Contours of a US-China Tech Grand Bargain

The first (known) phone call between Trump and Xi since the inauguration finally happened this week. Soon after the call, the new round of negotiation will take place in London on Monday, June 9th. A new wrinkle is added to the negotiation dynamic – Commerce Secretary Howard Lutnick is now part of the US negotiation team, joining Treasury Secretary Scott Bessent and USTR Jamieson Greer, who alone represented the US in Geneva last month

With Lutnick’s addition, smart observers have noted that technology restrictions are now firmly on the trade negotiating table, a significant shift from before. That’s because the Commerce Department is in charge of trade and technology, from semiconductor export control to AI deployment. Its policy portfolio also includes less sexy but no less important areas like IP patents and anti-dumping of steel, aluminum, and other raw materials.

It is no longer inconceivable that a US-China “tech grand bargain” of sorts may materialize to bring about a new steady state. I have been ruminating about the possibility of a grand bargain scenario since Trump was re-elected last November given how he talks about Xi and China. I have noted as much in my 2024 annual letter, though it felt too speculative back then. It is looking less speculative now – no longer a total waste of time to spill some digital ink on it.

In this post, I will give it my best effort to neutrally and analytically describe the contours of what I think a tech grand bargain between these two superpowers may look like – what will be on the table, what would each side want from each other, how might each item play out, and why. I’m approaching this exercise with the assumption that reducing the massive bilateral trade deficit in goods – the core imbalance that animated “Liberation Day” – is still the main goal of the US.

I’m sticking with tech and tech-adjacent issues only, broken down into four different buckets, because those are the issues I know best. Of course, a comprehensive bilateral grand bargain would include many other non-tech issues, from a steady tariff rate on all goods and student visas, to military exchange and even Taiwan. But technology is becoming by far the largest point of friction and concern between the two countries. Even the perennial Taiwan issue is now often framed within the “'chip war”, so addressing all the tech issues (somehow) may pave the way for other issues too.

Alright, let’s get to it!

Supply Chain Tech and Critical Materials

Advantage: China.

What does the US want and need from China, beyond everyday consumer goods, that China ultimately doesn’t mind selling but sees them as points of leverage? Supply chain goods and services up and down the advanced manufacturing stack.

Rare earth is the latest example, where a slower than expected outflow from China pushed Trump to call Xi. US automakers are facing factory shutdowns without a steady stream of Chinese rare earth imports. The situation became so dire that these automakers are legitimately considering setting up more operations in China to get their hands on more rare earth materials in order to keep their facilities in America humming – turning onshoring of American manufacturing on its head. China has since eased export restrictions for the three major American automakers – GM, Ford, Stellantis (Jeep) – since the Trump-Xi call, but it is illustrative of what the US depends on China and what China ultimately does not mind providing when the condition is most optimal.

There is a wide range of both materials and technologies that go into this bucket. Batteries, parts, and even know-hows are things that US automakers have made clear they want and their Chinese counterparts have also made clear they would like to sell. A deal that opens the door for IP licensing or knowledge transfer in these areas could be a “have your cake and eat it too” scenario. Not only do both sides have a willingness to deal on this issue, for the US side in particular, it would magically not exacerbate the trade deficit in terms of goods, because IP transfer typically would be considered “services”.

Ford has already struck a licensing deal with CATL in the past. Chinese companies and the government have shown willingness to include technology transfer into negotiations with the EU to reduce EV tariffs – a separate negotiation that is gaining momentum. US pharmaceutical companies are also licensing more from Chinese biotech firms, which have been starved by lack of VC funding and an anemic capital market. A stable technology licensing and transfer flow that won’t accidentally trip national security or political wires would be a rare win-win outcome.

There is an additional scenario where Chinese EV makers participate in joint ventures in the US to make and sell EVs directly to American consumers, similar to what foreign automakers had to do in China to gain access to the consumer market there decades ago. A BYD-GM plant in Georgia, a Xiaomi-Ford plant in Tennessee, you get the picture. It would create jobs, bring more competitive products to the American consumer market (to Tesla’s chagrin), and give Trump some big headline investment numbers to tout. It would also anger a lot of people in DC who find all things from China dangerous and toxic, so the political resistance would be fierce. Thus, this JV scenario is less likely than the IP licensing tech transfer scenario.

Semiconductor Manufacturing Equipments

Advantage: America.

Imposing export control on advanced semiconductor manufacturing equipment, and convincing allies like Japan and the Netherlands to do the same with their leading companies, has been one of the most effective policies in slowing down China’s homegrown semiconductor progress. With everything on the table, China will want this “chokehold” loosened the most during the London negotiation.

By the same token, the US side has most of the cards here. The new restrictions on EDA (electronic design automation) software that powers SMEs, which occurred after the Geneva Accord, are the latest expression of that leverage. Unlike China’s leverage in rare earth, however, the US government does not want to sell SMEs to China, as it represents one of the few remaining strategic edges that the US can maintain in its intensifying technology rivalry with China. (Whether you think such rivalry is valid or worthwhile is a separate question.)

China’s indigenous semiconductor manufacturing equipment is making progress, but the current gap is still wide and discernible enough that access to American SMEs would be a coveted outcome. A large SMEs purchase promise from China may be proposed and could be presented as a “fix” to the trade deficit in one fell swoop. These are very expensive equipment after all. But the strategic value of these SMEs are likely too high for the American negotiators to budge or they would demand too much in return from the Chinese side to reach an agreement. Therefore, I anticipate no significant change in this area from the status quo, despite it being on the negotiation table.

The wild card scenario outside the confines of the bilateral negotiation is if Japan and the Netherlands are no longer standing shoulder-to-shoulder with the Americans. This schism could happen because Trump has lost so much standing in Europe given his administration hostility towards the EU. Similarly, the US-Japan negotiation is going rather poorly at the moment with the same trio of American negotiators openly disagreeing in front of the Japanese.

If the multilateral aspect of SMEs restrictions toward China fall apart, where Tokyo Electron and ASML are being peeled off and becoming less restricted, then the American side may push an increasingly popular narrative that even for China, building on American tech is better than not.

Obviously, this narrative originated from and is most forcefully argued by Nvidia. Speaking of Nvidia…

AI Chips

Advantage: America.

The future of US export control on advanced GPUs to China may be the most high-profile item on the negotiating table. It is unthinkable that this issue would ever be on the table, given how both Biden and Trump’s teams have placed increasingly more stringent restrictions. But given the Trump administration’s new approach to AI chips exports – get rid of the Biden era AI diffusion framework and fold chips access into bilateral trade negotiations – plus Nvidia’s forceful lobbying both behind closed doors and in public, it is likely on the table with China as well, at least as a discussion item. (It is perhaps not a complete coincidence that London was chosen as the negotiation location, where Jensen Huang is also due to be in The City on the same day for London Tech Week!)

Similar to the SMEs bucket, American AI chips, specifically Nvidia ones, still have clear advantages over alternative Chinese products. But what is different, and important to emphasize, is that the Chinese side may no longer care for an offer to buy more modified Nvidia chips, nor does the US side really want to sell more Nvidia chips of any type to China.

China knows that even in a best case scenario, what it will get will be modified, less-performant products from Nvidia and AMD, but the supply will always be unreliable and subject to US policy whims. Simultaneously, Huawei and other domestic alternatives’ capabilities, though still lagging behind, are catching up fast enough that the country is feeling more confident about its own products. As I have written before, the Chinese government is not AGI-pilled enough to believe that generative AI is the be-all and end-all. It would rather focus on all the real world applications of AI in robotics, manufacturing, self-driving, supply chain management – all the things the country was already good at before ChatGPT came along. On the other hand, the US government, especially since Trump took over, is more AGI-pilled. This difference in attitude towards AGI creates a weird alignment between the US and China: neither side really cares to buy/sell jerry-rigged Nvidia chips!

In other words, the only party that wants more Nvidia chips to flow to China is Nvidia, and perhaps the usual list of Chinese big tech firms – Alibaba, Tencent, ByteDance, Baidu – all of which are trying to stay competitive in AI globally despite all the geopolitical headwinds. Even then, Chinese tech players are all hedging towards more reliable domestic options, as well as building their own custom ASICs to reduce reliance on Nvidia.

Could an opening still materialize out of London for more American AI chips to be sold to China? Unlikely. If it does happen, it will be because of Jensen Huang’s superhuman persuasion and lobbying ability rising to the occasion.

Consumer Tech Market Access

Advantage: China

Shortly after the Trump-Xi call, reports indicate that Trump plans to extend the TikTok ban deadline a third time, even though the law only allows for one extension. So whatever the shape that a tech grand bargain may take, if any at all, will definitely include a TikTok deal.

However, a TikTok deal is less meaningful on its own. The Chinese government never cared that much for ByteDance, the company, nor does it consider it of much strategic value. Its attitude towards TikTok is about the same as any concerned American parent grimacing at their scroll-addicted child. Otherwise, Zhang YIming would not be allowed to live peacefully in Singapore and pursue AGI on his own. (By contrast, a Robin Zeng or Wang Chuanfu relocating abroad would be unthinkable). It is not letting TikTok go so easily, because of the arbitrariness of the ban and the precedent that could set for other Chinese companies operating in the US. Trump’s obsession with saving TikTok is a welcoming surprise.

So the most meaningful outcome of a TikTok deal is actually a template of how to do business in the US. What would be an acceptable size of foreign ownership stake? What are the rules of the game for data governance and algorithmic control? Which cloud platform is the whitelisted and politically palatable infrastructure partner? (Probably Oracle)

Such a template would clarify things for Shein, Temu, and many of the Chinese EV brands that still aspire to take a crack at the US market. Chinese biotech firms are another industry looking for US market access. All of them would still face perception issues. But a TikTok deal would at least provide a roadmap of what to do when it comes to technical infrastructure, policy compliance, and corporate governance.

One of Trump’s most consistent views on China and on trade in general is the notions of reciprocity and fair market access. If China wants the US to open up its market for Chinese EV makers, what has to happen in return is more opening up or less restrictions on US-made goods flowing into China. However, after each respective countries’ goods make it to the other side, staying competitive against the domestic market’s alternatives would be a different story.

In a hypothetical “soybeans for EVs” swap, it is anybody’s guess that if the fight is fair, would Xiaomi or BYD win more American consumers’ hearts than Tesla and Ford? On the flip side, would Iowan soybeans win more Chinese consumers’ stomach than Brazilian soybeans?

As a believer of free market capitalism (with national security caveats), it is a fight I would rather see than the fight we currently have. As far as present day advantage is concerned, consumer product dominance tends to eventually go to the country that makes more stuff. That is clearly China, which is also how we ended up with this trade deficit and a trade war in the first place.

The London negotiation may be looming large, but I doubt it will be the last negotiation. Whatever the outcome may be next week, more haggling will happen, as they always do, all the way until the minute before Trump and Xi sit down to personally sign the deal.

When that deal is sealed, perhaps all or perhaps none of the tech issues described in the contours of a tech grand bargain will be included. But there is clearly more on the table than ever before, and more at stake.