asset, manufacturing. It should have a renewed focus on leadership in manufacturing, and open Intel’s fabs to non-x86 architecture manufacturing. It will require discipline. This enables Intel to better utilize its current capacity, while developing newer nodes (7, 5, and 3nm). Intel has technologies that are better than its two leading competitors—be it quality of transistors, density, or stacking (chiplets).
2. Invest, acquire, or both. Focus M&A activities around improving manufacturing capabilities, acquire or invest in companies that can help Intel gain edge, and get back into the lead position
3. Don’t lose the seat at the table. Intel had an aversion to ARM, and the boat sailed with NVIDIA buying ARM. Intel could still make a play with RISC-V. It’s a low-investment opportunity to have an alternative. It is better that companies such as Microsoft, Google, and Amazon build chips on an architecture that Intel can at least have a say in.
4. Think beyond the trillion-dollar goliaths. No more should Intel rely solely upon supplying Xeon (it’s money-making product) to the Amazons, Microsofts, Googles, Facebooks, Dells, and HPEs of the world. These cloud companies will make their own chips because they can afford to; they have even stopped buying machines from OEMs. Intel needs to identify areas and companies that have something to contribute to the industry and still need Intel.
5. Enlist help. Given the importance of Intel to national security and economics, Intel should work with the United States government to invest in R&D and associated technologies to make semiconductors an issue of national importance. Not only does this allow Intel to reduce its own cost, but also puts Intel back in the driver’s seat for all things semiconductor in the Western market—and possibly globally.

?Intel’s prescription

Intel has been the flagship for innovation, but it’s time to get to work. Intel must have a renewed focus on “getting stuff done” and finding people to put in lead roles who care deeply about the end goal. What Intel does is science, and scientists cannot be run by KPIs—they must be motivated to innovate.

It can take a page from hard disk makers and sell compute modules versus processors. This reduced the cost of doing business for companies that are not hardcore technology companies and will enable Intel to control more of the value chain. And while Intel has delved into it with its Edison and Galileo boards, it’s time to take the full leap.

What’s more: outsourcing its design and manufacturing cannot be the solution for long run. And while this may be needed in the short run, it will not solve Intel’s long-term objectives. The world is evolving, and so too must Intel. It’s time to rekindle the culture that made Intel the technology leader that kids aspired to work for in college. This is an achievable task and totally doable.

Disclaimer: this write-up represents my personal views and doesn’t necessarily represent the views of Kearney. In fact, the strength of our partnership lies in our ability to respect one another’s opinions—and this one is solely mine.