Why Is Biden Restricting More Chip Tech To China? It’s Complicated.

The Biden administration plans to broaden curbs on U.S. shipments of some advanced semiconductor chips and some of the most sophisticated chipmaking equipment to China, Reuters recently reported.

The goal presumably is to slow the development and deployment of the most advanced computing and chip production technologies by Chinese companies. This is a very complex issue and the eventual outcome is not obvious. But here are three questions to ask.

1. Will export curbs be successful?

From the perspective of the White House, export controls are a useful tool to protect the the country’s technological advantages, especially against undesired military uses. But the key question is how hard will it be for Chinese companies to develop alternatives. Ultimately this comes down to whether there is some technological chokepoint that has to be overcome, or are their alternative pathways that can get you there?

In the case of the super expensive ASML lithography machines that are needed to make the most sophisticated digital logic chips, ASML and its supplier Zeiss spent decades learning how to make the complex optical system and extreme UV light source, and tie the whole thing together. Though there could be alternative routes, the problem is so demanding that it would take a comparable scale of investment and time. Cutting off access to Chinese companies will likely hobble them for some time. Access to leading edge lithography tools have long been restricted, hence companies like Shanghai-based Semiconductor Manufacturing International Company (SMIC) have consistently struggled to keep up with the latest chip generations.

In contrast, if there are alternative pathways that can take you there, export restrictions end up channeling investment towards them. A 2007 study conducted by the U.S. Department of Commerce Bureau of Industry and Security (BIS) together with the U.S. Air Force Research Laboratory examined the impact of export controls on the American space industry. Evaluating the impact of International Traffic in Arms Regulations (ITAR) export controls, it found that the U.S. share of global satellite manufacturing revenues dropped from 63% in 1996-1998 to 41% from 2002-2005 as a result of these controls. The report concluded that “ITAR has impacted U.S. competitiveness by encouraging other nations, in many cases our allies, to develop indigenous space capabilities and industries that now market globally.”

Many people in Washington assume that American companies enjoy an overwhelming technological lead and that countries on the receiving end of restrictions don’t have the R&D wherewithal to enable them to develop alternatives. That might have been the case in the late 1980s and 1990s, but as we know many countries including China have developed impressive capabilities in many fields. The U.S. really needs to avoid unintended consequences like damaging domestic firms who today are global leaders.

2. How will new or additional restrictions affect manufacturers?

For tool manufacturers like Applied MaterialsAMAT
or Lam ResearchLRCX
, new licensing requirements on exports will make their job more difficult. It will also foster increased investment by the Chinese government in alternatives, which means potential future competition. China will work very hard to circumvent attempts to hold them back, and is already investing across a broad front to develop domestic capabilities. The Chinese market for semiconductor tools is enormous, thanks to the breakneck levels of investment in new fabs. Companies like Applied and Lam benefit greatly from selling into this market. It is more than just the positive impact on their bottom lines, as many in Washington may proclaim. Selling there gives them increased scale efficiencies and the profits necessary to invest in future technologies. Shrinking the size of their available markets risks hurting these companies over the longer term.

Tool manufacturers will likely also tell you that they benefit considerably by working closely with Chinese (and all) customers. This is because they can see and understand what those customers are working on. My sense is that this facet is widely under-appreciated.

Chip companies like AMD and Nvidia face a somewhat different dilemma. Restricting their sales will detract from the scale advantages they gain from the size of that market, and it will foster the growth of domestic Chinese alternatives. We should not assume that China is a developing country that lacks the knowhow to create innovative designs. Quite the contrary, there are a lot of smart engineers there who know how to use the latest electronic design automation tools and can and will design alternative chips.

3. Is this a sustainable strategy over the longer term?

The key to effective export controls is understanding where the U.S. might possess technological advantages, how we arrived at them, and how we sustain them. More often than not, this arises from the foundation provided by huge past investments in basic scientific research and deep understanding and tacit knowledge gained from them that are not easily reproduced. It also means working closely with allies and coordinating policy on this front.

Ultimately the way to win this kind of race is to run faster. As a chemist by training, I often suggest to my more technically-inclined colleagues that when the Universe cools to four degrees Kelvin, all gross margins will trend to zero. New innovations, in this model, are like hot spots that naturally tend to cool (I usually cite the Boltzmann Equation for this, but I won’t trouble you with the details here). You can wrap that innovation in blankets to keep it warm and slow down the cooling, but eventually all that knowledge seeps out and the innovation commoditizes. This is just another way of saying that almost all great proprietary innovations eventually become commodities, and the only path forward is to keep inventing the new. In the end, that is the great hope for the U.S., that we can keep doing that.