Intel and Qualcomm: This ‘odd’ couple won’t solve US chipmaking problems

Intel has been the best at this for decades. Then, around 2018, its leadership began to crumble, thanks to a series of missteps. TSMC, an upstart from Taiwan, has steadily won and is now arguably the world’s best chip maker.

Veteran Bloomberg chip reporter Ian King wrote about this in late 2018. Since then, Intel has declined and TSMC has risen. Intel is now worth less than $100 billion. It doesn’t even crack the top 150 companies by that measure. TSMC is worth nearly $1 trillion these days, putting it in the top 10.

Why countries must have fabs now

This is a stunning fall from grace for Intel and a huge strategic and geopolitical problem for the US.

If you want the best chips, you have to go to Taiwan to make them. Or maybe South Korea, where Samsung has built an impressive so-called foundry business that makes semiconductors for other companies.

Many of the famous “chip makers” we think of in the US today it doesn’t make chips. Nvidia, Qualcomm, AMD and all the others design chips, then usually have TSMC manufacture them. Apple and a host of other big tech giants also have TSMC create the chips they design in-house.

Again, actually producing these complex products on a massive scale without tiny imperfections is incredibly difficult.

If China invades Taiwan or somehow exerts much more control over this island nation, it will be a disaster for the US and Europe for that matter. No more fast iPhones for a while. Nvidia should look elsewhere to make their GPUs. AI progress could stall.

This is why, today, “real countries have fabs” is a better phrase. Chips power the modern economy. If you have to source these components overseas, you are vulnerable.

Qualcomm won’t fix Intel’s worrying decline

This is also why Intel’s decline is so worrying. It is the only company in the US that knows how to produce powerful chips at scale. (Globalfoundries also makes semiconductors, but not the latest ones).

The WSJ reported Friday that Qualcomm recently approached Intel about a deal. This followed a report from Reuters two weeks ago that Qualcomm was considering buying Intel parts.

If a deal materializes, it won’t solve America’s chip manufacturing problem. One analyst called a potential tie-up “odd.”

Qualcomm probably isn’t interested in Intel’s manufacturing operations. As reported by Reuters, she is passionate about some of the chip design operations.

Intel has 2 main businesses. In one, it designs semiconductors for PCs, data center servers and other uses. The other party produces these models.

For decades, Intel’s design and manufacturing operations were tightly integrated. This has worked great for ages. The company could set up its factories to the exact specifications of its in-house chip designers.

Then the world began to move to a different approach, pioneered by TSMC. Instead of designing and making your own chips, why not run factories and make chips for other companies?

In the late 1980s, when TSMC started, this idea was laughed at. (Hence the sexist comment from AMD’s Sanders about real men owning fabs).

But TSMC’s approach gradually caught on, helped by Intel’s mistakes and the march of modern technology.

The biggest happened when Intel missed out on making chips for the iPhone when that revolutionary device first came out. Apple ended up going with TSMC. Qualcomm is also a big smartphone chip designer and has TSMC making most of these components. Other chip designers, including AMD, have also begun turning to TSMC.

This gave TSMC the huge and diverse production volumes it needed to learn how to make better chips than anyone else. Ian King’s 2018 article describes this virtuous feedback loop well:

With billions of transistors on chips, a problem with a small number of those tiny switches can render the entire component useless. Production can take up to six months and involves hundreds of steps that require obsessive attention to detail. Every time a mistake occurs, the plant operator has a chance to make adjustments and try a new approach. If the change works, that information is kept to try the next challenge. The more productions running, the better. And TSMC has the most these days.

While TSMC was learning from this wide variety of huge customers, Intel’s manufacturing operation was stuck with just one customer: itself.

As smartphone chips became the biggest game in town, Intel simply didn’t have the volume to keep up with TSMC in the manufacturing race. AI has made this situation worse. Nvidia leads by a mile here, and TSMC manufactures its GPUs.

A miasma of stubborn production

Raising the manufacturing miasma around Intel will be an expensive, risky and complex endeavor.

Intel even started paying TSMC to make some of its chips. With that as a starting point, there is a very long way to go.

The American company took a major step recently when it separated its foundry business from its chip design business. This makes it easier for external customers to trust Intel to manufacture their chips without competing with them.

The next challenge is the biggest: Actually being very good at making chips again.

Intel’s foundry business won’t really be able to challenge TSMC until it gets some big customers. Again, to get good at making chips, you need huge and diverse volumes so you can identify defects, fix processes, and feed that knowledge back into your factories.

It’s a chicken and egg situation. Without high volumes, external customers will be wary of having Intel make their precious chip designs. But without customers, Intel cannot improve.

Does everyone love Raimondo?

One way to break this deadlock: get the US government to convince other companies to use Intel factories. That’s exactly what’s happening now, according to CNBC.

Commerce Secretary Gina Raimondo has tried to urge shareholders of companies like Nvidia and Apple to recognize the economic benefits of a US foundry that can produce AI chips, CNBC recently reported.

About 4 days later, Intel announced a partnership to make an AI chip for Amazon Web Services. Intel shares jumped on the news because AWS is the largest cloud provider and designs a lot of chips for use in its huge data centers. That’s the kind of volume Intel needs.

18 To hope

From a technological point of view, Intel has a new process node, called 18A. This is a set of chip design rules and an accompanying manufacturing system. If all goes well in the coming years, this could make Intel more competitive with TSMC’s core nodes again.

The AWS partnership is based on this 18A technology, and Microsoft said earlier this year that it will also have an in-house chip built on this process node.

Intel’s foundry business just needs a lot more of these customers. And the 18A technology has to work really well for that to happen.

“18A had better be spectacular as it bets the company on its success,” Bernstein chip analyst Stacy Rasgon wrote in a recent note to investors.

Would a split work?

Meanwhile, Qualcomm is unlikely to want to acquire this part of Intel.

This raises the specter of Intel being destroyed, something that has been speculated about for the past few months. Reuters said Qualcomm was interested in some of its chip design operations, while the WSJ reported on Friday that Qualcomm would likely sell some parts of Intel to other buyers.

How would the Intel Foundry business work as a separate company, separate from the design parts?

The problem here, again, is volume. Right now, Intel’s foundry business needs the production volume it gets from the company’s own chip designs. Without that, they would have very little production to learn from.

“We don’t see how that could really be done now,” Rasgon wrote in a note in early September. “The manufacturing arm cannot stand on its own now given the heavy losses and lack of scale.”

“A spin-off only really makes sense once it attracts substantial third-party business, which seems years away (if ever),” he added.