CHINA: The View from Washington

Earlier this week we attended a closed-door conference in Washington, D.C. that convened leading China experts and members of the policy community. The discussions were held under Chatham House rule, so we are not revealing the identity or affiliation of specific speakers. Below are our key takeaways for investors.

US-CHINA OUTLOOK

• U.S. officials see Trump’s visit to China as leaving the two sides in a much better position to manage frictions in the relationship. Communication is improving on economic issues, including through a Board of Trade and Board of Investment, and a new dialogue structure around AI guardrails (more below). Xi Jinping’s visit to the U.S. in September will extend this period of regular dialogue.
• Expert opinion is divided on whether the US-China relationship has truly entered a new and sustainable phase of “constructive strategic stability” (Xi’s framing, endorsed by Trump). One view posits that China’s military and techno/industrial prowess (including dominance of rare earths) – and U.S. recognition of these capabilities – provides a durable basis for the two sides to lower tensions. However, some experts worry that this is not a politically sustainable equilibrium, and that U.S. policy could snap back to a more contentious footing (as soon as after U.S. midterms). Related to this concern is a view that Beijing is highly confident in China’s long-term advantages over the United States and unlikely to make major accommodations to Washington.
• Our Take: We have argued that mistrust between the two sides, and the structural dynamic of strategic competition, will put continual pressure on the relationship. However, our base case is that the two sides avoid sharp re-escalation in 2026, and quite possibly 2027, as it would pose economic and political risks for each leader.

AI COMPETITION AND GUARDRAILS

There is a consensus that the US and China each have advantages across the AI tech stack. Using Nvidia’s metaphor of a “five-layer cake”:

• Energy: Advantage China.
• Chips: Advantage U.S. and its allied partners (given the central role of Netherlands’ ASML lithography and Taiwan’s TSMC).
• Infrastructure (cooling, networking, etc.): Both competitive.
• Models: Advantage U.S., though with China able to remain close (especially in building cost-effective models).
• Applications: China has the overall advantage in the pace of AI diffusion (especially to industrial applications) but with the U.S. also pushing aggressively in coding and related fields.
• As one expert warned, U.S. officials/investors should not forget that many of the “basic ingredients” of this cake also rely on supply chains from China, including critical minerals used in chips and chip-making tools as well as transformers and other power equipment.

The debate on what this means in practical and policy terms tends to break down into two main camps:

• A more hawkish camp sees US-China AI competition as defined by access to the most advanced AI models, and that capacity as dependent on high-density GPUs. In this view, the US has a sustainable lead that can be preserved so long as the US maintains tight controls on semiconductor technology exports to China.
• A rival camp argues that a clear U.S. lead is not sustainable, especially if competition is defined by AI diffusion rather than the most advanced models. This camp believes that China can continue to train and deploy low-cost models that are “good enough” for most applications, using clever software and hardware engineering and access to low-cost electricity. This camp sees U.S. export controls as hurting U.S. industry through lost sales, while also incentivizing China to develop an independent AI technology stack.
• Note that while those terms of debate are mainly geopolitical, the investment thesis behind the U.S. AI/data center buildout looks shakier if the second camp turns out to be right.

“Hawks” and “doves” both welcome a U.S.-China agreement to start a dialogue on AI safety guardrails but have limited expectations for outcomes:

• U.S. officials see both sides as having incentives to manage the safety risks around advanced AI models, particularly the ability of non-state actors to use AI to manufacture dangerous biological compounds. U.S. officials believe that U.S. leadership in AI, as demonstrated by Anthropic’s Mythos model, is key to persuading Beijing to engage.
• There is skepticism among experts as to whether Chinese national security/tech agencies will be fully committed to this dialogue, and whether a bilateral, as opposed to a broader grouping of countries is the most effective format.

Our Take: US-China competition in AI is a key geopolitical theme and one that investors should not ignore. However, we would also not lose sight of the extent to which these ecosystems will remain linked – including through the Chinese firms participating in the global data center buildout (see recent coverage HERE).

ECONOMIC OUTLOOK AND TRADE TENSIONS

• Most observers are pessimistic about the prospect of China’s growth pattern shifting anytime soon. That is, the expectation is that domestic demand will stay weak relative to production and investment, resulting in continued excess capacity, deflationary pressure, and a large trade surplus.
• While rhetorical support for consumption has picked up, experts see ideological as well as structural reasons to doubt that Beijing will aggressively pursue rebalancing. Beijing views technological upgrading as key to productivity growth as well as geopolitical resilience, but this agenda will mean a continued pattern of overinvestment (and low returns on investment). In the meantime, consumption is held back by negative wealth effects from property and a very weak labor market.
• In a comment that struck us as perceptive, one observer pushed back on the notion that China has a dual-track economy with a highly efficient tech sector and low productivity in traditional industry. Both high-tech and low-tech sectors are subject to distortions in the allocation of capital, as overcapacity in advanced sectors such as electric vehicles demonstrates.
• European governments are increasingly outspoken about the threat to industry from China’s exports, especially in autos. However, experts – including those from Europe – are skeptical that EU member states will soon reach the consensus and political will to push back aggressively. One dilemma – especially for German industry – is the attractiveness of producing in China to export to the rest of the world. See HERE for a recent take from Jacob Kirkegaard, 22V’s head of Europe Strategy and Geopolitics, on the EU-China outlook.
• Our Take: The above is broadly consistent with our views. Our cyclical outlook emphasizes the weak labor market as a particularly acute problem for demand (and for stabilizing the property sector). Aggressive fiscal stimulus is necessary to promote a faster macro recovery, but we do not expect this to come anytime soon (recent take HERE). We also agree that structural distortions in the allocation of credit lower the cost of capital for Chinese firms but also suppress equity market returns.

TAIWAN AND GEOPOLITICAL FLASHPOINTS

• While Taiwan is the most sensitive and potentially dangerous issue in the US-China relationship, experts do not see risks of conflict as particularly high over the next several years.
• Domestic politics in Taiwan, US policy towards Taiwan, and current US-China military capabilities broadly allow the CCP to plausibly claim that it can and should afford to be patient for now. However, this is not a stable equilibrium, and the risks rise if Beijing becomes highly confident in its military capabilities and/or decides that political developments are moving against it.
• Taiwan’s next presidential election in 2028 will be a major risk watchpoint. If President Lai Ching-te wins a second term decisively, Beijing may run out of strategic patience. Political victory by the opposition KMT party would cheer Beijing but could also lead to domestic protests (and U.S. concerns) if KMT leadership pursues political dialogue with China more aggressively than Taiwan’s public tolerates.
• Trump’s willingness on his trip to discuss U.S. arms sales to Taiwan with Xi Jinping – a break with longstanding U.S. policy – is a source of consternation, with some experts warning that it risks undermining confidence in the U.S. in Taiwan and the broader region. However, U.S. officials are adamant that there has been no change in U.S. policy toward Taiwan.
• Experts assess that China-Japan trade tensions will remain high, though see China’s hosting of APEC in November as a reason for both sides to pursue a modest de-escalation.
• Our Take: We also view near-term Taiwan conflict risks as low, and will discuss these dynamics in a forthcoming report.

CHINESE ELITE POLITICS

• Experts are unanimous that Xi Jinping will seek a fourth term in power at the 21^st Party Congress in October/November 2027 and does not face significant obstacles. China’s political system is effective at many things, but a smooth succession of power has not been one of them.
• Xi (who turns 73 later this month) may decide to stay in power for another ten years or longer to groom a successor and preserve his legacy. That successor has not been identified, as there are numerous potential candidates in the next generation (or two) of officials.
• Xi has been effective in his pursuit of industrial dominance and technological self-reliance. But Xi’s political power does not translate to the ability to dictate outcomes at the local level, with problems such as excess capacity and high debt endemic to China’s political system.
• Importantly, political analysts do not believe that Xi will take reckless gambles to reunify China with Taiwan. If forceful reunification would endanger his broader ambitions for China’s national rejuvenation, he would not take that risk just to make this his legacy achievement.
• Our Take: Xi’s consolidation of power means that China’s policy outlook will see more continuity than change. However, the 21^st Party Congress next fall will still be a highly significant event, with Xi’s supporting cast gaining increased importance as he ages.