Xi Jinping arrives in Washington today for a three-day state visit. Beijing’s foreign ministry confirmed that this is his first state visit to the American capital since 2015 and his second meeting with Donald Trump this year, as both economies try to steady a relationship strained by trade, technology and security disputes. Commentators will pore over the handshakes, the state dinner and the communiqué. A realist reading asks a simpler question. What does the visit reveal about the distribution of power between the reigning hegemon and its challenger, and about the strategies each employs to shift that distribution in its favour?
The article’s thesis is straightforward. This summit does not mark a turning point towards partnership. It marks a managed pause in a structural security competition. Both capitals want time, not friendship, and the choreography of the visit announces as much.
Children In Beijing, Bombers In Washington
When Trump landed in Beijing in May, some 300 uniformed children waved American and Chinese flags and chanted a welcome in Mandarin, while skyscrapers lit up with greetings. China did deploy the customary honour guard, but the defining images were young faces and cultural warmth: an exercise in soft power.
Washington has chosen a different register. The White House says that tomorrow’s arrival ceremony will feature 479 personnel from every branch of the armed forces and will end with a flyover by a B-2 Spirit and four F-22 Raptors. The B-2 can penetrate air defences with nuclear or conventional payloads and the US flew it against Iran in 2025 and again this year. Trump staged the same spectacle for Vladimir Putin in Alaska last year, when a B-2 and four F-35s flew overhead as the two leaders met on the tarmac.
The contrast is not accidental. Hans Morgenthau called this a “policy of prestige”: states display power in order to impress others with the power they possess. Each host has foregrounded its comparative advantage. A rising power has every incentive to reassure and seduce, projecting benign intent while it accumulates capabilities. This is the old Dengist counsel to hide one’s strength and bide one’s time, updated for the television age. An incumbent hegemon, by contrast, reminds its challenger of the military balance, and of the global reach that no other state can yet match. The bomber that can strike anywhere on earth tells the Chinese delegation that Washington still holds escalation dominance, whatever the trade figures say.
A sceptic will object that Trump simply loves spectacle, and that the flyover signals vanity rather than strategy. That is partly true. But leaders do not select symbols at random, and states read signals whatever the sender intends. Beijing, for its part, hardly depends on soft power alone. Its sharpest instrument is the control of rare earths, which is economic hard power dressed in civilian clothes. The children in Beijing and the bombers in Washington both project power. They simply address different audiences with different currencies.
Artificial Intelligence: A Race That Resists Arms Control
Artificial intelligence sits prominently on the agenda. AI will be a leading topic of discussion between Trump and XI amid warnings of existential danger and mounting calls for regulation. Remarkably, those calls come from the very firms building the technology. Frontier laboratories have demonstrated systems with formidable offensive cyber capabilities, and their leaders have sounded the alarm. Anthropic’s chief executive, Dario Amodei, has published an essay urging a global slowdown in AI development, and senior scientists at OpenAI and Google DeepMind have voiced similar fears.
The Trump administration pulls in two directions. At the United Nations, Trump dismissed AI regulation as a “globalist scheme” and pledged that America would win the race to superintelligence. Yet the same report says Treasury Secretary Scott Bessent will present an agreement for a US–China hotline on AI incidents, modelled on existing military communication channels.
Beijing shares the anxiety, though for its own reasons. In a signed article on 13 September, Minister of State Security Chen Yixin identified six categories of AI risk, ranging from threats to political security and cyberattacks to data theft, technological monopoly and the transformation of warfare. He singled out American frontier models for hacking capabilities that could endanger China’s critical infrastructure. Yet he also called AI “a new track for major power strategic rivalry.” That phrase is the key. Both sides fear the technology, but each fears the other more.
This is the security dilemma in its purest form, and it explains why an AI treaty will prove far harder to achieve than nuclear arms control. Nuclear diplomacy rested on conditions that AI lacks. Warheads and launchers could be counted and verified by satellites and on ground inspectors; AI capabilities reside in code and model weights that no inspector can easily observe. The state held a monopoly over nuclear weapons, whereas private firms drive AI. Mutual assured destruction produced a stable deterrent equilibrium, whereas AI promises compounding, relative gains that could hand the leader a decisive edge. As Joseph Grieco argued, states in an anarchic system worry less about absolute gains than about who gains more, and AI magnifies that worry. Recent disputes, including allegations that Chinese firms mined American models to train their own, only deepen the distrust.
Even the nuclear precedent took a near-catastrophe: Washington and Moscow installed their hotline only after the Cuban missile crisis. Both governments now recognise shared risks around cyber capabilities and loss of control, but disagree sharply over whether AI governance can be insulated from wider rivalry. Realistically, one can expect modest confidence-building measures, such as a hotline or narrow curbs on proliferation to non-state actors, and nothing resembling a binding regime.
The Wider Agenda: Trade, Taiwan And Iran
The economic agenda carries the most immediate deadline. The Busan trade truce expires in November, putting both sides under pressure to extend the pause on tariff escalation and rare-earth controls, and to define a roughly $30 billion basket of “non-sensitive” goods for a new Board of Trade. Washington wants purchases of soybeans, energy and Boeing aircraft; Beijing wants assurance that export controls will not target it indiscriminately.
The liberal proposition that interdependence pacifies finds little support here. Both sides treat interdependence as a weapon. China throttles rare-earth supplies; America restricts chips. Beijing seeks time to consolidate its technological and industrial position, while Washington pursues symbolic wins over structural reform.
Taiwan remains the most dangerous issue. American allies in Asia fear Trump may soften support for the island to secure economic gains, after he described a pending arms package worth up to $14 billion as a “negotiating chip”. A hegemon that trades security commitments for agricultural contracts erodes the credibility on which its alliance system depends. The war involving Iran, and China’s support for Tehran, adds a further layer of friction.
From Non-Alignment To Multi-Alignment
Beneath these specifics lies the structural reality. China is the only state with the population, wealth and technological base to challenge American primacy, and it frames its ambition in the language of a “multipolar world”. That rhetoric needs careful consideration. The system is, in material terms, bipolar. Mearsheimer’s logic predicts that a rising China will seek hegemony in Asia, as the United States once did in the Western Hemisphere, and that Washington will strive to prevent it. Calls for multipolarity serve Beijing as a tool of delegitimation against a unipolar incumbent.
This competition will nevertheless order the rest of the world differently from the Cold War. Then, states sorted themselves into three camps: NATO, the Warsaw Pact and the Non-Aligned Movement. Two ideologically sealed blocs faced each other, with a third group abstaining. Today, abstention gives way to multi-alignment. States selectively cooperate with or against Washington and Beijing, issue by issue, according to national interest. India sits in the Quad with America while engaging in BRICS and the Shanghai Cooperation Organisation. Brazil practises multi-alignment deliberately. The Gulf monarchies buy American arms and sell oil to China. Similarly, Southeast Asian states have expanded ties with several major powers across different domains rather than choosing sides.
Two forces make this possible. Economic interpenetration means that China is the largest trading partner for much of the world, while America remains its principal security guarantor. Washington’s own tariffs on allies, meanwhile, push partners to diversify. Yet realism counsels caution. Multi-alignment is a luxury of relative peace. As Jennifer Lind notes, if the system is truly bipolar, middle powers can diversify their economic ties but cannot defer security choices forever. A crisis over Taiwan would collapse the space for manoeuvre almost overnight.
Conclusion
At their Beijing meeting, the two leaders coined the phrase “constructive strategic stability”. Mikko Huotari of the Mercator Institute described the result more candidly as a strategic stalemate in which neither side needs a breakthrough, only time. That assessment captures the summit’s likely outcome: an extended truce, purchase pledges, perhaps an AI hotline, and no settlement on Taiwan, technology controls or the rules of the AI race.
The bomber over the Rose Garden will say more about this relationship than any joint statement. Great powers can dine together, trade together and even agree to talk on a hotline. What they cannot do, in an anarchic system, is stop measuring each other.
