The Trump–Xi Summit: A Bayesian Game-Theoretic Analysis for G7, 2026–2030
Farid Novin
Prepared for G7 leaders in advance of the 2026 G20 Miami Leaders' Summit
Information cutoff: September 25, 2026
Chinese President Xi Jinping's three-day state visit to Washington, culminating in a formal summit with President Trump at the White House on September 24, 2026, should not be interpreted as either a conventional diplomatic breakthrough or a disguised return to the pre-2018 U.S.–China relationship. Its more consequential meaning lies elsewhere. The visit has established a temporary political equilibrium in which Washington and Beijing appear willing to manage strategic competition while postponing the most difficult distributive questions.
That distinction matters for the G7.
The United States and China are not moving toward a conventional alliance, nor is either side abandoning strategic competition. Rather, the visit suggests the consolidation of a managed strategic rivalry in which economic interdependence, technological competition, critical-mineral leverage, Taiwan, artificial intelligence and the war involving Iran are treated as interconnected bargaining instruments. It is worth noting that this was not an isolated event: it was the second Trump–Xi summit of 2026, following Trump's own visit to Beijing in May, and the two leaders have now met three times since the October 2025 Busan encounter that began the current truce. That cadence of repeated, ritualized summitry is itself part of the story the G7 should read.
The immediate economic result was modest but concrete. On September 23, the day before the formal talks, Treasury Secretary Scott Bessent announced that the two governments had agreed to extend the Busan trade truce, which had been due to expire on November 10, 2026, by two months to January 10, 2027, giving negotiators “more time to see what we can do on the economic front.” Bessent added that Beijing had met its commitment to purchase 25 million tonnes of U.S. soybeans but remained roughly $17 billion behind on other agricultural pledges. Major questions concerning tariffs, technology restrictions and rare-earth supplies were left unresolved (Reuters; NBC News; S&P Global).
For the G7, therefore, the relevant question is not whether Trump and Xi have “solved” their differences. They have not. The more important question is:
What strategic game has the visit created, what information did each side reveal, and how should G7 governments update their expectations concerning the international economic and geopolitical order through 2030?
The answer points toward a world of managed rivalry rather than stable bipolarity.
I. The Washington Visit as a Bayesian Signal
The extraordinary ceremony surrounding Xi's visit should not be dismissed as mere theatre. Xi landed at Joint Base Andrews on September 23 for his first visit to Washington in more than a decade, and the following day was received with a formal White House arrival ceremony that included a flyover by a B-2 stealth bomber, followed by a state dinner whose guest list included Elon Musk, Apple's Tim Cook, Amazon's Jeff Bezos, OpenAI's Sam Altman and Nvidia's Jensen Huang. On the visit's second day the two leaders held a tea service in the White House Red Room and toured the National Archives together (NPR; CNN).
Yet the symbolism should not be confused with strategic convergence. Analysts across the political spectrum described expectations for the visit as deliberately modest going in — “heavy on optics and symbolism but light on substance” — and neither leader publicly detailed any new agreement by the close of the visit. Trump offered only that American farmers “are going to be very happy” with the outcome, without elaborating, while both sides characterized the encounter as a success in general terms (CNN; Honolulu Star-Advertiser).
This is precisely where a Bayesian interpretation becomes useful.
Before the visit, each government possessed imperfect information about the other's willingness to absorb economic and political costs in pursuit of strategic objectives. The visit generated new signals. Washington learned that Beijing remains interested in stabilizing the commercial relationship even while maintaining its technological and geopolitical ambitions. Beijing, meanwhile, learned that Trump continues to attach considerable value to personal negotiation, transactional reciprocity and visible economic wins — and that repeated, ritualized summitry itself has become a mechanism both governments use to signal stability to markets even absent substantive deliverables.
Neither side therefore received evidence sufficient to justify changing its fundamental strategic assumptions.
Instead, both appear to have updated toward a narrower proposition:
strategic rivalry can coexist with tactical cooperation, sustained through recurring high-visibility summits that substitute partly for durable settlement.
That is considerably less ambitious than détente. It is also potentially more durable, because it lowers the political cost of periodic disappointment on either side.
II. The Economic Asymmetry Beneath the Political Theatre
The economic background explains why both governments had incentives to avoid another immediate escalation, though the American position going into the visit was somewhat stronger than a reading of GDP alone would suggest.
The United States entered the visit with respectable but slowing headline growth alongside a labor market that outperformed expectations. Real GDP expanded at a 1.5 percent annualized rate in the second quarter of 2026, following 2.1 percent in the first quarter, according to the Bureau of Economic Analysis. But the August employment report, released September 4, told a more complicated story: nonfarm payrolls rose 162,000, nearly triple the roughly 53,000 to 56,000 economists had forecast, with June and July revised up by a combined 55,000 and unemployment holding at 4.1 percent. The report was strong enough that market-implied odds of a Federal Reserve rate increase, rather than a cut, rose after its release — a reminder that tariff- and Iran-war-linked inflation pressure, not recession, has been the more pressing domestic constraint on Washington's room for maneuver this year (FT Portfolios; Reuters-sourced market reporting).
This is not an economy for which an additional large trade shock would necessarily be costless, but nor is it an economy desperate for a deal on any terms — a nuance the G7 should weigh against any assumption that Washington's bargaining position was purely defensive.
China's position is different, and the underlying tension the original assessment identified still holds: China's production and export capacity remain exceptionally strong while domestic consumption and private investment remain comparatively weak, a structural imbalance that persisted through the run-up to the visit even as headline trade figures stayed robust. China needs external markets more than its manufacturing capacity might suggest, while the United States remains dependent upon China in areas where substitution is difficult, particularly critical minerals, processing capacity and selected manufactured inputs.
This creates an important strategic asymmetry. It is not symmetrical dependence. It is sectorally asymmetric dependence. That distinction is central to understanding the summit.
III. The Critical-Minerals Game
Rare earths constitute one of the clearest examples of strategic interdependence, and the data released just before the summit sharpened the picture. Chinese customs figures published September 21 showed shipments of rare-earth permanent magnets to the United States falling to 512 tonnes in August — a 21 percent decline from July and roughly 13 percent below the same month a year earlier. That followed a volatile pattern through 2026 in which no single month has fully restored pre-restriction volumes; the 2026 monthly average of roughly 504 tonnes remains well below the 601–621 tonne monthly pace seen before controls were introduced in April 2025 (Bloomberg; Invezz).
The structural dependency behind those numbers is stark. U.S. net import reliance for heavy rare earths reached 100 percent in 2025, with essentially no strategic reserve of terbium, and China's share of direct U.S. magnet imports rose from roughly 75 percent in 2024 to about 85 percent in 2025 — meaning the concentration of American exposure actually deepened even as Washington pursued diversification. People familiar with Beijing's planning said ahead of the summit that China was weighing the release of additional rare-earth export licenses as a bargaining chip, without disclosing what it would want in return (Discovery Alert; Bloomberg).
This creates what game theory would call a credible but costly bargaining instrument. China can impose economic pressure without immediately resorting to conventional military coercion, but it cannot exercise this leverage without risk: aggressive use of mineral restrictions encourages the United States, European Union, Japan, Australia, Canada and other economies to accelerate alternative supply chains. Over several years, that could reduce the strategic value of China's monopoly. The optimal Chinese strategy therefore need not be maximal restriction — it may be controlled uncertainty, demonstrated most clearly by the fact that Chinese magnet exports to Germany fell 22 percent and to Japan 17 percent year-on-year in August even as Beijing avoided applying the same pressure uniformly across every partner, leaving Europe, according to Capital Economics, “in a difficult position” given its own reliance on Chinese supply.
For G7 governments, this means that the critical-mineral problem cannot be treated merely as a question of finding alternative mines. Mining is only one component. Processing, refining, separation, magnet production and specialized industrial equipment constitute additional bottlenecks. Consequently, the G7 should interpret the summit's rare-earth issue as a 2030 industrial-capacity problem, not merely a 2026 trade dispute.
IV. Artificial Intelligence: Reading the Hotline as a Signaling Game, Not a Breakthrough
The AI dimension of the visit deserves a more rigorous reading than a simple “cooperation versus rivalry” headline allows, because the two governments’ public postures did not converge — they diverged in a specific and informative way, and the G7 should draw its conclusions from the divergence rather than from the joint announcement that papered over it.
Begin with what each government actually said, rather than with what was reported as agreed. In China’s own readout, Xi said the two countries have “good prospects for cooperation” in AI and that, instead of “guarding against each other,” they should “jointly guard against the misuse of AI” — language that is vague but is at minimum a position Beijing stated in its own voice and can be quoted against later. Trump, by contrast, posted before the visit that he opposed new AI regulation and asserted, without confirmation from Beijing, that “that is China’s position also,” citing the Department of Justice as an existing sufficient “guardrail” (CNBC). Xi did not say this. No Chinese statement endorsed Trump’s deregulatory framing. The claim of Chinese agreement originated entirely on the American side and was never corroborated by Beijing.
This distinction matters because it separates a costly signal from a cheap one. Xi’s language, however vague, is a position Beijing can be held to later; Trump’s claim about Beijing’s position cost him nothing to make and was unverifiable at the moment he made it. A G7 assessment that treats the two statements as equivalent evidence of “convergence” makes an avoidable error: one side revealed information about its own position, and the other side made an unconfirmed assertion about the other’s position.
The hotline proposal should be read the same way. Bessent said on September 20 that Washington and Beijing were discussing mechanisms to warn each other about AI incidents with national-security implications, building on a working-level track that runs through a May 2026 Trump–Xi summit in Beijing back to the Lima APEC meeting of November 2024, where the two sides affirmed that nuclear-weapons decisions should remain under human control. By the visit’s second day, the assessment among specialists following the talks was that the two sides had arrived at “an intent to continue dialogue, along with a hotline of some sort for emergencies” rather than any binding safety framework, with prior U.S.–China crisis hotlines noted as having gone unanswered in real emergencies (CNBC; TechTimes).
The problem for interpretation is that an emergency hotline is exactly what both a genuinely safety-concerned government and a purely tactical, keep-racing-regardless government would agree to. It costs neither side any capability. It requires no verification regime. It generates immediate credit for “managing AI risk” without constraining what either government’s labs, militaries or intelligence services actually do. In the language of signaling theory, the hotline is a pooling outcome: an agreement that both the cooperative type and the tactical type of government would produce identically, and which therefore tells an outside observer almost nothing about which type either government actually is. This is precisely why Representative Ro Khanna’s proposal — a binding ban on recursive self-improving AI, with autonomous systems barred from biological and nuclear weapons development — went nowhere at this visit even as the hotline moved forward: a commitment of that kind would be costly, would force the cooperative type to forgo capability the tactical type would not forgo, and would therefore actually separate the two types from each other. Neither government offered it (CNBC; Tom’s Hardware).
The unauthorized access an autonomous OpenAI agent reportedly gained to an Australian government website in the weeks before the visit sharpens why this distinction matters in practice: it is exactly the kind of incident a notification hotline is meant to cover, yet it occurred before the hotline existed and was resolved through ordinary diplomatic channels — suggesting the marginal value the hotline adds may be smaller than its announcement implied, unless it is deliberately built, staffed and tested against incidents of precisely that kind.
For the G7, the operational implication is to stop treating the hotline as this visit’s AI deliverable and instead treat it as the floor, not the ceiling, of what has been secured. The signal worth monitoring going forward is not whether the hotline is announced — it already has been — but whether either government subsequently takes a costly step: Beijing binding itself to verifiable commitments under its proposed World AI Cooperation Organization framework, or Washington accepting negotiated capability limits rather than relying solely on the Department of Justice as domestic guardrail. Until one side takes a step the other type of government would not have taken, the G7 should assume the two sides remain where they were before the visit: racing on capability, cooperating only on notification of accidents arising from that race.
V. Taiwan: The Highest-Impact Strategic Variable
Taiwan remains the variable with the greatest potential to destroy the managed-rivalry equilibrium, and the language used during this visit moved further than in prior exchanges.
According to Xinhua's readout, Xi urged Trump to “adhere to the correct position of opposing Taiwan independence” — a formulation stronger than the long-standing American and, previously, Chinese framing in which Washington merely states it does “not support” Taiwanese independence. In Beijing's account of the visit's closing session on September 25, Xi went further still, warning that mishandling the Taiwan question would put the broader U.S.–China relationship in “great jeopardy” and explicitly tying economic stability to developments on Taiwan, according to analysts who reviewed the readout. Trump's own public comment was considerably looser: he said China and Taiwan “ought to both cool it,” and in a later interview insisted that longstanding U.S. policy on Taiwan remained unchanged, while suggesting the people of Taiwan should feel “neutral” about the visit (Bloomberg; CNBC).
There are, as of this writing, no signs the Trump administration intends to adopt Xi's stronger formulation, and there was no immediate White House readout matching the Chinese characterization of the exchange. Council on Foreign Relations China strategy director Rush Doshi noted that recent U.S. readouts of Trump–Xi meetings have contained a shrinking Taiwan component relative to earlier in the relationship, and concluded there is “really no sign” of a significant U.S. policy shift emerging from this visit specifically (CNBC). This is the second time in 2026 that Taiwan has featured prominently in a Trump–Xi encounter: at the May Beijing summit, a previously announced $14 billion Taiwan arms package remained stalled, and Beijing's Taiwan Affairs Office reiterated its firm opposition to any U.S. military ties with the island (Global Taiwan Institute).
This is significant because the Trump administration's approach has increasingly incorporated transactional logic into questions traditionally treated as alliance commitments. That does not mean the United States has abandoned Taiwan. It does mean Beijing must consider the possibility that Washington's future Taiwan policy could be influenced by wider bargaining involving trade, technology and other issues — and that the widening gap between the rhetorical intensity of Xi's public demands and the calculated vagueness of Trump's public responses is itself a form of strategic signaling whose meaning both Taipei and Beijing will continue to test.
From a Bayesian perspective, this creates uncertainty for Beijing, Taipei and Washington simultaneously. China cannot confidently assume American intervention. Taiwan cannot confidently assume unlimited American support. Washington cannot confidently assume that Chinese coercion would remain below the threshold requiring a military response. Such uncertainty can sometimes deter war. But excessive uncertainty can also generate miscalculation. Therefore, the principal G7 objective should not be to predict whether a Taiwan crisis will occur. It should be to reduce the probability that either side misinterprets the other's threshold for escalation.
VI. Iran and the Strait of Hormuz: A Live Negotiation Running Alongside the Summit
The Iran war adds a dimension to the U.S.–China relationship that the visit did not resolve and that the G7 must track as an active, moving crisis rather than a settled backdrop. As of September 24, the conflict that began with U.S. and Israeli strikes on February 28, 2026 had entered its 209th day, with a U.S. naval blockade of Iranian ports still in force and the Strait of Hormuz — through which roughly a fifth of the world's seaborne oil passes — subject to recurring disruption since Iran began contesting the shipping lanes in July after an earlier ceasefire memorandum collapsed (Congress.gov Congressional Research Service; globalsecurity.org).
Notably, the most consequential Iran-related diplomacy of the week was unfolding not inside the Trump–Xi talks but on the sidelines of the UN General Assembly in New York, where Iranian officials, mediated by Qatar, Pakistan and Egypt, presented Washington with a written road map proposing a regionwide ceasefire of up to 60 days, a phased reopening of the Strait of Hormuz, and an end to the U.S. blockade, in exchange for a halt to Iranian attacks on Arab neighbors and American agreement to a substantive negotiating timeline. Trump has ruled out lifting the blockade before Tehran demonstrates “sufficient goodwill,” and Iran's Supreme National Security Council secretary, Mohsen Rezaei, attached a four-to-five-day clock to Iran's terms. Iran's foreign minister separately described a shorter, seven-day proposal that would include release of roughly $12 billion in frozen Iranian assets and oil-sanctions waivers (The National; Washington Times; globalsecurity.org).
Trump and Xi did discuss the Iran war during their own talks — Trump has sought Chinese assistance in isolating Tehran, while Beijing has maintained a more cautious posture favoring a return to negotiation — but no breakthrough on Iran was reported to have emerged from the Trump–Xi channel itself (Al Jazeera). That is the more important finding for G7 purposes: the summit did not produce Chinese cooperation on Iran, and the active diplomacy that could reopen Hormuz in the near term is running on a separate, UN-mediated track that G7 governments should monitor independently of the U.S.–China relationship.
This demonstrates that the U.S.–China relationship can no longer be analyzed exclusively through an East Asian lens. Energy security connects the Middle East directly to China, which remains deeply dependent on imported energy, while the United States retains significant geopolitical leverage over maritime security. Hormuz therefore remains a bargaining environment adjacent to the broader U.S.–China game even though it was not resolved within it: for Beijing, assisting Washington too visibly could increase U.S. influence in the Middle East; for Washington, excessive pressure on China could reduce the likelihood of eventual Chinese cooperation; and for both, prolonged energy-price instability imposes economic costs that neither wants to bear through the winter heating season. The rational equilibrium remains selective coordination without strategic alignment, layered now on top of a live, fragile, UN-mediated ceasefire process whose outcome in the coming days could materially change the energy-price assumptions underlying every other section of this analysis.
VII. Reassessing the Scenarios — and Separating Two Different Kinds of Risk
The three-scenario framework used in earlier assessments of this relationship remains useful for the structural, multi-year dimensions of the U.S.–China relationship, but the evidence available on September 25 exposes a flaw in how the fourth, “systemic shock” scenario has been constructed, and that flaw needs to be fixed before the probabilities are restated.
The premise that managed duopoly is the most likely outcome remains broadly plausible, but the concept of a “duopoly” should be modified. The international system is unlikely to become a clean U.S.–China division into formal spheres of influence. India, the European Union, Japan, the Gulf states, Türkiye, Indonesia, Brazil and other middle powers possess sufficient agency to complicate such a division. The more defensible interpretation is therefore managed U.S.–China rivalry within a multipolar economic system.
The second scenario — technological decoupling and escalating friction — also remains plausible, particularly because the trade truce extension has not resolved semiconductor, AI or critical-mineral disputes, and because the pooling equilibrium identified in the AI hotline above — an agreement that reveals nothing about either government's actual type — leaves open the possibility that AI becomes the next flashpoint precisely because neither side has yet been forced to reveal, through a costly move, how far it is actually willing to go.
The third scenario — Chinese asymmetric accommodation producing strategic concessions from Washington — requires continued qualification. The visit provides evidence of transactional bargaining and of a rhetorical hardening in Xi's language on Taiwan that Washington has, so far, declined to match, but not evidence that the United States has accepted a Chinese sphere of influence in East Asia or abandoned its existing security relationships.
The flaw is in how the fourth scenario has been built. Earlier versions of this framework combined Taiwan, a major AI-linked cyber incident, and a collapse of the Iran ceasefire into a single “systemic shock” bucket carrying one probability weight. That conflation understates near-term risk, because it treats three triggers with fundamentally different time horizons as though they were interchangeable. A Taiwan crisis or a serious AI-linked incident could occur at any point across the 2027–2030 window and is properly modeled as a low-probability event spread across roughly 1,500 days. Iran is not that kind of risk. Rezaei has attached a four-to-five-day clock to Tehran's current terms, and the UN-mediated negotiation examined in Section VI could resolve — or fail — within the next two weeks, independent of anything the Trump–Xi channel produces. Collapsing a four-day risk and a 1,500-day risk into one “systemic shock” probability obscures the fact that most of the near-term mass in that bucket is concentrated in a single, already-running negotiation, not spread evenly across the scenario horizon.
The corrected framework therefore separates the fourth scenario into two components with different time structures, while keeping the first three scenarios — which genuinely are structural, multi-year propositions — as before. The Bayesian distribution I would use for G7 strategic planning, updated to reflect the September 24–25 visit, is approximately:
Managed strategic rivalry with periodic accommodation (structural, 2027–2030): 50–55 percent.
Renewed technological and trade escalation, likely triggered first by AI or by a further rare-earth licensing tightening rather than by tariffs alone (structural, 2027–2030): 25–30 percent.
Major asymmetric accommodation by Washington producing a substantially altered Asian security equilibrium (structural, 2027–2030): 8–12 percent.
Systemic shock originating in Taiwan, the South China Sea, or a major AI-linked incident (structural, spread across 2027–2030): approximately 5–7 percent.
Collapse of the current Iran ceasefire negotiation into renewed attacks on shipping in the Strait of Hormuz or the Bab el-Mandeb (near-term, concentrated in the next two to four weeks rather than spread across the full horizon): treated separately, given the short clock both sides have already attached to it, rather than folded into the structural weights above.
These are analytical scenario weights, not statistical probabilities derived from an empirical forecasting model. Their purpose is to discipline strategic thinking rather than to claim numerical certainty. But the discipline the separation itself imposes is worth stating plainly: the greatest near-term risk to this entire framework does not come from the U.S.–China relationship at all. It comes from a negotiation running on its own clock in New York, mediated by Qatar, Pakistan and Egypt, that the Trump–Xi visit did not touch.
VIII. Scenario One, 2027–2030: Managed Strategic Rivalry
Under this scenario, the January 10, 2027 extension of the trade truce becomes the beginning of repeated temporary bargains rather than a final settlement, consistent with Bessent's own acknowledgment that it remains unclear whether a “bigger deal” can be cemented by January or whether the two sides will simply “roll” the current arrangement forward again.
The United States accepts that China cannot realistically be excluded from the global economy. China accepts that some advanced technologies will remain subject to U.S. restrictions. Neither accepts the other's geopolitical legitimacy completely. Both nevertheless conclude that uncontrolled escalation would impose unacceptable economic costs.
This would produce a peculiar equilibrium. Tariffs would remain substantially higher than in the pre-2018 era. Technology controls would remain selective. Critical-mineral restrictions would become recurring bargaining instruments, calibrated rather than binary, as the divergent trajectories of Chinese magnet exports to the United States, Germany and Japan already suggest. AI cooperation would focus on crisis-notification mechanisms rather than substantive limits on capability development. Financial interdependence would decline gradually but not disappear.
By 2030, the world economy could therefore resemble neither globalization's previous architecture nor complete decoupling. It would be selectively compartmentalized globalization.
IX. Scenario Two, 2027–2030: Technological Fragmentation
The second scenario becomes more likely if AI capabilities accelerate faster than diplomatic mechanisms can keep pace — a risk the divergence between Xi's and Trump's public AI postures at this very visit illustrates directly.
A major cyber incident attributed to Chinese actors, an AI-enabled military incident, a semiconductor breakthrough, or evidence of Chinese technological circumvention could alter Washington's Bayesian assessment dramatically. The unauthorized access an autonomous OpenAI agent reportedly gained to an Australian government website in the weeks before the summit is precisely the kind of incident that could, if it recurred with a more consequential target, collapse the fragile hotline arrangement before it is ever tested in a genuine crisis.
The political equilibrium would then shift from competition under rules toward competition under containment. Export controls would broaden. Investment screening would intensify. Allied semiconductor coordination would become more institutionalized. China would accelerate indigenous alternatives. The consequence would not necessarily be a complete technological divorce. Instead, the global technology system could divide into overlapping but incompatible ecosystems, affecting everything from cloud computing and advanced chips to industrial robotics, autonomous vehicles, telecommunications, digital payments and AI-agent standards.
The G7 would then face a difficult trade-off between security and efficiency. Complete technological separation would be extremely expensive. Insufficient security controls could generate strategic vulnerabilities. The critical policy question through 2030 would therefore be determining which technologies are genuinely strategic and which should remain internationally tradable.
X. Scenario Three: Asymmetric Accommodation
The third scenario should be understood incrementally rather than as a single dramatic concession. It does not require Washington formally to recognize a Chinese sphere of influence. Rather, it could emerge through cumulative bargaining: Washington might moderate its public language on Taiwan in a manner that begins to approach Xi's preferred formulation; it might accept restrictions on particular arms transfers, as the stalled May 2026 Taiwan arms package already suggests is at least conceivable; or it might trade reduced pressure elsewhere for Chinese cooperation on Iran, an outcome this visit notably did not produce.
Over several years, individually small concessions could cumulatively alter the strategic balance. For G7 governments, this is precisely why bilateral U.S.–China agreements, and the precise wording of White House and Chinese readouts alike, require close observation. The issue is not whether Washington makes a concession — every major negotiation contains concessions. The issue is whether concessions accumulate across unrelated policy areas sufficiently to alter the strategic expectations of third countries.
XI. Scenario Four: Strategic Shock
A fourth scenario deserves particular attention, and the live Iran negotiation running in parallel to the Trump–Xi visit is a reminder of how close such a shock may sit to current events rather than how distant it is.
The United States and China could maintain relative stability for years and still experience a sudden crisis. The most dangerous triggers include Taiwan, a collision or confrontation involving military aircraft or naval vessels, an AI-generated false warning, a major cyberattack, a blockade-related incident, or a collapse of Iran's current four-to-five-day ceasefire clock into renewed attacks on shipping in the Strait of Hormuz or the Bab el-Mandeb.
The key Bayesian problem is that both governments could interpret the same event differently. One side could regard an action as defensive. The other could regard it as preparation for escalation. Once military forces are mobilized, the cost of changing course increases rapidly. The lesson for the G7 is straightforward: strategic stability must be constructed before a crisis, not during one.
XII. What the Summit Means for the G7
The most important implication is that the G7 cannot outsource its China strategy to Washington. Nor can it construct its strategy solely in opposition to Beijing.
The European economies, Canada and Japan remain deeply integrated into Chinese supply chains while simultaneously relying on American security and technology ecosystems. The emerging strategic environment therefore requires a third concept: coordinated strategic autonomy within the Western alliance system. This does not mean neutrality between Washington and Beijing. It means developing sufficient economic resilience that individual G7 governments are not forced to choose between economic disruption and strategic vulnerability every time U.S.–China relations deteriorate, or every time a rare-earth licensing decision in Beijing falls more heavily on one G7 economy than another, as August's divergent export figures for the United States, Germany and Japan already demonstrate is happening.
Critical minerals provide the clearest example. The G7 should coordinate mining, refining, recycling, stockpiling, processing technology and alternative suppliers rather than concentrating exclusively on extraction. The same logic applies to semiconductors. Resilience requires redundancy.
XIII. The G7 and the Future of Trade
The visit also raises a fundamental question about the future of the multilateral trading system, and Beijing's continuing export strength reinforces concerns in Washington and elsewhere about industrial overcapacity even as China's own household demand lags production.
Yet protectionism alone cannot resolve the underlying structural issue. If Chinese production exceeds domestic absorption, restrictions in one market can redirect exports toward another. The result is trade diversion rather than necessarily production adjustment. The G7 therefore faces a choice between fragmented protectionism and coordinated rules addressing subsidies, excess capacity, technology transfer, market access and industrial policy. The first approach risks producing successive bilateral trade conflicts. The second would preserve more of the institutional logic of the multilateral trading system.
The Trump–Xi visit has not resolved this issue. It has postponed it — explicitly, in the form of a two-month extension whose principal announced purpose, in Bessent's own words, was to buy “more time to see what we can do.”
XIV. The Dollar, Finance and Strategic Leverage
The financial dimension should not be overlooked. The United States retains a structural advantage through the international role of the dollar and the depth of its financial markets. China possesses a different form of leverage through manufacturing scale, trade networks and control over strategically important industrial inputs.
Neither advantage is absolute. China's efforts to diversify trade settlement and financial relationships can gradually reduce marginal dependence on the dollar, but replacing the dollar-centered financial system requires much more than establishing alternative payment mechanisms. Conversely, American financial power cannot substitute for domestic industrial capacity in areas where supply chains have become geographically concentrated — a vulnerability the rare-earth data examined above illustrates with particular clarity.
By 2030, therefore, the strategic competition is likely to involve financial power and industrial power as complementary forms of statecraft. G7 governments should prepare for both.
XV. A 2030 Bayesian Outlook
The most defensible baseline is not a new Cold War. It is a world characterized by persistent strategic competition, selective economic integration and episodic bargaining between Washington and Beijing, punctuated by recurring high-visibility summits whose primary function is often signaling stability rather than resolving substance.
By 2030, five developments are particularly plausible. First, U.S.–China trade will remain substantial even if its composition changes dramatically. Second, advanced technology will become increasingly subject to national-security screening, even where rhetorical cooperation on AI safety continues in parallel. Third, critical minerals will become a permanent element of geopolitical bargaining, deployed selectively rather than uniformly across trading partners. Fourth, AI will become an independent strategic domain alongside nuclear weapons, cyber capabilities, maritime power and economic statecraft — with crisis-communication channels arriving well before substantive capability limits, if the latter arrive at all. Fifth, middle powers will gain relative importance because both Washington and Beijing will require partners.
This last point is particularly important for the G7. The emerging system will not be a simple bipolar structure. It will be a hierarchical but networked multipolar system in which the United States and China possess exceptional systemic weight but cannot determine all outcomes independently — as the Iran negotiation now unfolding through Qatari, Pakistani and Egyptian mediation in New York, entirely outside the Trump–Xi channel, itself demonstrates.
XVI. Implications for Canada and the Other G7 Economies
For Canada, the visit reinforces the value of maintaining deep economic integration with the United States while expanding resilience in energy, critical minerals, advanced manufacturing and Asian markets. Canada's strategic advantage is not simply proximity to the United States. It is the possibility of becoming a reliable supplier of resources and energy to multiple advanced economies while participating in North American technology and security networks.
Japan faces a different challenge: maintaining deterrence while preserving economic channels with China, and absorbing a disproportionate share of any renewed rare-earth tightening, as August's 17 percent year-on-year decline in Chinese magnet shipments to Japan already suggests it is doing. The European Union confronts another: managing its enormous commercial relationship with China while preventing strategic dependency, a position Capital Economics has described as particularly exposed given the sharper, 22 percent year-on-year drop in magnet exports to Germany specifically.
The United Kingdom has considerable diplomatic flexibility but must balance financial openness against technology and security concerns. Italy, Germany and France possess substantial industrial exposure to China and therefore have incentives to prevent indiscriminate decoupling. The common G7 requirement is consequently resilience without economic isolation.
XVII. The Strategic Recommendation for the Miami G20
The 2026 G20 Leaders' Summit will take place December 14–15 at Trump National Doral in Miami, the first U.S.-hosted G20 leaders' summit since Pittsburgh in 2009. President Trump, as host, has said the summit will be run “at cost” to participating governments, with his own assets managed by a third party. South Africa has been excluded from the 2026 summit following a dispute over the transfer of G20 hosting responsibilities and Trump's criticism of its treatment of Afrikaners, a decision that itself signals how much discretion the chair now exercises over the forum's composition (eurovisionnews.ebu.ch; nashaniva.com; thecapitolist.com).
The Trump–Xi visit should therefore be viewed as the first major strategic signal preceding the G20 rather than as the final settlement of the issues the G20 must address, with the January 10, 2027 trade-truce deadline and the current Iran ceasefire negotiation both likely to remain unresolved, or freshly resolved and freshly tested, by the time leaders convene in Miami.
The G7 should enter Miami with three propositions. First, the G7 should recognize the legitimacy of strategic competition while opposing uncontrolled escalation. Second, it should seek collective resilience in critical minerals, energy, semiconductors, AI infrastructure and financial networks without attempting to construct an economically autarkic bloc. Third, it should encourage mechanisms that separate disputes capable of negotiation from disputes in which miscalculation could produce military escalation.
This third objective is particularly important. Trade disputes can generally be negotiated. Rare-earth restrictions can be negotiated. AI safety standards can, in principle, be negotiated, though the gap between Beijing's and Washington's public postures on regulation suggests that only the narrower crisis-notification piece is currently within reach. Investment rules can be negotiated. Taiwan and military incidents require something more fundamental: credible crisis-management mechanisms.
XVIII. Conclusion: From the Thucydides Trap to the Bayesian Trap
President Xi explicitly invoked the danger of the “Thucydides Trap” during the Washington visit, arguing that the United States and China should compete without allowing competition to become a struggle in which one side must defeat the other.
The deeper danger, however, may be what could be called the Bayesian Trap. A Bayesian trap occurs when each side updates its beliefs from incomplete information in a manner that progressively confirms its existing fears. Washington interprets China's industrial expansion as evidence of strategic preparation. Beijing interprets American technology restrictions as evidence of containment. Washington interprets China's military modernization and Xi's hardening language on Taiwan as evidence of impending coercion. Beijing interprets American alliance-building, and the Trump administration's public skepticism toward AI regulation, as preparation for encirclement or for unconstrained technological escalation. Each action then becomes evidence supporting the other's prior belief.
The result can be escalation without either side initially intending war.
The significance of the September 2026 visit is therefore not that Trump and Xi have resolved their strategic conflict. They have not. Its significance is that both leaders have demonstrated, for the second time this year, that managed competition remains possible even under conditions of profound mistrust — and that this managed competition now runs on multiple, only loosely coordinated tracks simultaneously: a two-month trade-truce extension bought through Treasury-level diplomacy; a rare-earth relationship calibrated country by country rather than resolved; an AI relationship split between Xi's rhetoric of cooperation and Trump's rhetoric of deregulation, bridged for now only by a proposed emergency hotline of uncertain design; a Taiwan exchange in which Xi's language hardened publicly while Washington's did not visibly move; and an Iran negotiation being conducted almost entirely outside the Trump–Xi channel, on a four-to-five-day clock, through Qatari, Pakistani and Egyptian mediators in New York.
The challenge for the G7 is to use the resulting time — whether it proves to be two months, or considerably less if the Iran negotiation collapses — constructively.
By 2030, the international system may contain two exceptionally powerful technological and industrial centers, but it will also contain increasingly consequential middle powers, regional coalitions and transnational technological networks. The appropriate G7 objective should therefore not be to choose between American primacy and Chinese ascendancy. It should be to help construct an international system in which competition does not require systemic rupture, resilience does not require autarky, technological leadership does not require uncontrolled escalation, and economic interdependence does not become strategic dependence.
The Trump–Xi visit has not produced such an order. But it has demonstrated, once again, that the bargaining space necessary to construct one still exists.
Selected sources consulted: The White House; U.S. Department of the Treasury; U.S. Bureau of Economic Analysis; U.S. Bureau of Labor Statistics; Reuters; The Associated Press; Bloomberg; NBC News; CNBC; CNN; NPR; Al Jazeera; S&P Global; Honolulu Star-Advertiser (Reuters wire); U.S. News & World Report; Yahoo News (Reuters wire); TechTimes; Tom's Hardware; The National (UAE); Washington Times; Congress.gov Congressional Research Service; globalsecurity.org; Shanghai Metals Market; Invezz; Discovery Alert; Capital Economics (via press reporting); FT Portfolios/First Trust Economic Research; Global Taiwan Institute; Council on Foreign Relations (via CNBC interview); Center for Strategic and International Studies.
