Xi–Trump Summit: The Limits of Transactional Diplomacy

Summary

The Trump–Xi summit will test whether transactional diplomacy can contain an increasingly structural and competitive US–China rivalry. The more consequential questions concern Taiwan and the shifting US military posture in the Indo-Pacific, where concessions could have consequences beyond the immediate bargain.

The 24 September 2026 summit between Presidents Donald Trump and Xi Jinping will be closely watched for answers on the future of international order, technology and great-power competition. The last meeting between the two in May 2026 produced problematic developments, with the US seemingly capitulating to Chinese demands for equality and parity through ‘constructive strategic stability’, and attempting to build consensus around negotiated bipolarity. Great power competition has not eased despite these efforts; both nations have continued to use economic leverage, domestic laws and diplomacy to undercut the other. Therefore, the contention points—Taiwan, AI, rare earths and chips, trade and a collapsing security order—have deepened since the last meeting.

When Xi Jinping steps off his plane in Washington, the imagery of his interactions will do much of the work the substance of the visit cannot. Xinhua already gave us a glimpse into what the Chinese side may be thinking by repeating the questions asked by Xi during the May meeting, where China appeared as an equal partner in maintaining global peace and stability:

Can China and the United States overcome the Thucydides Trap and create a new paradigm of major-country relations? Can we meet global challenges together and provide greater stability for the world? Can we build a bright future together for our bilateral relations in the interest of the well-being of the two peoples and the future of humanity?^[1]

Xi is very likely to use the same assertive, confident tone he used in May. Trump, on the other hand, would like to record this meeting as “great” but with a number attached. Essentially, this meeting may not yield real results on the ground; the optics will matter more to both leaders. The current consensus is that this meeting is not a reset, as nothing underlying has actually reset.

The trade truce struck in Busan in October 2025 suspended export controls on rare earths, which expires on 10 November 2026. If the truce is not extended or made permanent, the US will suffer from a supply shortage, hurting its semiconductor industry. Reportedly, China’s shipments of rare earth magnets to the US fell 13 per cent year on year to 512 tonnes in August.^[2] This followed Beijing’s resumption of some export approvals after trade negotiations, and shipments recovered from their lows in 2025.^[3]

This highlights that the relationship is trying to be transactional but is limited by strategic constraints. While rare earths give China significant leverage in negotiating the terms of the relationship, the US is not without leverage. But since Busan, the bilateral relationship has changed little, except for a sharp increase in senior meetings and phone calls, suggesting both leaders have concluded the relationship is precarious and dangerous but cannot be left unmanaged.

Trade talks are likely to yield the most visible results. China is now over halfway to its soybean target, politically positive for Trump’s farmer base, but its non-soybean agricultural pledge of US$ 17 billion has largely stalled.^[4]  This is not an accidental bureaucratic error; it is a calculated move of selective compliance, a pattern visible in rare-earth licensing. China wants to keep the relationship manageable by buying enough headline-grabbing commodities, while letting harder negotiations slide. This is the tone of the relationship: manage it with non-strategic purchases to keep the optics positive for both audiences.

Taiwan will be one of the most crucial talking points. In December 2017, Trump’s first National Security Strategy broke with a half-century of engagement by naming China a ‘revisionist power’ and a ‘challenger to American power’ seeking to ‘shift the regional balance of power in its favour’.^[5] This historic moment would change the international order forever. Trump, in other words, has already once correctly, or at least consequentially, enabled a structural shift in the international order before the rest of the world could understand and hedge against the ramifications of these changes. Therefore, the question remains whether the Taiwan arms sales are about to become the second such historic move, albeit in the wrong direction.

Trump has been diffusing the deterrence that US presence had achieved by undermining its commitments to Taiwan. For instance, Trump called the arms sales to Taiwan “a very good negotiating chip”, referring to the US$ 14 billion arms transfer to Taiwan, which Trump has held in abeyance since May 2026, after his summit with Xi. Such a framing by the head of state of a principal ally, explicitly linking treaty commitments with negotiations with Beijing, highlights that he is not rigid and can be convinced.^[6] Recently, Trump has taken a rather soft stand on China. Referring to China’s alleged support for Iran, Trump replied, “When they say China spies on us, I say ​you’re right, and we spy on them too”, minimising national security concerns, troop safety and US interests.^[7]

Given that Trump wants to negotiate over Taiwan with Xi, the latter will likely push to halt the arms sales, or at least delay them by months or a year.^[8] If successful, even an implicit understanding would erode deterrence, making China more belligerent, the region more unstable, and the US an untrustworthy ally. The catch here is that capitulating to Chinese demands on geopolitics will increase the likelihood of conflict, not reduce it.

Additionally, the ambiguity itself is a vulnerability; it is not just what gets conceded, but the fact that concession is even on the table hurts US security interests in the Indo-Pacific, unless those interests have changed. The three things to pay attention to are: whether the US$ 14 billion arms sales package is passed, which includes 400–500 of the newest PAC-3 MSE missiles, Lower Tier Air and Missile Defence Sensor (LTAMDS) and Integrated Battle Command System (IBCS) technology;^[9] if previously purchased F-16s arrive on schedule in September; and whether President Lai Ching-te is permitted to transit the US.

The hemispheric division highlights this problem as well. The ‘Donroe Doctrine’ underscores Trump’s Western Hemispheric assertiveness, with explicit demands that Venezuela sever ties with China, Iran, Russia and Cuba as a precondition for resuming oil production after Maduro’s capture.^[10] It tells Beijing that it can secure the First and Second Island Chains through negotiations, as the highest-level American strategic thinkers were contemplating the same.

The ongoing war in Iran made this calculation more plausible, as it is no longer abstract; assets from the Indo-Pacific have actually been moved to West Asia. The 31^st Marine Expeditionary Unit, the only Marine unit permanently forward-deployed in the Indo-Pacific, with 2,500 Marines, has been in West Asia since March 2026.^[11] Other developments include the USS Abraham Lincoln carrier strike group departing the South China Sea in January 2026;[12] THAAD batteries and interceptors being removed from South Korea, a symbol of the ‘pivot to Asia’ strategy that China and North Korea fought against in 2017. To sum it up: rhetorical ambiguity on arms sales, a live-hemispheric-swap discourse, and a degraded force posture on the ground suggest a significant change in the international order is on the horizon.

Something even more interesting is happening on artificial intelligence (AI). There is an ongoing debate in the US on arresting the pace of AI innovation. Recently, Dario Amodei, CEO of Anthropic, wrote a piece urging the US and China to “slow the pace at which we improve the capabilities of AI models”.^[13] OpenAI CEO Sam Altman and Tesla CEO Elon Musk joined the call; this three-way convergence among the heads of rival labs was rare. While Trump and Beijing both denied the slowdown, a signed article by Chen Yixin, Minister of State Security, China’s most senior intelligence official, named AI a “systemic” threat to CPC’s political security, describing it as “the main battlefield of global scientific and technological competition and a new arena for strategic rivalry among major powers”.^[14]

The timing and register are the main takeaway: only days after Chinese officials publicly dismissed Western AI-safety warnings as alarmist, the criticism appears. It shows the zero-sum trap: China dismissed Western calls for a slowdown, while its top intelligence official identified the same AI risks. It underscores that both governments are aware, at the highest level, that AI carries a genuine, potentially catastrophic risk. Still, neither can afford to act on that knowledge by actually slowing down, as in this great power competition, unilateral restraint hands the other side a model capable of “rapidly discovering vulnerabilities in batches, automatically chaining together attack paths”, to quote Chen.^[15]

This brings in the next major contention point–Chips, where the US holds leverage. This relationship has moved further than any single track since Busan, which also reveals most clearly why full convergence remains impossible. China sees the US offer of H200 chips as a strategic trap, not a concession. China’s focus on chip self-reliance has delivered results, as domestic chip production rose from 16 per cent to 28 per cent in one year.^[16] While the US retains its dominance in chip manufacturing, China is set to close the gap in the next few years. China is choosing strategic patience to let its industries develop and remove an important American chokehold. But if China cannot close the gap, its AI innovation would suffer, allowing the US to take a colossal lead. Therefore, a deal allowing H20/H200 in exchange for rare earth supply agreements is possible.

Another likely talking point will be rare earths, where China has an almost absolute monopoly, controlling 70–85 per cent of global mining and 90 per cent of refining and processing capacity.^[17] The April 2025 export-license agreement on seven heavy rare earths remains in force today, amplified by the October 2025 controls, extending Chinese jurisdiction to any foreign product containing even 0.1 per cent Chinese-origin rare earth content.^[18] The US has been unable to diversify away from Chinese rare-earth supply. US officials have complained vociferously against Chinese export licensing, as exports of permanent magnets continued to drop, which “creates ongoing uncertainty, making full confidence premature”.^[19]

The Xi–Trump summit, therefore, is unlikely to resolve the contradictions that have spread across the bilateral relationship; if anything, this meeting will highlight those differences. Trade can be negotiated because both sides have incentives to preserve commercial flows; rare earths can be traded because China’s dominance gives it leverage and US dependence gives it a reason to bargain; AI can be discussed because both governments recognise the risks of uncontrolled competition. But none of these negotiations alters the underlying strategic contest. Taiwan remains a question of sovereignty and deterrence, technology remains a question of national power, and the distribution of military and economic capabilities remains contested.

This is why, like the last meeting, the optics of the summit may ultimately matter more than the substance. Xi is likely to reiterate China as an equal, building on the language of “constructive strategic stability”. Meanwhile, Trump will want visible material gains to impress his domestic voters ahead of the upcoming mid-term elections in November. The two leaders therefore need not agree on the meaning of the relationship to make the summit successful. They only need enough transactional progress to demonstrate that confrontation remains manageable.

Taiwan will be a clear test of where the international order is heading. If arms sales become a bargaining instrument between Washington and Beijing, the immediate transaction may be presented as part of a broader effort to stabilise relations. But the precedent matters more than the individual package. If Taiwan is placed on the negotiating table, the US will have capitulated to China, with far-reaching consequences.

The same logic applies to the military balance. The diversion of US attention and assets towards West Asia, combined with uncertainty surrounding the future force posture in the Indo-Pacific, creates a strategic environment in which perceptions can matter as much as actual capabilities. For Beijing, any reduction in American attention to the region may create an opportunity to test the limits of US commitment. For Washington, preserving deterrence requires demonstrating that other crises have not displaced competition with China. The question is therefore whether the US can manage multiple theatres simultaneously without allowing strategic ambiguity to become a strategic opportunity for its principal competitor.

^[1] “Xi Steers China-U.S. Relations Toward Stability”, Xinhua, 21 September 2026.

^[2] Nayan Seth, “Rare Earths Friction Threatens to Stall Trade Truce Extension Ahead of Xi-Trump Summit”, South China Morning Post, 22 September 2026.

^[3] Ibid.

^[4] Hallie Gu and Erin Ailworth, “China’s Selective Crop Buying Tests US Trade Truce Before Summit”, Bloomberg, 21 September 2026.

^[5] “National Security Strategy of the United States of America”, The White House, December 2017, pp. 2, 25.

^[6] Chris Buckley, “Trump Makes a High Risk Move to Win Over Xi”, The New York Times, 16 May 2026.

^[7] “Trump Downplays Report China Entities Helped Iran Before Attack That Killed US Troops”, Reuters, 14 September 2026.

^[8] Trevor Hunnicutt, Yimou Lee and Ben Blanchard, “Xi to Press Trump to Halt Taiwan Arms Sales Under 1982 Agreement, Sources Say”, Taipei Times, 22 September 2026.

^[9] Lo Tien-pin and Jake Chung, “US Arms Package to Boost T-Dome Project: Source”, Taipei Times, 6 August 2026.

^[10] “China Finds Risks, Opportunities as Trump Pushes for ‘Spheres of Influence’”, Al Jazeera, 8 January 2026.

^[11] “Marine Commandant Acknowledges Indo-Pacific Gap After Middle East Deployment”, Stars and Stripes, 18 May 2026.

[12] “US Military Assets Exit Asia”, National Security Desk, Substack, 31 May 2026.

^[13] “We Must Pace the Frontier”, Dario Amodei, September 2026.

^[14] Lucy Quaggin, “Trump Condemns ‘Sick Conspiracy’ Against AI, Says He is Only Guardrail Needed”, South China Morning Post, 14 September 2026; “On September 14, 2026, Foreign Ministry Spokesperson Guo Jia-kun Hosted a Regular Press Conference”, Ministry of Foreign Affairs, People’s Republic of China, 14 September 2026; Chen Yixin, “全面筑牢人工智能安全屏障 推动人工智能健康有序发展” [Building a Comprehensive AI Security Barrier and Promoting the Healthy and Orderly Development of Artificial Intelligence], China Cyberspace, No. 8, 2026, reproduced by Security Reference (安全内参), 13 September 2026.

^[15] Ibid.

^[16] Cohen Trace, “US-China Tech Decoupling in 2026: Nvidia’s Export Ban, SMIC’s 7nm Chips, and What’s Actually Split”, Value Add VC, 19 July 2026.

^[17] “China Rare-Earth Magnet Exports Fall 13% as Trump-Xi Summit Tests Beijing’s Leverage”, Rare Earth Exchanges, 21 September 2026; Chen Jing,  “How China Weaponised Its Rare Earth Monopoly”, Think China, 21 July 2026.

^[18] “The Ministry of Commerce Announcement No. 61 of 2025 Announced the Decision to Impose Export Controls on Relevant Rare Earth Items from Overseas”, Business Policy and Regulation Database, Ministry of Commerce of the People’s Republic of China, 9 October 2025.

^[19] “Rare Earths Friction Threatens to Stall Trade Truce Extension Ahead of Xi-Trump Summit”, South China Morning Post, 22 September 2026.