Print the page
Increase font size

Hello, Mr. CHIPS

John Schroeter

Posted August 04, 2022

John Schroeter

Last week Congress finally funded the “Creating Helpful Incentives to Produce Semiconductors” (CHIPS) for America Act.

Helpful?

You bet it is. But not in the way you might think. The fact is the CHIPS Act is the biggest package of pork our government has ever foisted upon the American taxpayer. But they’ve managed to put a gloss of lipstick on this pig that would make Monika Lewinsky envious.

In short, the Act earmarks $52 billion for chipmakers to build semiconductor facilities in the U.S., which, we’re told, is essential to shoring up American supply chains and to countering China in the great semiconductor arms race. It’s made all the more urgent by the semiconductor shortage we’ve all been living through. So, we need to do something, right? Yeah, we do, but it’s not this. Let’s find out why—and see who wins, which, in the end, will explain quite a lot.

Our first clue is the list of key provisions the bill addresses. H.R.7178 – CHIPS for America Act – includes investment tax credits related to R&D, equipment purchases, and production; Federal grants to match state incentives; funding for the development of secure microelectronics supply chains; a national strategy on semiconductor research and development; skills training to build a capable workforce; and Department of Defense support for semiconductor technologies.

At first blush, this sounds reasonably reasonable. But the reality behind the gloss is enough to make an industry insider blush. Note that what’s not on the list is the one thing we actually could use: pure-play foundry capacity to serve for the long haul.

While it’s true that America’s share of the world’s chip-making capacity has fallen to as low as 10% from nearly 40% in 1990, the U.S. leads the world by far in equally (if not more) valuable areas such as chip design, electronic design automation (EDA), and chip-making equipment. Note also that seven of the world’s 10 biggest semiconductor companies are American companies. That didn’t happen by accident. A look at the dynamics of global economics provides some insight here.

For starters, why is most of the chip production sourced offshore? Could it be because it’s as much as 40% cheaper to make chips overseas? The fact is chips are low-margin commodity products. So, is the answer to this “imbalance” found in manufacturing them here, by American companies, where they will be FAR more expensive to produce? Even if we could compete on cost (we never will), don’t be fooled: if we build new plants in America and hire tens of thousands of people to staff them, where do you suppose that workforce will come from? Uh, China, Taiwan, South Korea. (If we build those fabs, they will come.) We’ve not cultivated that talent base in America. Do we really think we’re going to develop it here, natively?

Now, besides all this nonsense, chipmakers are busy expanding and diversifying capacity globally anyway. Take Samsung. The company plans to build a $17 billion factory in… Texas. TSMC has broken ground on a $12 billion facility in… Arizona. Both moves are intended to address supply chain issues and to improve logistics simply by locating manufacturing closer to their customers. It’s a model that makes sense for them. Indeed, TSMC tripled capital spending between 2019 and 2022, while Intel nearly doubled its capital spending in the same timeframe. All sans multi-billion-dollar boondoggles.

Consider also that American companies export billions of dollars’ worth of semiconductor products every year. It’s how we’ve come to own nearly 50% of the global semiconductor market. In fact, semiconductors are among the top-three manufactured U.S. export products, along with automobiles and aircraft.

Would you be surprised to learn that global semiconductor capacity actually increased between 2020 and 2021? Even without CHIPS Act funds to bloat its sails, it’ll grow another 9% this year. It’s beginning to look like we’ve got an over-capacity situation looming, especially as Asia’s collective massive investments pour more fuel on the fire. Investments that are sure to continue to benefit from American investments, as the last bit of drama in the life of the Act involved Chuck Schumer’s pulling of anti-China security measures. It’s true. American semiconductor companies are demanding such loopholes—to stay competitive!

But American interests aren’t the only ones lobbying for those CHIPS Act funds. Samsung is pushing for an equal shot at competing for state incentives, demanding that Washington “ensure that all qualifying companies should be able to compete for CHIPS Act incentives to pursue semiconductor projects in the U.S. on an even playing field, irrespective of their country of incorporation.”

And that’s because the Asian companies know that even after all this spending, we’ll still be dependent on Asian sources: $52 billion hardly dents the market requirements. In any case, many aspects of the industry will simply not be sufficiently re-shored. And yet, American politicians on both sides of the aisle (they all have pockets, after all, yearning to be filled by special interest dollars) are telling us that our national competitiveness will demand greater domestic capacity, and that government should manage these resources rather than leaving them to free market forces.

Well, let’s put this notion to the test with a handy historical reference. Remember SEMATECH? This was the brilliant organization that was formed to respond not to Chinese, but Japanese advances in the sector. At that time, the American government was somewhat more constrained: it squandered only half a billion on that particular porkfest. In the end, the free government money yielded not better chips or a stronger market position but wasted spending, as it is wont to produce, without contributing any value. Will it really be different this time?

Of course, the semiconductor companies will have their hands out for free money. After all, it’s in their shareholders’ interest for them to do so. They’ll even do a bit of public posturing to get their fair share. Intel CEO Pat Gelsinger, for example, famously held the company’s $20 billion new Ohio plant hostage to the funding of the Act. Moreover, he threatened to divert Intel’s investment dollars to Europe if he couldn’t get his hands on the CHIPS funding. And why not? Earlier this year, the EU’s version of the CHIPS Act allocated $7.3 billion to subsidizing Intel’s new fab in Germany. Micron and GlobalWafers have taken a page from Intel’s book and threatened likewise. What’s a politician to do?

Now, never mind that Intel has $39 billion sitting in the bank. And never mind that throwing government money has never built a semiconductor company. The SEMATECH intervention actually reduced domestic investment in the semiconductor sector! No doubt, CHIPS will outdo it. After all, it’s getting funded at 100X the SEMATECH level!

In the end, the $52 billion CHIPS Act is nothing more than 52 billion ways the American taxpayer gets screwed by filling corporate welfare slush funds—and the industry harmed for short-term corporate gains. The American semiconductor industry does not need public funds to outdo the Chinese. As it is we’re investing FAR more in R&D than the Chinese. It’s not even close. The private sector is doing just fine, thank you.

And then there are the claims that the Act will give a needed boost to economic competitiveness. Again, it’ll do just the opposite. Easy government money consistently undermines competitiveness. And then it makes things worse…

As we’ve stated previously, the chip shortage is about to come to a close. Which means that this $52 billion pig will kick in right when we’ll need it least. The timing of the Act will ensure that all that money is squandered in a misguided attempt to solve last year’s problem. Indeed, yesterday’s shortage is about to become transformed into tomorrow’s glut—and potential market weakness.

It is no secret that chip supply is cyclical. The Act will do nothing to assuage that fact of life. Now throw into the mix the fact that the aforementioned Samsung plant isn’t expected to become operational until the second half of 2024—in the midst of what will not only be a glut but a likely deepening recession. Or that Intel’s Ohio plant—if ultimately built—won’t come online until 2025.

Of course, Texas governor Abbott has been a major booster of the CHIPS Act. To put it graciously, like many politicians, he’s between the proverbial rock and a hard spot. It is a convoluted mess, for sure, with many competing and sometimes contradictory market dynamics at work. But don’t worry: the geniuses in Washington are sorting it all out for us.

But back to reality. No doubt, we’ll be telling a different tale next year, with a growth rate receding to the low single digits. Especially as increasing percentages of the household budget get consumed by rising food, energy, and housing costs. After all, there’s only so much discretionary spending to go around. It is, in fact, the consumer segment that will drag things down. Automotive, industrial, medical, IoT, and other sectors will continue to hold nicely. And therein lies yet another rub: it’s the consumer markets that are largely driving the smaller geometry nodes, e.g., 7nm and below—the very processes that the CHIPS Act is targeting. Note, however, that the industrial market applications (automotive, IoT, etc.) are dominated by nodes at 28nm and above! And these processes are booked solid. Such is the magic of your tax dollars at work!

As usual, our governmental leadership has got everything exactly backward. And as usual, they have no grasp of the nuances and interdependencies of a complex system that responds most optimally to market drivers, as opposed to politically motivated interventions. The externalities involved are massive. Consider, for example, that Taiwan was recently forced into a devil’s choice between supplying water for its rice growers or supplying the water demands of TSMC. Rice irrigation accounts for 70% of the total water consumption in Taiwan, where up to 10 million gallons per day are required to operate a large fab. TSMC is a good reference here: the company also consumes more than 7% of all of Taiwan’s electricity.

What’s more, building complex fabs is a serious undertaking fraught with risk. One analyst characterized it as fitting somewhere between a nuclear power plant and the International Space Station.

Most concerning, though, is the broad brush approach the CHIP Act is taking. It seems to assume that advanced nodes are all that matters. This couldn’t be further from the truth. All semiconductor products are not created equal. Certain processes applicable to one product class are utterly incompatible with those of another. The fact is a critical mass of semiconductor products are perfectly content—if not optimized—using far less exotic processes and equipment. What might be good for one is not good for all.

The economics of advanced fabs also seems to be ignored in the Act. David Fried of Lam Research (a maker of semiconductor processing equipment) observed, “The trend that we have seen lately is that fewer and fewer companies are able to monetize the value of the most advanced-scale technologies.” Indeed, as he points out, there are fewer customers at 5nm than there were at 7nm, and there were fewer at 7nm than at 10nm because a smaller number of companies can extract value from the large capital investments needed to develop these new products. “You are going to see that trend continue,” he concludes. “If you cannot capitalize financially on the value of scaling, be it power, performance, area, or yield, then you shouldn’t scale. This decision has to be made at the product level.”

So, in the end, what really matters with respect to American semiconductor superiority and security? intellectual property.

Note that six of the world’s top 10 most valuable companies are American tech giants that do not manufacture their own chips: Apple, Microsoft, Alphabet, Amazon, Tesla, and Meta all rely on TSMC for the lion’s share of their chip supply. In fact, more than 90% of the world’s most advanced chips come from Taiwan and Korea. A $52 billion investment is going to change that? Apple’s products might be “Designed in California” but for the most part, they’re “Assembled in China.” Does anyone give any thought to Apple’s weakness in the face of Chinese market aggression? No, not too much. That’s because intellectual property—creativity, innovation, knowledge, knowhow—is what ultimately leads to market dominance and economic power.

It was government meddling that got us into this CHIPS Act mess. More of it will not get us out of it.

The Controlled Demolition of American Prosperity

Posted August 03, 2022

By Jeffrey Tucker

That’s the ruling-class vision of our future: poor, foraging for food, out of gas, paying through the roof for electricity, unemployed and living off welfare, vaxxed up, and testing constantly for the presence of the Coronavirus. Or maybe Monkeypox.

Joe Manchin Needs Better Advice

Posted August 02, 2022

By Richard Vigilante

Republicans have not left Reagan behind. Just the opposite. They have gone all pre-Reagan.

How Inflation Changes Culture

Posted August 01, 2022

By Jeffrey Tucker

Here we are now with a preventable inflation pandemic and the realization that we have to learn to live with inflation. Soon we’ll realize that we have to live with recession at the same time.

Congress, the Fed, and the Great Scam

Posted July 29, 2022

By Jeffrey Tucker

We don’t even have a word yet for what might be coming. Depression is already used. How do you describe high inflation plus a manufactured depression?

In Praise of Failure

Posted July 28, 2022

By Jeffrey Tucker

The systems of safety built up over decades will not protect people for much longer. We are headed to a world that restores the natural order of things.

A Literal Depression Is Already Here

Posted July 27, 2022

By Jeffrey Tucker

For years now, we kept thinking things would get better, that someone would save us from the emerging hell. That is not happening.