Good morning,
I trust that everyone in the U.S. and Taiwan had a pleasant holiday yesterday, and that the rest of you enjoyed our absence from the Internet.
On Friday’s Dithering we covered Google’s Pixel Hardware event; on Sharp Tech we discussed Elon Musk and Twitter, streaming, and the validity of ad-supported business models.
On to the update:
The China Chip Ban
From Reuters:
The Biden administration published a sweeping set of export controls on Friday, including a measure to cut China off from certain semiconductor chips made anywhere in the world with U.S. equipment, vastly expanding its reach in its bid to slow Beijing’s technological and military advances. The rules, some of which take immediate effect, build on restrictions sent in letters this year to top toolmakers KLA Corp, Lam Research Corp and Applied Materials Inc, effectively requiring them to halt shipments of equipment to wholly Chinese-owned factories producing advanced logic chips.
The raft of measures could amount to the biggest shift in U.S. policy toward shipping technology to China since the 1990s. If effective, they could hobble China’s chip manufacturing industry by forcing American and foreign companies that use U.S. technology to cut off support for some of China’s leading factories and chip designers…
Earlier on Friday, the United States added China’s top memory chipmaker YMTC and 30 other Chinese entities to a list of companies that U.S. officials cannot inspect, ratcheting up tensions with Beijing and starting a 60 day-clock that could trigger much tougher penalties. Companies are added to the unverified list when U.S. authorities cannot complete on-site visits to determine if they can be trusted to receive sensitive U.S. technology, forcing U.S. suppliers to take greater care when shipping to them…if a government prevents U.S. officials from conducting site checks at companies placed on the unverified list, U.S. authorities will start the process for adding them to the entity list after 60 days.
This is a very big deal, albeit one that has been telegraphed for a while. I thought this paragraph in the New York Times write-up of the new rules was interesting:
Technology experts said the rules appeared to impose the broadest export controls issued in a decade. While similar to the Trump administration’s crackdown on the telecom giant Huawei, the new rules are far wider in scope, affecting dozens of Chinese firms. And unlike the Trump administration’s approach — which was viewed as aggressive but scattershot — the rules appear to establish a more comprehensive policy that will stop cutting-edge exports to a range of Chinese technology companies and cut off China’s nascent ability to produce advanced chips itself.
Huawei was a big deal, thanks in part to Huawei’s size and worldwide reach, but I think it was the Trump administration’s moves against ZTE for having violated Iran sanctions that were the real beginning of this story. I wrote at the time:
The other major impact of both the trade war and this action against ZTE will be the further bifurcation of Chinese and Western technology; already there is a major separation at the services layer: Facebook/WeChat, Google/Baidu, Amazon/Alibaba, etc. No, they aren’t perfect comparisons (particularly Amazon and Alibaba), but they are symbolic of how the Great Firewall has resulted in two Internets.
What seems likely to happen in the long run is a separation at the hardware layer as well; China is already investing heavily in chips, and this action will certainly spur the country to focus on the sort of relatively low-volume high-precision components that other countries like the U.S., Taiwan, and Japan specialize in (to date it has always made more sense for Chinese companies to focus on higher-volume lower-precision components). To catch up will certainly take time, but if this action harms ZTE as much as it seems it will I suspect the commitment will be even more significant than it already is.
To put this in broader context, China both benefited from and has been harmed by having grown its economy in a globalized world with mostly free trade. On one hand, the country has become the world’s manufacturer. That manufacturing, though, has long been biased towards labor-intensive work: originally this was because Chinese labor was cheap, but even once Chinese labor became expensive the entire ecosystem and skill base gave the country a big comparative advantage.
The implications of comparative advantage are the other hand: the rest of the world, particularly advanced economies like the U.S., Europe, Japan, and Taiwan focused on capital-intensive manufacturing. This all worked out well for the first two decades of the 21st century: China exported labor-intensive goods (like finished iPhones), and imported capital-intensive goods (like iPhone chips). The problem is that China is a bit like a body-builder who only works on their upper-body: the arms are super impressive but the legs are twigs. To that end ZTE was a flashing warning sign that the country needed to do some leg presses.
This increased motivation was one of the reasons many commentators argued the ZTE and Huawei moves were a mistake; Dan Wang made the astute point in a Stratechery interview that the ZTE and Huawei moves brought China’s private industry in line with government priorities:
In my view, the fundamental problem with Chinese industrial policy has been that this has mostly been a government-led affair in which the government counts on mostly government ministries as well as state-owned enterprises to be the buyers of obviously inferior Chinese technologies, and then hopes that the procurement process drives improvements…This time is different, and I think this time is different mostly because of US government actions…
So many of these Chinese leading companies are facing the prospect or the actual effects of US sanctions, and what the US sanctions have done is really aligned the interests of Chinese companies along with that of the Chinese state, to pursue self-sufficiency as well as technological greatness. So I’m now hearing from a bunch of smaller Chinese companies who could never get the time of day from a major company like Huawei, now they are getting and becoming qualified vendors because Huawei simply cannot buy from American technologies.
In my view, this is putting Huawei in the position that NASA was in the 1960s. When NASA and the US Air Force were the major buyers of semiconductors they more or less created the integrated circuit market. When NASA said, “Well, we will buy integrated circuits on the basis of performance not on cost. We need chips that can send three men to the moon and bring them safely back, we’ll pay whatever price for it.” Now Huawei today is a cash-rich organization, highly sophisticated technology company that is trying to bring up the capabilities of the Chinese ecosystem, not because it is trying to follow any government directives, but to save its own butt and maintain operations.
What wasn’t clear is if this approach would last past Trump; it was plausible that the Biden administration would reverse course on China, and that Chinese private industry would go back to buying U.S. chips and ignoring those smaller Chinese companies. After all, at the end of the day performance matters for high performance computing (it is in the name!), and U.S.-derived chips would always be the best option — if they were available, that is.
The Logic of the Ban
To that end, as late as this March, I was wondering if the U.S. would use access to advanced technology as a bargaining chip for increased Chinese cooperation on Russia specifically; it’s not clear if that was ever on the table, but what does seem clear is that China has chosen to remain relatively aligned with Russia. That perhaps influenced National Security Advisor Jake Sullivan’s comments last month that the Biden administration was going to fundamentally rethink the U.S. approach to technological exports:
On export controls, we have to revisit the longstanding premise of maintaining “relative” advantages over competitors in certain key technologies. We previously maintained a “sliding scale” approach that said we need to stay only a couple of generations ahead. That is not the strategic environment we are in today. Given the foundational nature of certain technologies, such as advanced logic and memory chips, we must maintain as large of a lead as possible.
Earlier this year, the United States and our allies and partners levied on Russia the most stringent technology restrictions ever imposed on a major economy. These measures have inflicted tremendous costs, forcing Russia to use chips from dishwashers in its military equipment. This has demonstrated that technology export controls can be more than just a preventative tool. If implemented in a way that is robust, durable, and comprehensive, they can be a new strategic asset in the U.S. and allied toolkit to impose costs on adversaries, and even over time degrade their battlefield capabilities.
This gets to the big takeaways from this announcement. First, in case there was any question, it is clear that China is being viewed as an adversary, and that that view is a bipartisan one. Any tech company with business in China would do well to note that any further investments are fraught with risk, and previous investments need to be diversified sooner rather than later.
Second is the point I started with: while Trump deserves credit for upsetting the apple cart in terms of conventional wisdom with regards to China relations, the Biden administration is correct to pursue those previous actions to their logical conclusion.
Go back to the bodybuilder analogy: before you can do leg presses you need weights. What this means in the case of chips is that manufacturing them is devilishly difficult: TSMC and Intel have access to the same equipment, but it is a fact that TSMC can build more advanced chips than Intel can. Said equipment, though, is itself incredibly difficult and complex to manufacture. In the case of China, its foundries, including SMIC in logic and YMTC in memory were figuring the manufacturing process out, albeit a few years behind their Western counterparts. Still, they were figuring out said process with U.S.-sourced (or, in the case of foreign suppliers like ASML, U.S. IP-infused) equipment.
To that end, what is significant about this move is not simply that it bans chips but it also bans equipment as well (and, given the restrictions it places on U.S.-persons, also bans the service of existing equipment). That certainly increases the motivation for China to build alternatives, but it is tough to get strong legs when you have to first figure out how to make weights (but the weights involve the most complex tools ever invented by humans). There is definitely room to argue that this is a bad idea in the long run, whenever China does catch up, but extending the restrictions in this way is at least a logical endpoint of the story that started with ZTE.
Chinese Retaliation, or Not
The obvious question is what China will do in response, and my suspicion is not much. The reason why can be gleaned from the answer Nvidia CEO Jensen Huang gave me when I asked about the Biden administration’s moves to ban Nvidia’s highest performing chips from China:
It’s excellent to know what the law is and it’s excellent to know what our limitations are. The limitations and the restrictions are very specific to a combination of computation level and multi-chip interconnection level. That restriction gives us plenty of envelope to go and run our business and for the vast majority of our customers in China, this is either unaffected because we have so many alternative products that they can use and they’re all architecturally compatible, many of them are socket compatible because we use industry standard sockets. So the ability for us to serve our customers in China with alternative products is quite acceptable and early response from all of our customers has been great.
For the customers that absolutely need that level of capability, we have to go for a license, that’s all, and it isn’t intended to keep Nvidia’s products from China, it isn’t intended to keep Chinese customers from those products. It’s intended to keep a specific use case of our products to be visible to the US government. I think the limitations and restrictions that is now known to us is an envelope that we can work with. For this particular quarter, we had to make a disclosure because it applied instantly, there was no grace period. So in our particular case, we’re right in the middle of the quarter and so our ability to execute and pivot was something that we wanted to make sure that people understood. But the vast majority of those customers that depend on our products will make sure that they had alternatives that are within the guidelines and the restrictions.
There are two takeaways embedded in this answer: first, the fact of the matter is that the capabilities provided by Nvidia chips in this case simply aren’t available elsewhere, particularly from domestic Chinese companies. This applies broadly. Second, China’s best advocates for gaining access to these capabilities are U.S. companies. This means it would be counterproductive to attack those companies in retaliation.
This principle also applies to non-chip-making sectors: hurting Apple, for example, means hurting the millions of Chinese employed in Apple manufacturing (either directly via Foxconn or throughout the Apple supply chain). More broadly, U.S. exports are a much larger part of the Chinese economy than Chinese imports are for the U.S. economy, which is why the Trump tariffs (which Biden has maintained) affected China more than it did the U.S.
These detentes are, admittedly, fragile; it’s plausible that China finds this naked exercise of U.S. technological might intolerable and retaliates anyways. These moves also reduce the cost of more extreme actions, like moving against Taiwan: while it is almost certainly unrealistic for China to gain control of TSMC via military action (fabs wouldn’t survive any sort of conflict), the cost of TSMC’s destruction for China is significantly reduced if China doesn’t have access to their chips anyways.
For now, I expect the industry to push on these new rules and find whatever loopholes exist to continue to supply Chinese companies. And, by the same token, it seems clear that more rules are coming over time. What is certain is the biggest takeaway of all: the free trade globalized world that China grew up in is gone, and the U.S. is determined to maintain the advantages it has for as long as possible.
This Update will be available as a podcast later today. To receive it in your podcast player, visit Stratechery.
The Stratechery Update is intended for a single recipient, but occasional forwarding is totally fine! If you would like to order multiple subscriptions for your team with a group discount (minimum 5), please contact me directly.
Thanks for being a subscriber, and have a great day!