Pondering Robert Tetlow’s Unsolicited Advice for Kevin Warsh: A Bayesian Perspective on Forward Guidance, Transparency, and Monetary-Policy Credibility

Tetlow's framework appears to assume, at least implicitly, that the central bank operates within a reasonably stable and well-behaved probabilistic environment in which the reaction function can be identified, communicated, and subsequently inferred by market participants. A Bayesian perspective begins from a more demanding premise: the reaction function itself is conditional on the information set, the underlying economic model, the estimated parameters, and the distribution of risks—and all of these can change.

This distinction matters enormously.

The central question is not simply whether the Federal Reserve should communicate its reaction function. It is whether there is, at any particular moment, a sufficiently stable and identifiable reaction function for communication of that function to provide reliable guidance about future policy.

The distinction I would therefore draw is between forward guidance as commitment and communication as disclosure of the central bank's evolving posterior beliefs.

I. The Central Bayesian Objection

Tetlow's argument can be summarized approximately as:

Transparency → understanding of the reaction function → formation of market expectations → improved monetary-policy transmission.

Within a relatively stable monetary-policy environment, this is an entirely reasonable proposition.

But the proposition becomes considerably less secure when the central bank confronts structural breaks, geopolitical shocks, fiscal transformations, financial instability, nonlinear inflation dynamics, changing productivity, or uncertainty about the parameters of the monetary-policy transmission mechanism.

Under such circumstances, the reaction function should not be regarded as a fixed object waiting to be discovered by financial markets.

A Bayesian central bank is continuously updating its beliefs about the economy.

In simplified notation, its evolving assessment can be represented as:

P(θ | It)

where θ represents the uncertain characteristics of the underlying economic structure and It represents the information available at time t.

The appropriate policy instrument therefore depends not merely on observed inflation or unemployment, but on the central bank's evolving assessment of the economic structure, risks, and uncertainty surrounding its models. In simplified form:

it = f[Et(θ | It), Rt, Ωt]

where:

• it = the policy interest rate;

• Et(θ | It) = the central bank's updated assessment of the economic structure;

• Rt = the relevant distribution of risks; and

• Ωt = uncertainty surrounding the model and its parameters.

The implication is profound:

A central bank should not be expected to remain faithful to yesterday's reaction function merely because yesterday's communication caused markets to form expectations around it.

Indeed, doing so can invert the proper relationship between information and policy.

Policy should respond to information. Information should not be subordinated to the preservation of a previously communicated policy path.

II. Forward Guidance: Commitment versus Conditional Belief

Tetlow provides a useful distinction between Odyssean and Delphic forward guidance. Odyssean guidance involves a commitment that can constrain future policy, while Delphic guidance communicates the central bank's economic outlook and the associated likely policy path.

This distinction is important, but a Bayesian interpretation takes the argument one step further.

The fundamental problem with strong forward guidance is not merely that the forecast may turn out to be wrong. Forecast errors are unavoidable. The deeper problem is that a sufficiently credible commitment can alter the policymaker's incentives after new information arrives.

Suppose a central bank communicates in January:

Given the information available today, we expect the policy rate to follow path X.

By April, however, new information materially changes the probability distribution over possible economic states.

A Bayesian policymaker should then be able to say:

The information has changed; our assessment has changed; therefore our policy has changed.

If instead the policymaker concludes:

We must continue toward X because we previously told financial markets that we would,

then forward guidance has ceased to be merely a communication device. It has become a constraint on Bayesian learning.

That is not necessarily credibility.

It can become institutional inertia disguised as credibility.

This is precisely where Christine Lagarde's observation at Sintra becomes particularly interesting. As Tetlow notes, Lagarde expressed regret that she had at one point felt "bound" or "compelled" by forward guidance.

From the conventional perspective, the lesson is that Odyssean forward guidance can tie the hands of policymakers.

From the Bayesian perspective, the lesson is deeper:

A commitment can increase the cost of responding rationally to new information.

Once markets have incorporated a policy commitment into asset prices, reversing course can become costly. The central bank may consequently find itself caught between two objectives:

New information → optimal policy may change

while simultaneously:

Previous guidance → cost of changing policy increases.

The institution can then become trapped between economic optimality and reputational consistency.

For a Bayesian policymaker, that is precisely the danger.

III. Does Tetlow Conflate Credibility with Consistency?

This is perhaps the most important conceptual difference between Tetlow's framework and the Bayesian perspective developed here.

Tetlow is rightly concerned that an abrupt retreat from forward guidance could weaken transparency and impair the monetary-policy transmission mechanism. But there are at least two different meanings of credibility.

Procedural credibility

The central bank does what it previously said it would do.

Epistemic credibility

The central bank changes what it does when sufficiently strong evidence changes its assessment—and clearly explains why.

These two forms of credibility are not necessarily identical.

Indeed, under conditions of radical uncertainty, epistemic credibility may be more valuable.

A central bank that rigidly follows an obsolete forecast merely to preserve consistency may appear predictable, but it may simultaneously lose credibility as a competent economic institution.

Conversely, a central bank that changes course whenever the latest data point changes would also be behaving badly.

Bayesian updating is not data-point chasing.

The relevant question is whether the weight of evidence has changed sufficiently to alter the posterior distribution of beliefs.

The Bayesian central banker therefore asks:

• How reliable is the new information?

• Is it persistent or transitory?

• Is it consistent with other evidence?

• Does it alter the estimated parameters of the model?

• Does it support one model over another?

• Has the probability of a major regime change increased?

• Has the balance of risks materially shifted?

This is very different from mechanically reacting to each incoming statistic.

IV. The Reaction Function Is Not Necessarily a Stable Object

Tetlow's proposed alternative to forward guidance is particularly interesting. If the Federal Reserve does not wish to commit itself to a future policy path, it can explain its reaction function sufficiently well that markets can infer future policy from incoming economic data.

The idea is elegant.

But it rests on an important assumption:

that the reaction function is sufficiently stable and identifiable for markets to infer it.

That assumption deserves considerably more scrutiny.

A conventional Taylor-type reaction function can be represented as:

it = r + πt + α(πt − π) + β(yt − y*t)**

where:

• it is the policy interest rate;

• r* is the equilibrium real interest rate;

• πt is inflation;

• π* is the inflation objective;

• yt − yt* is the output gap; and

• α and β represent the policy response to inflation and economic slack.

But almost every important variable in this equation raises a Bayesian question.

What is the true value of r*?

What is potential output?

What is the output gap?

What is the underlying Phillips curve?

How persistent is inflation?

Has the monetary transmission mechanism changed?

Has fiscal policy altered the equilibrium relationship?

Has productivity shifted?

How should geopolitical shocks be incorporated?

Has the relationship between unemployment and inflation changed?

These are not simply observed quantities.

They are latent variables, uncertain parameters, and model-dependent estimates.

Consequently, markets cannot simply "learn" the Federal Reserve's reaction function from incoming data as if they were observing a stable mechanical rule.

The data themselves must first be interpreted.

And interpretation requires a model.

V. From Parameter Uncertainty to Model Uncertainty

This is where the Bayesian argument becomes more consequential.

In a conventional framework, uncertainty is often represented as uncertainty about the values of parameters within an assumed model.

A more genuinely Bayesian approach also allows uncertainty about the models themselves.

Instead of asking only:

What are the values of θ within model M?

the central bank must sometimes ask:

Which model M is more plausible in the first place?

The relevant posterior can therefore be represented as:

P(Mj, θj | It)

where Mj represents competing models and θj represents their associated parameters.

The central bank is therefore updating not merely its estimates of economic parameters but potentially the probability assigned to competing representations of the economy itself.

This is a substantially more difficult problem.

It also explains why a simple reaction function can be useful as a benchmark without being an adequate description of actual monetary-policy decision-making.

Tetlow himself acknowledges an important qualification in his footnote: central banks with which he is familiar generally use simple monetary-policy rules as guides rather than mechanically adhering to them.

That observation is important because it actually brings his position closer to the Bayesian perspective.

If the reaction function is ultimately a guide rather than a binding rule, then the crucial issue becomes how policymakers update their judgment when the assumptions underlying that guide change.

VI. Transparency Is Not the Same as Revealing a Mechanical Rule

Tetlow argues, persuasively, that transparency is an important obligation of an independent central bank. I agree with the democratic and institutional logic of this argument.

But I would qualify the proposition that transparency necessarily requires the communication of a stable reaction function.

There is an important distinction between:

transparency about reasoning

and

transparency about a predetermined policy rule.

A Bayesian central bank can be highly transparent while explicitly acknowledging that it cannot provide a reliable mechanical rule for the future because the probability distribution over economic states is changing.

Indeed, such communication may sometimes be more intellectually honest than presenting a highly precise reaction function that conveys a false impression of certainty.

The central bank can explain:

1. what it currently believes;

2. which evidence has caused that assessment;

3. how its assessment differs from the previous assessment;

4. which risks could cause another revision;

5. which indicators it is monitoring;

6. how policy would respond under alternative economic states.

This is not opacity.

It is conditional transparency.

And conditional transparency may be particularly appropriate in a world characterized by structural change and radical uncertainty.

VII. The "Vacuum" Argument

Tetlow argues that if the Federal Reserve does not communicate, somebody else will fill the informational vacuum. This is a reasonable concern.

Financial markets will inevitably form expectations.

But the existence of market expectations does not establish that the central bank should attempt to control those expectations through detailed forward guidance.

The deeper question is:

Should the central bank tell markets what it expects to do, or should it teach markets how it learns?

The Bayesian answer is the latter.

The central bank should not attempt to tell markets what the future will be.

It should explain how its beliefs about the future are formed and how new information can change those beliefs.

The difference can be expressed simply.

Conventional forward guidance

"We expect rates to be X."

Rule-based guidance

"Rates will respond according to rule Y."

Bayesian conditional guidance

"Our current assessment assigns the greatest probability to state A. If incoming evidence materially increases the probability of state B, our policy response will change accordingly."

The third approach preserves flexibility without sacrificing transparency.

It also provides markets with something potentially more valuable than a forecast:

an understanding of the architecture of the central bank's learning process.

VIII. Where Tetlow's Criticism of Warsh Is Persuasive

The Bayesian critique of Tetlow should not be mistaken for an unconditional defense of Kevin Warsh.

Tetlow identifies an important issue when he argues that rejecting forward guidance should not mean abandoning communication about contemporaneous policy decisions.

I agree.

Rejecting forward guidance does not logically imply rejecting explanation.

If Warsh refuses to explain the reasoning behind a current policy decision, then Tetlow's criticism has force.

A Bayesian central banker should arguably be more communicative about the information process, not less.

The appropriate distinction is therefore:

Forward guidance

"We intend to do X in the future."

Bayesian transparency

"Given today's information, we assign particular probabilities to competing economic states. This assessment explains today's decision. New information can alter those probabilities and therefore alter our policy."

The latter may provide markets with something more valuable than a rigid forecast:

an understanding of the policy-learning process.

This is the area in which Tetlow's warning deserves to be taken seriously.

Warsh should not confuse skepticism toward forward guidance with skepticism toward accountability.

The two are entirely different propositions.

IX. Commitment, Time Consistency, and the Limits of the Bayesian Critique

There is, however, an important qualification.

It would be too strong to argue that a central banker should always abandon previous guidance when new information arrives.

There are circumstances in which maintaining a commitment despite new information can be rational.

Commitment can help solve a time-consistency problem. It can influence expectations, affect credibility, and produce economic benefits that would not arise under pure discretion.

The Bayesian argument therefore should not be reduced to:

"Always abandon forward guidance when new information arrives."

That would be as simplistic as the rigid guidance it criticizes.

The more defensible proposition is:

A central banker who treats previous forward guidance as binding even when sufficiently powerful new evidence has materially changed the posterior distribution is failing to distinguish commitment from evidence-based policy.

That statement is considerably stronger theoretically because it recognizes that commitment itself can have value.

And the converse is equally important:

A central banker who changes policy whenever the latest data point changes is also not practicing Bayesian policy.

Bayesian policy lies between dogmatism and data-point opportunism.

It requires:

commitment when commitment has genuine economic value;

flexibility when the information set materially changes the posterior distribution;

and

transparency about why the posterior has changed.

X. Two Conceptions of Monetary-Policy Credibility

The disagreement between Tetlow's perspective and the Bayesian perspective can ultimately be expressed as two different conceptions of credibility.

The conventional framework

Stable reaction function → transparent communication → predictable expectations → effective transmission.

This framework emphasizes predictability.

The Bayesian-adaptive framework

Information → belief updating → reassessment of models and risks → conditional policy → transparent explanation of the update.

This framework emphasizes epistemic adaptability.

Under ordinary conditions, the two frameworks can coexist.

Indeed, there is no reason why they should be regarded as mutually exclusive.

But under radical uncertainty, structural change, or model instability, they can diverge sharply.

The central difference is therefore not whether the central bank should be transparent.

It is what it should be transparent about.

The conventional approach emphasizes the policy reaction function.

The Bayesian approach emphasizes the information, beliefs, risks, and updating process that generate the policy decision.

XI. The Deeper Epistemological Issue

The most consequential difference may therefore lie in the treatment of uncertainty itself.

Tetlow's argument appears to treat uncertainty principally as uncertainty about the future values of economic variables.

A Bayesian central banker must confront a deeper problem:

uncertainty about the model generating those variables.

This distinction is fundamental.

Suppose inflation unexpectedly rises.

A conventional interpretation might ask:

How should the reaction function respond to higher inflation?

A Bayesian interpretation asks several questions first:

Is the inflation increase persistent?

Is it demand-driven or supply-driven?

Has the Phillips curve changed?

Has inflation expectations changed?

Has productivity changed?

Is this a temporary geopolitical shock?

Is the economy entering a different regime?

Has the estimated natural rate changed?

Is the historical model still reliable?

Only after those questions are considered does it make sense to determine the appropriate policy response.

Thus, the policy rule is downstream from the process of inference.

This is the central epistemological point that, in my view, is missing from much of the conventional debate over forward guidance.

XII. From Forward Guidance to Bayesian Conditional Guidance

The real alternative to conventional forward guidance is therefore not silence.

Nor is it a mechanical Taylor rule.

Nor is it simply a refusal to communicate.

The more promising alternative is what I would call Bayesian conditional guidance.

Under this approach, the central bank communicates four things:

First, its current assessment of the economy.

Second, the distribution of risks surrounding that assessment.

Third, the evidence that would cause it to revise its assessment.

Fourth, the broad policy implications of alternative states of the world.

This creates a communication framework that is simultaneously transparent and adaptive.

It tells markets:

We are not promising you a predetermined future.

But it also tells them:

We are not acting arbitrarily.

Instead:

Here is what we currently believe, here is why we believe it, here is what could change our minds, and here is how policy would respond if it did.

That may be the most credible form of communication available under radical uncertainty.

XIII. A Bayesian Defense of Warsh—With an Important Caveat

From this perspective, Warsh's skepticism toward conventional forward guidance is intellectually defensible.

Indeed, the historical experience cited by Tetlow himself demonstrates why policymakers can become uncomfortable with guidance that later constrains their ability to respond to changing circumstances. His discussion of the Federal Reserve's reluctance to invoke an escape clause during the inflation surge of 2021 illustrates the potential cost of excessive commitment.

But Warsh should be careful.

The rejection of forward guidance should not become a rejection of epistemic transparency.

A central banker does not need to tell markets precisely where interest rates will be six months from now.

But the central banker should explain:

• what the FOMC currently believes;

• what evidence supports that belief;

• what uncertainties surround it;

• which developments could change it; and

• how policy would respond to materially different states of the economy.

That is not forward guidance in the conventional sense.

It is transparent Bayesian conditionality.

XIV. Conclusion: Credibility Through Learning, Not Through Inertia

Robert Tetlow's essay is a valuable contribution to the debate over Federal Reserve communication. His distinction between Odyssean and Delphic forward guidance is useful, and his insistence that rejecting forward guidance should not become an excuse for abandoning transparency is entirely reasonable.

Where I respectfully part company with Bob is in the implicit epistemological architecture underlying the argument.

The monetary-policy reaction function is not necessarily a fixed object that markets can simply learn.

It is conditional on the central bank's information set, its model of the economy, its estimates of latent variables, and its assessment of risks. Those beliefs can change. The models can change. The parameters can change. And, in periods of structural transformation, the probability assigned to competing models can change.

Consequently, a serious central banker should never regard previous forward guidance as an obligation to ignore sufficiently powerful new Bayesian evidence.

Doing so would confuse credibility with consistency.

The credibility of an independent central bank should rest on something deeper:

A central bank's credibility should not derive from its willingness to remain faithful to forecasts that have been invalidated by new evidence. Its credibility should derive from demonstrating that it can distinguish commitment from dogmatism, update its beliefs when the evidence warrants it, and explain those updates transparently to the public and to financial markets.

This does not mean that every new data point should produce a policy reversal. Bayesian updating is neither mechanical nor instantaneous. Evidence must be evaluated according to its reliability, persistence, relevance, and consistency with competing explanations.

Nor does it mean that commitment has no value. On the contrary, commitment can be economically valuable when it solves a time-consistency problem or stabilizes expectations.

The appropriate principle is therefore neither "always commit" nor "always adapt."

It is:

Commit when commitment has economic value; adapt when sufficiently strong evidence changes the posterior distribution; and explain transparently why the assessment has changed.

In this sense, the true alternative to forward guidance is neither silence nor a mechanical reaction function.

It is Bayesian conditional guidance:

Communicate the information set, the distribution of risks, the current posterior assessment, and the conditions under which that assessment would change.

This approach reconciles the legitimate institutional demand for transparency emphasized by Tetlow with the epistemic flexibility emphasized by Warsh.

Most importantly, it recognizes a fundamental reality of modern monetary policymaking:

The central bank does not merely respond to the economy. It is continuously learning about the economy.

And when the learning changes, policy must retain the freedom to change with it.

That is not a failure of credibility.

Under genuine uncertainty, it may be the highest form of credibility.

A Bayesian Game Theoretic Study of Global Geostrategic and Political Economy Consequences Plus Their Regional Ramifications Until 2030

The September 2026 Trump–Xi Meeting, the December G20 Miami Summit, and the Emerging Architecture of Strategic Competition

Farid Novin

Analytical Report Prepared for the G20

Information and data cut-off: August 11, 2026

Executive Summary

The anticipated September 24, 2026 meeting between United States President Donald J. Trump and Chinese President Xi Jinping in Washington, followed by the United States-hosted G20 Leaders' Summit at Trump National Doral in Miami on December 14–15, forms one of the most consequential diplomatic sequences of the year. President Trump confirmed the September date directly, noting that the visit will fall during the week of the United Nations General Assembly, when Xi — who has attended the General Assembly only once since taking office in 2012 — will travel to Washington rather than New York. The significance of the encounter lies not in the possibility of a single trade agreement but in the fact that it will occur inside a rapidly reorganizing international system in which artificial intelligence, energy security, critical minerals, tactical and strategic nuclear risk, space competition, Taiwan, the war in Ukraine, the future of the Middle East, the Belt and Road Initiative, BRICS financial coordination and the emerging American doctrine of hemispheric primacy in Latin America are becoming inseparable from one another.

The May 2026 Trump–Xi summit in Beijing produced an important but incomplete stabilization mechanism. The two governments agreed to establish Boards of Trade and Investment, China committed to substantial annual purchases of American agricultural products, Beijing agreed to address American concerns regarding critical minerals, and the two leaders articulated a shared interest in strategic stability, including a joint statement that Iran should not possess a nuclear weapon and that the Strait of Hormuz should reopen. These commitments should not be read as the emergence of a stable U.S.–China condominium. They are better understood as signals in a Bayesian strategic game in which each side is continuously updating its beliefs about whether the other is pursuing genuine accommodation, tactical delay, or long-term strategic displacement.

The evidence accumulated since May points toward the latter interpretation being only partially correct. Barely five weeks after the Beijing summit, China placed two of the largest American recipients of federal rare-earth investment — MP Materials and USA Rare Earth — back onto its export-control blacklist, demonstrating how quickly a bilateral truce can be narrowed even as broader diplomatic engagement continues. Simultaneously, the People's Bank of China released, on August 10, 2026, a five-year reform blueprint for 2026–2030 that formally prioritizes renminbi internationalization, expansion of the Cross-Border Interbank Payment System, and the development of Shanghai and Hong Kong as competing international financial centers. Both developments confirm the report's central analytical premise: cooperation has become an instrument for managing rivalry, not a replacement for it.

Artificial intelligence has emerged as the dominant substantive theme of the September dialogue itself. Chinese commentary attributes the timing of the summit in part to Washington's concern that Beijing is closing the frontier AI gap — a concern reinforced by China's forging of AI partnerships with twenty-eight countries, mostly in the Global South, at the World Artificial Intelligence Conference in Shanghai in July, and by the rapid rise of Chinese open-weight models. By mid-2026, Chinese models accounted for a majority of tokens processed on major model-routing platforms, and independent U.S. government evaluation found the gap between the Chinese and American technological frontier narrowing to months rather than years. President Trump has himself described AI as the decisive technology of the era, stating that whichever country wins the AI race "is probably going to win" more broadly — language that elevates the September meeting from a trade negotiation into a strategic-technology summit.

The report identifies four principal trajectories through 2030: Managed Strategic Competition, assessed as the most probable outcome; Asymmetric Decoupling; Strategic Bloc Formation; and a low-probability but catastrophic Systemic Fracture arising from the interaction of several crises at once. As of August 11, 2026, several of those crises are not hypothetical but active. The United States and Israel have been at war with Iran since February 2026; the Strait of Hormuz has been effectively disrupted for most of the past five months; Taiwan has just completed its largest-ever Han Kuang military exercises amid explicit warnings from Beijing that the island could become a flashpoint; Israel's prime minister has publicly rejected the Trump administration's own fifteen-point Gaza framework days before this report's cut-off date; and the United States has, for the first time in the modern era, used direct military force in Latin America to remove a head of state, inaugurating what American officials themselves now call the "Donroe Doctrine." None of these developments make a comprehensive U.S.–China settlement more likely in September. They make the alternative case more urgent: the world does not require a grand bargain between Washington and Beijing to avoid systemic instability. It requires guardrails. The December G20 Summit — whose official American agenda is deliberately narrow, organized around deregulation, energy affordability and technological innovation — should nonetheless be used as the venue in which selected elements of bilateral U.S.–China stabilization are converted into multilateral safeguards on crisis communication, energy security, financial-system interoperability, AI safety and nuclear risk reduction. The objective is not to eliminate rivalry. It is to make rivalry predictable, bounded and survivable.

I. The Analytical Premise: A Bayesian Game Under Radical Uncertainty

The relationship between Washington and Beijing cannot be adequately understood through a conventional balance-of-power model. Both governments possess extensive intelligence capabilities, yet neither can know with certainty the strategic intentions, domestic constraints or future risk tolerance of the other. The essential uncertainty is therefore not simply uncertainty about facts. It is uncertainty about type. Washington must assess whether Beijing's economic concessions represent genuine stabilization, tactical delay, or preparation for a stronger future position. Beijing must make the corresponding judgment about Washington. Is President Trump's transactional diplomacy intended to build a durable strategic accommodation with China, or is economic negotiation being used to extract concessions while the United States simultaneously strengthens military, technological and alliance pressure through export controls, hemispheric consolidation in Latin America and expanded security partnerships across the Indo-Pacific? This is a Bayesian strategic environment in the fullest sense. Each side begins with prior beliefs shaped by decades of rivalry. New evidence then modifies those beliefs, but the same observable action can support multiple, mutually inconsistent interpretations. The creation of bilateral Boards of Trade and Investment, China's agricultural purchasing commitments and the leaders' joint declaration on Iranian nuclear weapons and Hormuz all constitute positive evidence for accommodation. Yet within weeks, China's decision to blacklist two politically significant American rare-earth firms, and its continued build-out of alternative financial and payment infrastructure through BRICS and the digital renminbi, constitute contradictory evidence. The correct synthesis is not that cooperation has replaced rivalry, but that cooperation has become the chosen instrument for managing rivalry — a distinction that governs the interpretation of every subsequent development analyzed in this report.

II. The Strategic Baseline Entering September

The September meeting will occur in a substantially more fragmented environment than existed even a year earlier. The United States and China remain economically interdependent, but supply chains are being reorganized around security considerations rather than efficiency alone. China is simultaneously attempting to expand exports and reduce its vulnerability to foreign technological restriction; Chinese exports have continued to rise even as domestic demand remains comparatively weak, a pattern some analysts describe as a second "China Shock," increasingly centered on electric vehicles, batteries, green technology and, now, artificial intelligence. This produces a structural paradox. China needs continued access to Western markets and capital; the United States needs Chinese manufacturing capacity, processed minerals and components. Yet each government increasingly regards dependence upon the other as a strategic vulnerability rather than a source of mutual gain. The result is neither conventional globalization nor complete decoupling. It is competitive interdependence — and it is the strategic environment in which the September summit, and the report that follows, must be understood. III. The September Trump–Xi Meeting: What Is Actually Known President Trump confirmed on multiple occasions following the May Beijing summit that he intended to host President Xi in Washington in the fall, and by mid-summer had settled on September 24 as the date, coinciding with the week of the United Nations General Assembly. This is now the best-documented and most operationally concrete element of the bilateral calendar, reported consistently across American, Chinese and European coverage. Reuters reported in July that the September encounter would include the first official U.S.–China AI dialogue of the Trump administration, to be led on the American side by Treasury Secretary Scott Bessent — an institutional choice that itself signals Washington's intention to treat frontier technology as inseparable from macroeconomic and financial strategy. The meeting inherits the substantive architecture created in Beijing in May: a bilateral Board of Trade intended to manage commerce in non-sensitive goods; a Board of Investment; a multi-year Chinese commitment to American agricultural purchases; an initial Chinese order for Boeing aircraft; a shared Chinese and American position that Iran should not acquire a nuclear weapon; and a joint call for the reopening of the Strait of Hormuz without tolls imposed by any country or organization. These commitments matter because they create repeated interaction. Repeated games are structurally more conducive to cooperation than one-shot confrontations, because today's behavior shapes tomorrow's credibility. The September meeting is therefore, among other things, a test of whether the May commitments generated sufficient credibility to sustain further cooperation — a test complicated by China's mid-June decision to reimpose targeted export controls on American rare-earth companies, discussed in Section XIII.

IV. The Central Bayesian Question: Accommodation or Strategic Delay?

The most consequential uncertainty facing the September summit remains the interpretation of concessions already made. Washington may read China's agricultural and Boeing commitments as evidence that Beijing will accept durable economic constraints. Beijing may read the same commitments as the price of maintaining market access while China accelerates technological self-sufficiency. Both interpretations remain rational, which is the essence of a game of incomplete information: the same observable action produces different posterior beliefs depending on each side's priors. The evidence available as of August 11 favors a synthesis of selective accommodation rather than either pure interpretation. Both powers cooperate where the costs of confrontation are highest — energy markets, agricultural trade, and, cautiously, nuclear-proliferation risk in Iran — while continuing to compete intensely wherever the expected strategic return from competition remains substantial, most visibly in critical minerals, semiconductors, frontier AI and the Western Hemisphere. V. Artificial Intelligence: The Central Strategic Technology

Artificial intelligence is now the most consequential single subject shaping the September dialogue, not merely one item among many. AI connects economics, military power, finance, intelligence, cyber capability and scientific productivity in ways no other technology does, and it is the domain in which the balance of perceived advantage has shifted most rapidly in 2026. The United States retains major advantages in frontier computing, semiconductor design, cloud infrastructure and capital markets. But China's progress in efficient, lower-cost open-weight models has altered Washington's strategic calculus. By mid-2026, Chinese models — led by DeepSeek, Alibaba's Qwen family, Zhipu AI's GLM series and Moonshot AI — accounted for a majority of tokens processed on major third-party model-routing platforms, and Qwen alone had surpassed one billion downloads, forming the base of a large share of new derivative open-weight models. Independent U.S. government technical evaluation found the capability gap between the Chinese and American frontier narrowing from years to months: one leading Chinese model was assessed in mid-2026 as roughly comparable in overall performance to an American frontier model released only about six months earlier. This progress has occurred despite, and to some extent because of, American export controls. Chinese engineers, constrained in access to the most advanced chips, have directed disproportionate effort toward algorithmic efficiency and inference optimization, extracting frontier-level performance from lower-tier and domestically produced hardware. Reporting on leaked comments from a leading Chinese AI executive indicates that compute scarcity — not algorithmic capability — remains the binding constraint on China's largest models, with domestic chip supply still running at a fraction of what leading Chinese labs say they require. The strategic implication is that technology controls can slow China's access to the highest tier of frontier computing, but cannot easily prevent the country from developing competitive alternative pathways, particularly in efficient, widely distributed open-weight models that are reshaping AI adoption across the Global South. Beijing's own posture toward its AI ecosystem is also shifting in a way directly relevant to the September summit. Reuters reported in July that Chinese authorities held meetings with Alibaba, ByteDance and Zhipu AI regarding possible new restrictions on foreign access to China's most advanced models, including models not yet released and some open-weight systems — a move that, if implemented, would mark a significant reversal of Beijing's open-model strategy and suggests China itself is beginning to treat frontier AI diffusion as a strategic asset to be controlled rather than freely exported.

The Bayesian equilibrium likely to emerge from September is therefore selective technological decoupling rather than either unrestricted technological isolation or full technological integration. The most sensitive technologies — advanced chips, military AI, semiconductor manufacturing equipment and dual-use computational systems — will remain heavily restricted on the American side, while an increasingly Chinese-controlled tier of commercial open-weight models spreads through markets the United States has been slower to reach. The September summit could plausibly produce a limited understanding on AI safety, cyber stability and human control over autonomous military systems, without resolving the underlying contest for technological leadership. VI. The AI–Nuclear Nexus

One of the most dangerous developments through 2030 may not be an AI war as such, but the incorporation of AI into nuclear decision-support systems. SIPRI's 2026 research finds that China, India and the United States are all examining or integrating AI into nuclear-related systems, even as the United States and China have separately expressed rhetorical support for preserving human control over nuclear-use decisions. This creates a new category of Bayesian danger: during a crisis, each government may have incomplete information about whether the other side's AI systems are purely analytical or capable of influencing command-and-control processes. A system one side considers defensive may appear offensive to the other, particularly when AI is combined with cyber operations, satellite surveillance, hypersonic weapons and automated early-warning systems. There is now an emerging international normative anchor for addressing this risk. In mid-2026, more than two hundred global leaders and former officials endorsed a declaration — informally described in reporting as "Humanity at the Threshold" — calling for prohibitions on fully autonomous recursive AI self-improvement and on AI involvement in nuclear launch decisions. While this declaration carries no binding legal force, it demonstrates that the principle Washington and Beijing might adopt bilaterally in September already has meaningful civil-society and expert-community support, lowering the political cost of formal adoption. Even if a comprehensive arms-control agreement remains impossible, an explicit U.S.–China statement that neither government will permit AI systems to make or materially influence nuclear-use decisions would constitute a modest but historically significant confidence-building measure — one that the two governments could plausibly agree to at comparatively low cost, since neither currently claims to seek fully autonomous nuclear command. VII. Nuclear Weapons and the Tactical-Nuclear Threshold

The nuclear dimension of the international system has deteriorated in structurally important ways. SIPRI's Yearbook 2026, released June 8, estimates the global nuclear inventory at approximately 12,187 warheads as of January 2026, with about 9,745 held in military stockpiles for potential use and roughly 4,012 deployed with missiles and aircraft. Russia and the United States continue to hold close to ninety percent of the world's nuclear weapons — roughly 5,420 and 5,042 warheads respectively — while China's arsenal has grown to an estimated 620 warheads, an increase of twenty in a single year, with SIPRI noting that Beijing may now be deploying a small number of warheads on mobile launchers during peacetime exercises for the first time. SIPRI projects that China could possess as many intercontinental ballistic missiles as either Russia or the United States by the end of the decade. The expiration of New START in February 2026 removed the last binding bilateral U.S.–Russian strategic arms-control framework, and no successor agreement is currently under negotiation. This creates an unusual strategic triangle in which the United States must weigh Russian and Chinese capabilities simultaneously, Russia remains focused on the United States and NATO, and China increasingly weighs American power in the Indo-Pacific against the long-term survivability of its own still-modernizing nuclear forces. The most dangerous pathway is not necessarily deliberate strategic nuclear war but nuclear escalation emerging out of conventional conflict, and by August 2026 two live theaters illustrate the risk directly.

Ukraine represents the clearer European pathway. SIPRI documents that Russia has continued fielding its dual-capable Oreshnik intermediate-range ballistic missile, including conventional-warhead strikes against Ukraine as recently as May 2026, and has begun constructing a forward-operating base for the system in Belarus. A major Russian battlefield reversal, an expansion of direct NATO involvement, or a perceived threat to Russian strategic assets could each alter Moscow's nuclear calculus, even absent any intention to use a weapon. The Middle East represents the second and, as of this writing, the more acute pathway. The war between the United States and Israel, on one side, and Iran, on the other, has continued since February 2026, and the shared Trump–Xi declaration in May that Iran must not acquire a nuclear weapon establishes a rare area of explicit U.S.–China convergence. Yet the current conflict — addressed in detail in the following section — demonstrates how rapidly a proliferation-focused understanding can become entangled with a live regional war involving direct strikes, naval blockade and a contested maritime chokepoint. The September summit should therefore treat nuclear risk reduction as both a direct and an indirect priority. The most realistic objective is not a comprehensive three-power nuclear treaty but incremental risk reduction: reinforced crisis-communication channels covering Taiwan, space assets, cyberattacks and nuclear signaling, alongside the narrower AI-nuclear commitment discussed above.

VIII. The Strait of Hormuz: The Most Immediate Global Economic Shock

The Strait of Hormuz has moved from a background risk factor to the most acute and continuously live crisis shaping the September calculus. The United States and Israel struck Iranian military targets, including the killing of Iran's Supreme Leader, in late February 2026, opening a war that has now continued for more than five months. Iran responded by attempting to close the Strait to foreign shipping and, in May 2026, establishing a self-declared "Persian Gulf Strait Authority," asserting that no vessel may transit without a passage permit issued by Tehran — a claim that runs counter to the strait's actual geography, since its navigable channels lie in both Iranian and Omani territorial waters, and one that Oman itself has pointedly declined to endorse. A June 17, 2026 memorandum of understanding between the United States and Iran briefly increased crossings, but the arrangement collapsed within weeks after renewed attacks on shipping in Omani waters on June 25 and again on July 7–8, each followed by American airstrikes on Iranian targets. By mid-July, the tempo of attacks and counterattacks exceeded any point since April, again suppressing cross-strait traffic. As of early August, the conflict has abated somewhat amid a fresh round of Iran–Oman negotiations, which Omani and Iranian officials both publicly described in the first week of August as "very close" to a new arrangement — though Iran's Foreign Minister explicitly conditioned any deal on additional concessions, including compensation from Washington and the lifting of the American naval blockade of Iran that has been in place since spring. Oil markets have moved accordingly: West Texas Intermediate has traded in a wide band around the high seventies to low eighties per barrel through the first half of August, spiking on individual attack reports and falling on signs of progress in the Oman talks, while the U.S. Energy Information Administration has raised its full-year Brent forecast into the mid-to-high eighties and now expects a portion of the roughly 5.5 million barrels per day of regional production shut in during the conflict to remain offline into 2027. This context reframes the strategic significance of the May Trump–Xi understanding on Hormuz. Washington and Beijing agreeing, in principle, that the strait should reopen without tolls does not translate automatically into Iranian compliance, since Iran is not a party to that understanding and has repeatedly linked reopening to conditions the United States has not yet met. The Hormuz problem is therefore not a single bilateral game but at least three layered games occurring simultaneously: a U.S.–China game focused on preventing energy disruption from becoming a source of bilateral strategic confrontation; a U.S.–Iran game centered on whether sanctions, military pressure, naval blockade and diplomacy can jointly produce a durable reopening; and a China–Iran game in which Beijing seeks to preserve access to Iranian and Persian Gulf energy while avoiding direct confrontation with Washington. China, as one of the largest purchasers of oil transiting the strait, has a particularly strong structural incentive to see the corridor stabilized, which is precisely the kind of shared material interest around which conditional U.S.–China cooperation is most likely to form even amid broader rivalry. IX. Energy Beyond Oil: The

The energy competition between Washington and Beijing is no longer simply about petroleum. It increasingly concerns an integrated technological energy system spanning oil and liquefied natural gas, nuclear power, electricity grids, battery storage, critical minerals, solar manufacturing, electric vehicles, hydrogen and the enormous and rapidly growing electricity demand of artificial-intelligence data centers. China's structural strength in batteries, solar manufacturing, electric vehicles and related supply chains provides Beijing with a form of leverage distinct from, and in some respects more durable than, traditional oil power. The United States retains major advantages in hydrocarbons, capital markets, advanced semiconductor-linked technology and, increasingly, in efforts to expand domestic and allied energy production to power its own AI buildout.

The competition is therefore evolving from a petroleum-centered order toward an electro-technological order in which access to processed minerals, battery chemistry, grid technology and computing power all function as forms of energy security. This creates a genuine opportunity for the December G20: energy security should be redefined at Miami to include not only access to fuel but access to the industrial technologies required to generate, store, transmit and consume energy at the scale the AI era demands. X. Taiwan: The Most Dangerous U.S.–China Regional Game

Taiwan remains the central military flashpoint in the relationship, and the August 2026 evidence base is unusually rich. Taiwan conducted its largest-ever Han Kuang exercises from August 5 to 14, explicitly designed, according to senior Taiwanese defense officials, to move the drills closer to real combat scenarios rather than scripted demonstrations. The exercises emphasized decentralized command and control — enabling commanders in the field to make operational decisions without waiting for orders from a potentially disrupted central command — alongside drone "kill-chain" operations combining reconnaissance and attack drones, urban resilience training, and civil-defense measures including deliberate, temporary slowdowns of mobile internet service and air-raid drills designed to expose the civilian population to the disruption a real conflict would bring. Taiwan is separately planning a substantial increase in 2027 defense spending, with proposals exceeding three percent of GDP. These developments carry a clear Bayesian interpretation. Taipei is signaling resolve and, through the decentralization of command, signaling that it is preparing specifically for a scenario in which Chinese action targets Taiwan's central leadership or communications. Washington is signaling that Taiwan must bear a larger share of its own defense burden. Beijing, for its part, interprets increased Taiwanese and American military preparedness as potential confirmation of its concern that Taiwan's political trajectory is moving toward permanent separation — a concern President Xi reportedly raised directly and explicitly with President Trump during the May Beijing summit, telling him that Taiwan could become a flashpoint between the two countries if not "handled properly." The danger is therefore not simply intentional aggression by either side but the mutual interpretation of defensive preparation as offensive preparation — a classic security-dilemma dynamic sharpened by the specific character of this year's exercises, which are unusually explicit in preparing for command disruption rather than merely demonstrating deterrent capability. The September summit should seek to reinforce several principles: that neither side use military exercises to create ambiguity about immediate offensive intent; that crisis communication channels be strengthened and tested rather than merely declared; that the distinction between deterrence and provocation be preserved in both countries' public signaling; and that economic coercion against Taiwan be treated as a component of escalation risk in its own right, not merely as a trade matter. The most probable trajectory through 2030 remains a tense status quo rather than invasion, but the probability of crisis will rise if military encounters become more frequent or if domestic political narratives in Washington, Beijing or Taipei further narrow the space for compromise.

XI. Taiwan and the Semiconductor System Taiwan cannot be separated from the AI competition described in Section V. Its semiconductor industry gives the island an extraordinary position in the global technological system, producing what might be called the Taiwan paradox: the more strategically important Taiwan becomes to advanced computing, the stronger Washington's incentive to protect it — but the more deeply Taiwan is incorporated into the U.S.-led technological system, the more Beijing may perceive semiconductor dependence itself as a strategic vulnerability requiring resolution. The optimal international strategy is therefore not simply to defend semiconductor production in Taiwan, but to build geographically diversified semiconductor capacity without making Taiwan economically irrelevant in the process. A more resilient global semiconductor system reduces the incentive for any actor to treat military control of Taiwan as the only means of controlling advanced computing — a strategic logic that should inform both American reshoring policy and G20-level discussion of supply-chain resilience. XII. Tariffs: From Trade Instrument to Strategic Weapon

Tariffs have evolved from instruments of trade policy into instruments of strategic bargaining. The May summit demonstrated that Washington can simultaneously threaten economic pressure and negotiate commercial concessions; China's multi-year agricultural purchasing commitment is the clearest example of this transactional equilibrium. The United States Trade Representative has continued to describe the bilateral Board of Trade explicitly as a mechanism for pursuing balance and reciprocity while retaining tariffs as instruments of both economic and national-security policy — language that signals tariffs will remain a standing feature of the relationship rather than a temporary crisis measure. China, in turn, continues to use export controls and supply-chain leverage, particularly in rare earths and other critical minerals, as its own strategic counterweight. The result is a strategic exchange in which Washington possesses tariff power and Beijing possesses supply-chain power, and neither instrument is costless: American tariffs raise prices and can redirect supply chains away from China over time, while Chinese export controls accelerate the very American and allied diversification efforts Beijing would prefer to prevent. This is a classic repeated-game structure — each side can impose costs on the other, but excessive or repeated use of coercion accelerates the counterparty's ability to escape dependence altogether. The rational equilibrium emerging from this dynamic is controlled economic coercion rather than unrestricted economic warfare, with both sides calibrating pressure to extract concessions without triggering an irreversible break. 

XIII. Critical Minerals: China's Structural Leverage

Rare earths remain the single clearest illustration of the gap between diplomatic language and strategic reality in the U.S.–China relationship. China commands roughly ninety percent of global rare-earth refining capacity and close to ninety-five percent of permanent-magnet production, and its October 2025 export-control regime — suspended for one year following the Busan summit as part of a broader tariff and trade-barrier rollback — was modeled explicitly on the extraterritorial reach of the American Foreign Direct Product Rule, requiring a license for any foreign-made product containing even trace amounts of Chinese-origin rare earths or processing technology.

The May Beijing summit raised hopes that this truce would be extended and deepened, with China committing to address American concerns over shortages and restrictions involving critical minerals and processing technology. That hope proved short-lived. Only five weeks after the Beijing summit, on June 22, 2026, China added MP Materials and USA Rare Earth — the two largest recipients of federal investment in American rare-earth independence, having received $550 million and $1.6 billion respectively — to its export-control blacklist, citing their links to the U.S. military-industrial base. Because China's controls carry extraterritorial reach, the blacklisting restricts not only direct exports to the two firms but any transfer of Chinese-origin inputs to them by any entity anywhere, reaching deep into allied supply chains. The episode functions as a textbook Bayesian signal: it does not indicate abandonment of the broader trade détente, but it demonstrates decisively that promises of supply are not equivalent to secure supply, and that Beijing retains, and is willing to use, its most powerful instrument of economic leverage even while pursuing a parallel diplomatic track with Washington.

The United States and its partners have continued building alternative capacity in response, though on a timeline measured in years rather than months. Washington and Australia signed a Critical Minerals Framework in October 2025 to leverage and expand mining and processing capacity, with Australia — the world's fourth-largest rare-earth producer and the recipient of nearly half of global rare-earth exploration investment in 2024 — emerging as the most important American partner in the sector, including a government-backed loan supporting a major new refinery tied to allied offtake. The United States has separately deepened cooperation with Japan through a bilateral critical-minerals framework, a border-adjusted price-floor initiative, and a deep-sea mining research agreement. New magnet-manufacturing capacity outside China began coming online in the summer of 2026, but self-sufficiency remains, by the assessment of specialists in the field, a long road rather than a near-term outcome. The G20's appropriate objective is therefore diversification rather than an unrealistic goal of eliminating Chinese participation in global mineral markets altogether; a genuinely diversified system reduces the incentive for any single actor to weaponize supply.

XIV. BRICS and the Emerging Financial Architecture

BRICS should not be interpreted as a conventional military alliance. Its strategic significance lies in providing an institutional environment in which emerging economies can reduce their dependence on Western financial infrastructure without necessarily abandoning the dollar altogether. The bloc's 2026 summit, hosted by India in New Delhi in September — occurring within days of the Trump–Xi meeting — has placed payment-system interoperability at the center of its agenda. India's central bank has led technical coordination on "BRICS Pay," conceived as a decentralized alternative to SWIFT linking Russia's SPFS, China's CIPS, India's UPI and Brazil's Pix, alongside a parallel proposal to interconnect the bloc's central-bank digital currencies — the digital yuan, India's e-Rupee, Brazil's DREX and the UAE's dirham CBDC. Reserve Bank of India Governor Sanjay Malhotra confirmed in early August that these discussions remain at an early, exploratory stage rather than having produced an operational unified system, but the direction of travel is unambiguous: faster, cheaper cross-border settlement that reduces reliance on dollar-denominated correspondent banking, pursued through practical infrastructure rather than symbolic calls for a single BRICS currency.

China is pursuing a closely related but distinct strategy. On August 10, 2026 — one day before this report's data cut-off — the People's Bank of China released a five-year reform and development plan for 2026–2030 that formally commits to "prudently advance high-level financial opening-up," explicitly naming expansion of the renminbi's international role in trade, investment and financing, deepened two-way market access, continued build-out of the Cross-Border Interbank Payment System, development of offshore renminbi markets, and the parallel positioning of Shanghai and Hong Kong as competing international financial centers. Digital-yuan development is named as a standing priority within the same plan.

None of this indicates that the renminbi will displace the dollar as the world's primary reserve currency by 2030; the dollar retains overwhelming advantages in liquidity, financial depth, legal predictability and the sheer size of global capital markets that trade and settle in it. The more plausible scenario is monetary diversification: the global financial system will remain dollar-centered while becoming measurably less exclusively dollar-dependent, particularly for trade among emerging economies. This distinction carries a direct policy implication for Washington. Excessive or indiscriminate use of dollar-based financial sanctions risks accelerating precisely the diversification the United States seeks to prevent, by validating the strategic logic — already well underway on both the Chinese and Indian sides — of building durable, dollar-independent settlement infrastructure before it becomes urgently necessary.

XV. The Belt and Road Initiative: From Infrastructure to Strategic Ecosystem

The Belt and Road Initiative has also changed in character. Its importance can no longer be measured simply by the number of roads, ports or railways constructed; its strategic weight increasingly lies in the creation of an integrated ecosystem connecting infrastructure, energy, digital networks, logistics, industrial production, finance, critical minerals and markets into a single interoperable system that reduces recipient countries' structural dependence on Western-controlled trade and financial channels.

Latin America has become an increasingly important theater within this system, and — as the next section details — increasingly contested. China's economic relationship with the region gives Beijing access to agricultural commodities, energy, minerals and strategic markets, while giving Latin American governments an alternative to dependence on Washington. The United States now faces a genuine strategic dilemma: attempting to remove China entirely from Latin America may be economically unrealistic given the depth of existing commercial ties, while ignoring China's continued expansion risks long-term structural dependence in a hemisphere Washington regards as its own strategic neighborhood. As discussed below, the Trump administration has resolved this dilemma in 2026 not through competitive economic provision alone but through direct assertion of hemispheric primacy, including the use of military force.

XVI. The "Donroe Doctrine" and Latin America

The single most consequential and least anticipated development in the Western Hemisphere since the original drafting of this report is the emergence — and active implementation — of what American officials, Chinese scholars and international media now uniformly call the "Donroe Doctrine," a term blending President Trump's name with the 1823 Monroe Doctrine. The phrase, first coined in commentary in 2025, was embraced by Trump himself in January 2026 following the American capture of Venezuelan President Nicolás Maduro in a direct military operation, after which the President declared that "American dominance in the Western Hemisphere will never be questioned again." The doctrine was subsequently formalized in the administration's National Security Strategy as what officials internally describe as a "Trump Corollary" to the original Monroe Doctrine, explicitly directed at ensuring the Americas remain "free of hostile foreign incursion or ownership of key assets" — language aimed, without naming China directly, squarely at Chinese commercial and infrastructure investment across the hemisphere.

The practical consequences have moved quickly. Chinese access to critical-mineral projects in Brazil, Argentina, Chile and Bolivia has come under sustained pressure; the United States has expanded security and defense-sales relationships across the region, including fighter-jet sales to Peru and Argentina; Washington has reinforced its position on the sovereignty and security of Panama's ports — a flashpoint symbolically underscored by the controversial December 2025 demolition of a monument commemorating a century and a half of Chinese presence in Panama — and has for the first time established a dedicated regional space-force initiative explicitly framed as a counter to Chinese space strategy in the hemisphere. Chinese scholars affiliated with state research institutions have themselves characterized the doctrine as marking a shift from American influence exercised through the language of rules and values toward influence exercised more directly through force and power, and have concluded that Latin American governments now face simultaneous pressure to preserve strategic autonomy while navigating renewed great-power competition on their own soil.

The doctrine's implementation coincides with, and has been reinforced by, an independent rightward political shift across parts of the region, with Trump-aligned governments gaining power from Costa Rica to Peru and, most prominently, in Chile's 2026 election. This alignment gives Washington a more receptive political environment than it has had in the hemisphere in decades, but it also raises the risk of backlash: polling through 2026 shows a marked decline in favorable views of the United States in several Latin American countries, suggesting that even where governments align with Washington, public sentiment may not follow automatically. For China, the Bayesian problem is now unusually stark: Beijing must decide whether to contest American primacy in a region where the United States has demonstrated a willingness to use direct military force, or to accept a diminished but still commercially significant role and redirect strategic emphasis toward Africa, Southeast Asia and the Belt and Road's traditional Eurasian corridor. The evidence to date suggests Beijing is choosing the latter path — de-escalating direct confrontation in the hemisphere while preserving commercial relationships wherever politically possible — which itself constitutes an important data point regarding the limits Beijing currently places on great-power competition outside its immediate periphery.

XVII. Ukraine: The European Dimension of the U.S.–China Game

The war in Ukraine remains a critical variable even though China is not a formal belligerent. As of early August 2026, no ceasefire is in place. President Zelenskyy has stated publicly that Ukraine remains in daily contact with the Trump administration and expects an intensified diplomatic push through August, with envoys Steve Witkoff and Jared Kushner expected to play an active role alongside continued, partly undisclosed engagement with Persian Gulf partners. At the same time, Ukrainian officials have accused Russia of deliberately dragging out negotiations while intensifying long-range strikes — including reports in early August of sixty-one missiles launched against Ukraine in a single week, the majority on ballistic trajectories — and have expressed concern that Moscow may use Russia's own September parliamentary elections as a pretext or occasion for renewed mobilization, timed to increase pressure on both Kyiv and European capitals simultaneously.

For China, Ukraine represents a persistent balancing problem rather than a fixed position. Beijing continues to benefit economically from discounted Russian energy and from Russia's alignment against American strategic predominance, but it also has a clear interest in preventing the conflict from escalating into an uncontrollable European war involving NATO directly, or from generating a nuclear incident that would radically destabilize the international system Beijing depends on for trade and investment. Washington's objective differs: an end to the war that does not reward aggression while avoiding indefinite military and financial escalation.

China can therefore play a genuinely useful but deliberately ambiguous role. A formal joint U.S.–China peace initiative on Ukraine remains unlikely, but narrower convergence is plausible around preventing nuclear escalation, maintaining humanitarian channels, discouraging attacks on nuclear facilities, and preventing geographic expansion of the conflict. The decisive Bayesian variable remains Russia's own calculation: both Washington and Beijing must independently assess whether Moscow's incentives currently favor a negotiated settlement or continued war in pursuit of further battlefield gains. If both conclude that Russia's position favors continued fighting, coordinated pressure becomes more plausible; if either believes Moscow can still achieve meaningful gains, the incentive for restraint on all sides weakens correspondingly.

XVIII. Tactical Nuclear Weapons and Ukraine

The tactical-nuclear dimension of the Ukraine war deserves separate treatment because the distinction between tactical and strategic nuclear use is politically and psychologically far more fragile than the formal doctrinal categories suggest. A nuclear detonation in Ukraine, however limited in yield or intended purpose, would generate overwhelming pressure for retaliation, escalation and direct alliance involvement, collapsing the careful distinctions policymakers currently draw between tactical and strategic weapons. Russia's continued deployment of dual-capable systems such as the Oreshnik missile, documented in SIPRI's 2026 assessment, keeps this risk structurally present even absent any specific indication of imminent nuclear use.

For Beijing, such an event would be dangerous because it could destabilize the European economy, disrupt global trade and financial markets, and create irresistible pressure for China to align more explicitly with one side or the other — precisely the kind of forced binary choice Beijing has worked consistently to avoid. For Washington, it would represent a fundamental challenge to the post-Cold War nuclear order that neither American nor Chinese strategic planning currently assumes. The rational Bayesian objective for both governments is therefore to make tactical nuclear use an overwhelmingly unattractive strategy for Moscow specifically: private, coordinated signaling to Russia that any nuclear use would radically and permanently alter the international strategic environment, combined with functioning crisis-communication channels capable of operating even amid the confusion of an active nuclear scare.

XIX. Israel and the Middle East

The Middle East adds a distinct layer to the U.S.–China strategic game, one that has become considerably more volatile since the original drafting of this report. The United States remains the dominant external security actor for Israel and several Persian Gulf states; China has developed substantial economic relationships throughout the region and has attempted to position itself as a diplomatic alternative to Washington, though without displacing core American security guarantees.

The most significant recent development is the near-collapse of the Trump administration's own Gaza framework. Following an October 2025 ceasefire that established a U.S.-led "Board of Peace" — chaired for life by President Trump and formally designated under United Nations Security Council Resolution 2803 as the body responsible for overseeing the Gaza peace process — the administration announced on July 30, 2026 what it called a "historic" breakthrough, asserting that Hamas had agreed to the complete disarmament of its heavy weapons, production sites, depots and tunnel network under a new fifteen-point plan. Prime Minister Benjamin Netanyahu, however, publicly and unambiguously rejected that fifteen-point document on August 9, 2026, stating that Israeli forces would not withdraw from any additional Gaza territory until Hamas is "genuinely," not "fictitiously," disarmed, and reaffirming that Israel would not accept the establishment of a Palestinian state under his government. Israel continues to control roughly seventy percent of Gaza under the ceasefire's existing terms. Hamas, for its part, has stated it remains committed to the broader framework while calling on mediators and guarantor states to prevent violations that could derail it. Netanyahu's rejection lands less than three months before Israel's own October 27, 2026 election, in which polling indicates his governing coalition is at genuine risk of losing its majority — a domestic political constraint that is now inseparable from his negotiating posture toward Washington.

The strategic consequence for this report's framework is that the Middle East cannot be reduced to a simple U.S.–China contest, and increasingly cannot even be reduced to a simple U.S.–Israel alignment. Regional and domestic actors possess their own independent strategies and constraints: Iran seeks regime survival and continued strategic leverage amid an active war; Israel's government seeks both security guarantees and, under current leadership, the practical dismantling of Hamas's military capacity, even at the cost of visible friction with its principal security patron; Persian Gulf states seek economic stability and protection from regional spillover; China seeks energy security and continued commercial access without deep entanglement in the war itself; and the United States seeks to preserve both Israeli security and its own broader regional architecture, including the stabilization of Hormuz addressed in Section VIII. The resulting equilibrium is likely to remain genuinely multipolar within an overarching American security architecture, rather than resolving cleanly in any single direction before 2030.

XX. Iran, Nuclear Proliferation and China's Strategic Dilemma

The May Trump–Xi understanding that Iran should not obtain a nuclear weapon provides a rare area of genuine common strategic interest between Washington and Beijing. But the current war, and the unresolved status of the Strait of Hormuz described in Section VIII, demonstrates the practical limits of that convergence. China does not want Iran to become nuclear-armed, both because of the direct proliferation risk and because a nuclear-armed Iran could trigger a wider regional cascade, with Saudi Arabia, Türkiye and other regional powers reassessing their own nuclear options. At the same time, Beijing does not want Iran's economic collapse, nor a prolonged conflict that threatens the flow of Iranian and wider Persian Gulf energy on which China remains heavily dependent.

China's optimal strategy through 2030 is therefore likely to remain one of stabilizing without fully aligning: continued diplomatic pressure against Iranian nuclear weapons development, combined with preserved economic relationships and studied avoidance of direct involvement in the U.S.–Israel–Iran war itself. This is a strategy that will be tested repeatedly as the Hormuz situation evolves, since any prolonged closure directly threatens Chinese energy security in a way few other elements of the current crisis environment do.

XXI. Space: The Next Strategic Frontier

Space is an increasingly critical component of U.S.–China strategic competition. Satellites provide communications, navigation, intelligence, missile warning and targeting, meaning that any attack on space infrastructure could plausibly be interpreted as preparation for a larger military conflict rather than an isolated incident. The Bayesian danger here mirrors the AI-nuclear problem described in Section VI: if one country develops or demonstrates an anti-satellite capability, the other cannot easily determine whether it is intended for deterrence or first-strike preparation, and the same ambiguity applies to satellite inspection, close-proximity maneuvering and cyber operations directed at space assets. The danger is compounded when space systems are connected to nuclear command-and-control networks, a linkage SIPRI's 2026 Yearbook addresses directly in its dedicated treatment of space security governance alongside its nuclear-forces and AI chapters — itself a reflection of how closely these three domains have become intertwined in serious strategic analysis over the past year.

Latin America has also become an unexpected new front in this competition: as noted in Section XVI, the United States has established a dedicated regional space-force initiative explicitly designed to counter Chinese space activity in the Western Hemisphere, illustrating that the space competition is no longer confined to the Indo-Pacific and cislunar domains but is now actively regionalizing alongside the broader hemispheric contest.

The September meeting should consider a narrowly defined set of space confidence-building principles: advance notification of dangerous proximity operations; direct military-to-military communication channels dedicated to space incidents; non-interference with designated nuclear early-warning satellites; and a shared vocabulary distinguishing reversible signaling from irreversible attack. A comprehensive space treaty remains unlikely in the near term; a narrower space crisis-prevention regime, of the kind both governments have shown some willingness to discuss in other domains, is considerably more achievable and would meaningfully reduce the risk of inadvertent escalation.

XXII. The Convergence of AI, Space and Nuclear Risk

The greatest strategic risk through 2030 may come not from any single technology but from the interaction among artificial intelligence, space systems, cyber operations and nuclear weapons together. AI increases the speed of decision-making; space systems provide the information on which that decision-making depends; cyber operations can disrupt or corrupt the communications linking the two; and nuclear weapons provide catastrophic destructive capability at the end of the chain. Together, these systems compress the time available for human judgment precisely when judgment matters most. A false signal transmitted through a compromised satellite network could be interpreted by an AI-supported command system as evidence of an impending attack — a scenario the "Humanity at the Threshold" declaration referenced in Section VI was explicitly designed to address.

This is precisely the environment in which Bayesian updating can become dangerous rather than stabilizing. Rapid updating is not always rational updating: a government may receive ambiguous evidence, assign it excessive weight under crisis time pressure, and act before contradictory evidence becomes available. The G20 should therefore treat the preservation of human decision-time itself as a strategic public good — a principle that can be operationalized through the AI-nuclear commitment discussed in Section VI, the space confidence-building measures discussed in Section XXI, and reinforced crisis-communication infrastructure discussed throughout this report.

XXIII. Four Principal Scenarios Through 2030

Scenario One: Managed Strategic Competition — approximately 45 percent

This remains the baseline scenario. The September meeting produces a political reaffirmation and modest expansion of the May framework. The Boards of Trade and Investment become functional rather than symbolic. China continues agricultural purchases despite periodic frictions such as the June rare-earth blacklisting episode, and critical-mineral disputes are managed through licensing and negotiated access rather than escalating into a full embargo. Tariffs remain a standing feature of the relationship but are selectively modified rather than expanded. AI restrictions remain concentrated on genuinely sensitive military and dual-use technologies while commercial AI markets remain partially interconnected. Taiwan remains tense — including further large-scale exercises on both sides — but avoids open conflict. Hormuz gradually stabilizes through the Iran–Oman negotiating track even without full American–Iranian reconciliation. Ukraine moves toward negotiations or a partially frozen conflict rather than either decisive victory or nuclear escalation. BRICS expands financial cooperation and payment interoperability without displacing the dollar's central role. The United States and China continue competing throughout Latin America under the Donroe Doctrine but avoid direct military confrontation with each other. Under this scenario, the global economy remains fragmented but integrated, and the world enters a period best described as managed multipolarity.

Scenario Two: Asymmetric Decoupling — approximately 30 percent

The September meeting produces limited, largely symbolic progress but fails to resolve fundamental disagreements. Tariffs remain elevated or increase further. Export controls expand on both sides — Washington tightening restrictions on advanced chips and manufacturing equipment, Beijing tightening restrictions on both rare earths and, per the emerging pattern discussed in Section V, its own frontier and open-weight AI models. China accelerates technological self-sufficiency across semiconductors, AI and critical minerals simultaneously, while the United States expands domestic and allied capacity in the same three domains. Both countries increasingly require third countries and companies to choose, or at minimum align with, competing technological standards. Global supply chains become regional rather than global. Europe, India, Japan, Korea, Southeast Asia, Latin America and the Middle East each attempt to preserve strategic autonomy amid this bifurcation rather than aligning fully with either pole. The world does not divide into two wholly separate blocs but develops multiple overlapping and increasingly friction-laden economic spheres. This scenario is economically costly but remains structurally manageable.

Scenario Three: Strategic Bloc Formation — approximately 18 percent

A series of compounding crises — most plausibly involving Taiwan, Hormuz, or a serious military incident — causes Washington and Beijing to begin treating economic interdependence itself as a security threat rather than a source of mutual restraint. The United States deepens technological and financial alignment with allies at an accelerated pace; China deepens integration with BRICS, the Belt and Road network and non-Western financial systems, with renminbi settlement and BRICS Pay-style alternative payment infrastructure expanding meaningfully beyond the exploratory stage described in Section XIV. Critical-mineral supply chains bifurcate more completely than under Scenario Two. AI ecosystems become increasingly incompatible, with China formally restricting overseas access to its most advanced open-weight models along the lines already under internal discussion in Beijing. Space cooperation collapses entirely. Taiwan becomes the explicit organizing principle of Indo-Pacific military planning on both sides. Latin America becomes even more actively contested under an intensified Donroe Doctrine. The result would resemble a new Cold War, but with one critical structural difference from the twentieth-century precedent: the two economies would remain far more interconnected than the Soviet and American blocs ever were, making the new confrontation simultaneously

The result would resemble a new Cold War, but with one critical structural difference from the twentieth-century precedent: the two economies would remain far more interconnected than the Soviet and American blocs ever were, making the new confrontation simultaneously more economically costly and more difficult to manage cleanly.

Scenario Four: Systemic Fracture — approximately 7 percent

This is the low-probability, high-impact scenario, and unlike in earlier drafts of this analysis, several of its component triggers are no longer hypothetical but are active or recently active as of August 2026: a major Taiwan military incident; prolonged or renewed closure of the Strait of Hormuz beyond what current Iran–Oman negotiations resolve; tactical nuclear use in Ukraine or another theater; a direct U.S.–China military collision, whether in the Indo-Pacific or, newly plausible, in the contested space and Latin American domains discussed above; a major cyberattack against strategic infrastructure; an attack on critical satellites; or a breakdown of nuclear communications amid one of the several live crises this report documents. Once multiple crises interact, governments may begin interpreting every subsequent event as evidence of hostile intent, and the international system could move rapidly from competition into confrontation. The global economic consequences would include a compounded energy shock layered atop the current Hormuz disruption, semiconductor and AI-supply disruption, financial fragmentation, capital flight and potentially severe recession. This scenario must receive disproportionate policy attention precisely because its probability, while modest, is no longer negligible against the backdrop of an active Iran war, the largest Taiwanese military exercises on record, and the first direct American use of force against a head of state in the Western Hemisphere in a generation.

 XXIV. Regional Ramifications Through 2030

Indo-Pacific

The Indo-Pacific will remain the primary theater of U.S.–China strategic competition. Japan, South Korea, Australia, India, Indonesia and the ASEAN states will continue seeking to strengthen their own security without becoming fully subordinate to either Washington or Beijing. Taiwan will remain the principal flashpoint, with the South China Sea a secondary but genuinely dangerous theater in its own right.

Europe

Europe's strategic position will continue to be dominated by the unresolved Ukraine war, energy security and the future of NATO. The disappearance of the previous U.S.–Russian arms-control framework following New START's expiration increases structural uncertainty. Europe will face stronger incentives to develop independent defense-industrial capability while retaining the underlying American security relationship, and will remain exposed to Chinese rare-earth leverage in ways that increasingly mirror the American experience, as reflected in the European Commission's own defensive trade measures adopted in 2026.

Middle East

The Middle East will remain a genuinely multipolar strategic environment rather than one reducible to great-power competition. Iran, Israel, Saudi Arabia, Türkiye,  Persian  Gulf monarchies and other regional actors will continue pursuing largely autonomous strategies. Hormuz will remain the critical global energy chokepoint through at least the near term of this report's horizon. China will increasingly be viewed as an indispensable economic actor across the region even as the United States remains the dominant security power, and the credibility of American-led diplomatic initiatives — including the Gaza framework itself — will remain contingent on domestic Israeli politics in ways that are not fully within Washington's control.

Latin America

Latin America has moved, more decisively than any other region covered in this report, from a zone of competitive economic engagement to one of active, occasionally military, great-power contestation. The Donroe Doctrine will continue to push Washington toward direct strategic engagement, including further use of security and, potentially, military instruments. China's most likely response is measured recalibration rather than confrontation — preserving commercial relationships where politically possible while avoiding direct contestation of American primacy in a region where Washington has already demonstrated a willingness to use force. Latin American governments will continue to exploit this competition to maximize their own bargaining leverage, even as several have shifted toward political alignment with Washington on largely independent domestic grounds.

Africa

Africa will grow increasingly important for critical minerals, energy, infrastructure and emerging consumer markets. China's existing commercial presence gives Beijing significant continuing advantages. The United States and other G20 partners therefore have a genuine incentive to offer credible infrastructure and financing alternatives, rather than approaching the continent primarily through a security lens, if they wish to compete effectively for long-term strategic access rather than ceding the field by default.

XXV. The G20 as the Necessary Third Layer of the U.S.–China Game

The September Trump–Xi meeting will be bilateral. The December G20 summit cannot be, and should not attempt to substitute for bilateral stabilization. A bilateral U.S.–China agreement can stabilize the relationship between the two largest powers, but it cannot by itself resolve the structural problems facing the wider international economy — from mineral-supply concentration to AI safety to the financial architecture questions raised by BRICS. The G20 must therefore function as a multilateral stabilizer operating alongside, rather than in place of, the bilateral track.

The United States, as 2026 G20 host, has structured its presidency deliberately narrowly. Official American priorities center on three themes: unleashing economic prosperity by limiting regulatory burdens; unlocking affordable and secure energy supply chains; and pioneering innovation in artificial intelligence and emerging technologies. Washington has explicitly signaled a "back to basics" approach, streamlining the summit's workstreams, narrowing the expected guest list relative to recent years, and excluding South Africa from this year's gathering amid an unresolved bilateral dispute. This is a narrower agenda than the crisis-management architecture this report recommends, and reconciling the two is itself a central strategic question for the Miami summit.

The resolution need not require expanding the G20's formal mandate. Each of this report's recommended priorities maps naturally onto the American presidency's own declared themes: energy security and critical-mineral resilience fit squarely within "affordable and secure energy supply chains"; AI safety and human control over nuclear and autonomous systems fit squarely within "pioneering innovations in artificial intelligence"; and financial-system interoperability, addressed through a lens of reducing regulatory friction rather than confrontation, fits within "unleashing economic prosperity." The practical task for the Miami summit is therefore less to add new agenda items than to ensure that the existing, narrower American themes are pursued with explicit awareness of the strategic-stability stakes documented throughout this report.

XXVI. Recommended G20 Agenda

1. Establish a Global Energy Security Mechanism

The G20 should develop emergency protocols for disruptions affecting Hormuz, Bab el-Mandeb, the Suez Canal and other critical energy corridors, informed directly by the operational lessons of the ongoing Hormuz crisis. Energy security should be defined broadly enough to include oil, gas, electricity, nuclear fuel, batteries and critical minerals together, reflecting the electro-technological transition described in Section IX.

2. Create an AI Strategic Stability Compact

Building directly on the September dialogue's own AI focus, the G20 should establish minimum shared principles concerning human control over nuclear decisions, military AI verification, defenses against AI-enabled cyberattacks, constraints on fully autonomous weapons, and dedicated emergency communication channels between major AI powers — extending the logic of the "Humanity at the Threshold" declaration into an intergovernmental commitment.

3. Establish a Space Crisis-Prevention Mechanism

The G20 should promote advance-notification procedures and emergency communication protocols for dangerous satellite encounters and incidents involving strategic space infrastructure, recognizing that the space competition described in Sections XXI and XXII is no longer confined to a single region.

4. Strengthen Critical-Mineral Resilience

The objective should not be complete decoupling from China, which the June 2026 blacklisting episode demonstrates remains a costly and disruptive instrument for all sides when deployed unilaterally. The objective should be diversification: a genuinely resilient global mineral market reduces the incentive for any country to weaponize supply in the first place.

5. Protect the Global Financial System from Strategic Fragmentation

The G20 should explicitly recognize the growth of alternative payment systems and central-bank digital currencies documented in Section XIV, and pursue interoperability between these emerging systems and existing dollar-based infrastructure rather than treating their expansion as inherently adversarial.

6. Maintain Open Channels on Ukraine

The G20 should support the negotiating track described in Section XVII while preserving dedicated humanitarian, nuclear-safety and energy-security channels that can continue functioning independent of the broader negotiation's pace.

7. Develop a Taiwan Crisis Communication Protocol

The G20 cannot determine Taiwan's political future. It can, however, reduce the economic consequences of military escalation by preparing emergency communication and financial-stability mechanisms in advance, informed by the specific command-disruption scenarios Taiwan's own 2026 exercises have highlighted.

8. Promote Competitive Infrastructure Rather Than Economic Exclusion

The G20 should not require developing countries to choose categorically between Chinese and Western infrastructure. The international community should instead create genuinely competing financing and infrastructure options, a lesson the Donroe Doctrine's early implementation in Latin America has already illustrated can otherwise generate significant local political backlash even among aligned governments.

9. Establish Rules for Economic Coercion

Tariffs, sanctions, export controls and investment restrictions should remain legitimate instruments of national policy, but their systemic consequences should be formally recognized. The G20 should develop mechanisms for transparency, notification and consultation concerning measures with major global spillovers, of the kind the June 2026 rare-earth blacklisting episode demonstrated can occur with little warning even amid an active diplomatic thaw.

XXVII. Bayesian Indicators to Monitor Between September and December

The probability of each scenario in Section XXIII should not be treated as fixed. The following observable developments should cause substantial updating between the September summit and the December G20.

A durable Iran–Oman agreement on Hormuz that survives beyond an initial announcement, and a corresponding sustained decline in oil-price volatility, would increase confidence in the Managed Competition scenario. Continued implementation of China's agricultural purchasing commitments despite the June rare-earth episode would constitute a second positive signal, as would rapid resolution of outstanding rare-earth licensing disputes. A substantial reduction in military activity around Taiwan following the conclusion of the Han Kuang exercises would also increase the probability of strategic stabilization.

Conversely, expanded Chinese military exercises around Taiwan — particularly if timed close to the September or December summits and accompanied by coercive economic measures — would increase the probability of Asymmetric Decoupling or Strategic Bloc Formation. A prolonged or renewed Hormuz closure beyond the current Iran–Oman negotiating track would sharply increase systemic risk. A tactical nuclear incident in Ukraine or elsewhere would trigger an immediate reassessment of the entire scenario structure. A significant breakthrough in the Ukraine negotiations Zelenskyy has described for August would reduce European nuclear risk, while major Russian escalation, including any mobilization timed around Russia's own September elections, would increase the probability of systemic fracture.

Expansion of BRICS payment-system interoperability beyond the exploratory stage described in Section XIV, or a formal operational launch of BRICS Pay or CBDC linkage, would increase the probability of accelerating long-term monetary diversification. Confirmation that China intends to formally restrict overseas access to its advanced open-weight AI models, following the exploratory discussions with Alibaba, ByteDance and Zhipu reported in July, would signal that Beijing is preparing for a more technologically fragmented world and should be read as a meaningful escalatory data point in its own right. Finally, any expansion of the Donroe Doctrine into direct military action against a second Latin American government, or any Chinese decision to contest that doctrine more assertively rather than accommodate it, would materially raise the probability of Strategic Bloc Formation.

XXVIII. The Strategic Paradox of 2026–2030

The central paradox of the period remains that the United States and China need one another most precisely when they trust one another least. China requires access to major consumer markets, advanced technologies and global financial networks; the United States benefits from Chinese manufacturing capacity, critical-mineral supply chains and an enormous consumer market. Neither side can easily replace the other without incurring substantial costs, even as both governments have concluded, on the evidence of 2026 alone, that excessive dependence constitutes a strategic vulnerability rather than simply an economic efficiency.

This produces a self-reinforcing contradiction that this year's evidence has sharpened rather than resolved. Economic interdependence continues to encourage cooperation, most visibly in the May summit's trade and investment architecture. Security competition continues to encourage separation, most visibly in the June rare-earth blacklisting, the ongoing Han Kuang exercises, and the accelerating hemispheric contest in Latin America. Technological competition accelerates both dynamics simultaneously, since AI leadership is simultaneously the domain of greatest potential cooperation — through the AI-nuclear commitment discussed in Section VI — and the domain of most intense unresolved rivalry, as described in Section V. The resulting equilibrium is therefore unlikely to be either full globalization or total decoupling. It will be managed interdependence under strategic rivalry, sustained less by trust than by the shared recognition, on both sides, of what an uncontrolled break would cost.

XXIX. Final Assessment

The September 2026 Trump–Xi meeting should not be judged by whether it produces a historic, comprehensive settlement. That is the wrong benchmark, and the events of the months preceding it — an active war with Iran, the largest Taiwanese military exercises on record, a Chinese rare-earth blacklisting episode that undercut the May détente within five weeks, and the collapse of the administration's own Gaza framework at the hands of its closest regional ally — make clear that no single meeting could plausibly resolve the accumulated tensions this report documents. Its success should instead be measured by whether it prevents competition from crossing thresholds that neither government could subsequently control.

The most important achievable outcomes are therefore relatively modest in appearance even if strategically significant in substance: maintaining trade channels despite periodic disruptions of the kind seen in June; reducing critical-mineral uncertainty through licensing predictability rather than renewed confrontation; preventing further tariff escalation; establishing the AI dialogue announced for September as a genuine, sustained channel rather than a single event; preserving Taiwan crisis-communication channels through and beyond this year's exercises; supporting the Iran–Oman track toward Hormuz stabilization even without full American–Iranian reconciliation; taking initial, narrow steps on nuclear and AI-nuclear risk reduction; and preserving space-related communication amid the domain's rapid regionalization into new theaters including Latin America.

These achievements would appear less dramatic than a comprehensive geopolitical settlement. They would nevertheless be strategically more valuable, and considerably more attainable given the evidence available as of this report's cut-off date. The December G20 Summit then becomes the second stage of the process: the bilateral Trump–Xi meeting can establish or reaffirm a framework for U.S.–China stability, while the G20 — pursued through its own deliberately narrow official themes of deregulation, energy security and AI innovation — can convert selected elements of that bilateral understanding into multilateral mechanisms without requiring an expansion of its formal mandate.

The emerging international order will not be defined by the disappearance of great-power rivalry. The evidence compiled in this report — an active U.S.–Israel war with Iran, the first direct American military removal of a head of state in Latin America in a generation, a rare-earth blacklisting that unraveled a trade détente within five weeks, and a Gaza ceasefire framework publicly rejected by its own principal regional partner — makes unmistakably clear that rivalry, in multiple registers and multiple regions simultaneously, is not receding in 2026. It will instead be defined by whether rivalry can coexist with a functioning global economy. That is the central Bayesian judgment of this report, and the most dangerous pathway toward its failure is not a single crisis but cumulative miscalculation: a Taiwan incident that alters semiconductor markets; a semiconductor crisis that accelerates technological decoupling; technological decoupling that intensifies financial fragmentation; financial fragmentation that strengthens BRICS alternatives; BRICS expansion that increases American strategic pressure; strategic pressure that intensifies Chinese military hedging; military hedging that increases the risk of an incident in Taiwan, space, or now Latin America; and a regional military incident that transforms an economic rivalry into a systemic confrontation.

The strategic objective of the September meeting and the Miami G20 should therefore be to prevent these feedback loops from becoming self-sustaining. The appropriate doctrine is neither naïve cooperation nor permanent confrontation. It is guarded cooperation under conditions of strategic competition. For the United States, this means preserving deterrence — in the Indo-Pacific, in Latin America, and now explicitly in the AI and space domains — without converting every economic relationship into a security confrontation. For China, it means pursuing technological and strategic autonomy without transforming every American measure, including the Donroe Doctrine's implementation in a hemisphere Beijing has spent decades cultivating, into an existential challenge requiring direct confrontation. For the G20, it means constructing institutions capable of absorbing shocks before they become crises, within an American presidency that has deliberately chosen a narrower official mandate than the moment may ultimately require.

The international system entering 2030 will almost certainly be more multipolar, more technologically competitive and less economically integrated than the system that existed before the current period of strategic rivalry began. But multipolarity does not necessarily imply disorder. The decisive question is whether the major powers can learn to distinguish competition from confrontation, deterrence from provocation, and strategic autonomy from economic isolation — in the Indo-Pacific, across an increasingly fractured Middle East, in a Western Hemisphere now defined by direct American assertion of primacy, and in the still-forming governance of artificial intelligence, space and nuclear risk together.

The September Trump–Xi meeting is therefore best understood not as a summit that will decide the future of the world, but as a critical Bayesian updating event within a much larger repeated game whose stakes have, if anything, grown since this report's initial drafting. The December G20 Summit in Miami will provide the broader institutional test, conducted under an American presidency that has chosen to frame its priorities narrowly around deregulation, energy and technology even as the strategic environment surrounding those themes has grown considerably more volatile. If Washington and Beijing can establish sufficient confidence to keep channels open while competing intensely — as they have, imperfectly, managed to do through a war, a military intervention in Latin America, a rare-earth blacklisting episode and a collapsing regional peace framework, all within the months preceding this report's cut-off date — the international system can move toward a difficult but manageable equilibrium. If they instead conclude that every concession is weakness and every defensive measure is preparation for aggression, the world could move toward strategic blocs, economic fragmentation and an increasingly dangerous security dilemma.

The difference between those futures may depend less upon the grand declarations made at either summit than upon whether both governments can preserve credible signals, reversible actions and functioning communication channels precisely when uncertainty is greatest. That is the central strategic imperative for the G20 through 2030.

Selected Sources and Evidence Base

This report deliberately excludes Wikipedia, Encyclopaedia Britannica and unverifiable secondary references. All factual claims have been checked against official government sources, established news organizations and recognized research institutions, current as of the August 11, 2026 data cut-off.

White House statements and contemporaneous American and Chinese official readouts of the May 13–15, 2026 Trump–Xi Beijing summit, including the Boards of Trade and Investment, agricultural purchasing commitments, the Boeing order, critical-minerals language, and the joint statement on Iranian nuclear weapons and the Strait of Hormuz, as corroborated by Forbes, the South China Morning Post and CSIS summit coverage.

United States Trade Representative documentation and public statements characterizing the bilateral Board of Trade and the continued role of tariffs in U.S. trade and national-security policy.

Asia Times, Brussels Signal, CSIS and contemporaneous wire coverage, July–August 2026, on the confirmed September 24, 2026 Trump–Xi meeting in Washington, its timing relative to the UN General Assembly, and its artificial-intelligence-centered agenda, including the planned U.S.–China AI dialogue led by Treasury Secretary Scott Bessent.

Congressional Research Service (Library of Congress), early August 2026, report on Strait of Hormuz security developments and their impact on oil, gas and other commodities.

Bloomberg, CNBC and CNN reporting, August 6–10, 2026, on continued attacks in the Strait of Hormuz, Iran–Oman negotiations, oil-price movements and the status of the U.S. naval blockade of Iran.

U.S. Energy Information Administration data cited in contemporaneous financial-press reporting on 2026–27 Middle Eastern oil-production shut-ins and Brent price forecasts.

Agence France-Presse, France 24, Arab News and regional wire coverage, August 5, 2026, on Taiwan's Han Kuang military exercises, their focus on decentralized command and drone operations, and Taiwan's planned 2027 defense-spending increase.

Stockholm International Peace Research Institute, SIPRI Yearbook 2026 (released June 8, 2026), including its assessments of global nuclear-warhead inventories, China's expanding arsenal, the expiration of New START, Russian dual-capable missile deployments, and its dedicated chapters on artificial intelligence and space security governance.

Foundation for Defense of Democracies, Foreign Policy magazine and CSIS analysis, June–August 2026, on China's June 22, 2026 rare-earth export-control blacklisting of MP Materials and USA Rare Earth, and on the broader evolution of China's export-control architecture since October 2025.

Reuters reporting and Reserve Bank of India statements, August 2026, on BRICS discussions of linked payment systems and central-bank digital currencies ahead of the September 2026 New Delhi summit.

People's Bank of China announcement and Xinhua, People's Daily and CGTN coverage, August 10, 2026, of the PBOC's 2026–2030 financial reform and development plan and its renminbi-internationalization priorities.

CSIS "Two Loops" analysis (U.S.-China Economic and Security Review Commission), Center for Strategic and International Studies reporting on DeepSeek, Sheppard Mullin legal analysis, and Transformer News reporting, 2026, on the state of Chinese open-weight AI development, compute constraints, and Beijing's own deliberations on restricting overseas access to advanced Chinese models.

NBC News, Al Jazeera, NPR, the Washington Post and CNN, August 9, 2026, on Prime Minister Netanyahu's rejection of the Trump administration's fifteen-point Gaza plan and the status of the Board of Peace under UN Security Council Resolution 2803.

The China-Global South Project, CSIS Interpret: China, CNBC and Foreign Affairs, January–July 2026, on the origins and implementation of the "Donroe Doctrine," including the January 2026 capture of Venezuelan President Nicolás Maduro and its regional consequences for Chinese commercial and infrastructure interests.

U.S. Department of State releases, December 2025, and CBS News reporting, on the United States' assumption of the 2026 G20 presidency, its declared priorities, and the December 14–15, 2026 Leaders' Summit at Trump National Doral Miami.

Ukrainska Pravda and RBC-Ukraine, early August 2026, on the status of Ukraine–Russia diplomatic engagement and continued Russian strikes.

Editorial Note on Probabilities

The scenario probabilities in this report are analytical posterior judgments, not statistical forecasts. They are deliberately presented as Bayesian assessments rather than as objectively measurable probabilities, and they should be updated whenever material evidence changes the strategic beliefs of Washington, Beijing, Taipei, Moscow, Tehran, Jerusalem or other major actors identified in this report. The principal methodological premise remains Pragmatic Bayesianism: probabilities are not treated as permanent numerical truths but as disciplined representations of current belief that must change when credible new evidence arrives.