Xi Jinping left Washington on September 25 and was back in Beijing the next afternoon. The talks were held, the state dinner was served, and the coverage split along predictable lines, which was predicted and confirmed. The implementation audit — the four-track behavioral audit published here on October 4 — is locked. Those questions are settled — or at least staged for settlement in December.
There is a further question the coverage did not address, because it requires looking in a different direction: not at Washington's record of the summit, but at Beijing's. The Xi that Washington saw and the Xi that China was shown are not the same figure — and the gap between them is worth measuring.
The gap has a name in this framework. The operating system runs separate channels simultaneously: what the leadership says to the world, what it says to the Party, and what it allows the domestic public to hear are three distinct messages, calibrated for three distinct audiences. A summit visit to Washington is raw material for all three at once. How Beijing processed that material in the days after Xi landed home says something about the summit that neither side's press release does.
The summit produced no joint statement, and that turns out to be useful. Each side published its own account: a White House fact sheet on September 25, a Chinese foreign ministry readout of the September 24 talks, and, on September 26, an eight-point “outcomes consensus” (八点成果共识) released through Xinhua and carried across state media. Two official records of the same meeting, written for two different audiences, can be compared line by line. The comparison is the method of this piece.
Beijing's domestic record led with status. The first of the eight points is the agreement to build a “constructive strategic stability relationship” based on “respect, fairness, and reciprocity” (尊重、公平、对等). A Xinhua commentary republished by the Cyberspace Administration on September 29 explained why that point comes first, glossing reciprocity as “the two countries are sovereign equals, with no higher or lower” (两国主权平等，没有高低之分). The foreign ministry readout of the talks themselves uses the same register — “the spirit of equality, respect, and mutual benefit” and “the US and China are both major countries.” The domestic narrative does not require that anything concrete was agreed. It requires that equality was performed and recorded.
That framing is not new. It is the standard Chinese treatment of senior US-China meetings in the Xi era, and on its own it says little. The informative part is what each record kept and what each one dropped.
The concessions that changed shape are the clearest signal, and there are three.
First, fentanyl. The White House fact sheet names fentanyl precursors specifically: it notes China's new export controls on two precursor chemicals, cites 21 arrests in August, and encourages China to schedule these and other substances permanently. China's eight-point record does not use the word fentanyl. Its sixth point reports that drug-enforcement cooperation has produced “visible results” (可视化成果), with cases jointly solved and dozens of suspects arrested “in both countries.” The item is on Beijing's record, but recast: a one-way compliance demand becomes a two-way law-enforcement partnership. The word Washington cares most about does not appear in either the readout or the consensus.
Second, artificial intelligence. The US fact sheet announces a “U.S.-China Super Intelligence (SI) Dialogue” and a bilateral channel for “SI incidents,” and states that the two sides agreed to use the term super intelligence. China's record calls the same mechanism a China-US “artificial intelligence dialogue” (中美人工智能对话) and an “AI incident” channel. Xi's own language in the readout — that “AI must be kept under human control” and that both countries are “AI major powers” — frames the dialogue as a meeting of peers in a field where both are major powers, not as a safety constraint accepted at American urging. Same mechanism, two names, two stories.
Third, purchases. The White House fact sheet lists commitments that run in one direction: China will import at least 10 million metric tons of US coal in each of 2027 and 2028, an agricultural market-access working group is launched, and a bilateral Board of Investment is established. Beijing's eight points mention the Board of Trade and the reciprocal tariff arrangement on $30 billion of goods each way — framed as reciprocal — but not the coal volumes, and, as CSIS analysts noted, not the Board of Investment. What survived into the domestic record is what can be described as mutual.
Some items are missing from both records, and those prove nothing about domestic filtering. Neither the US fact sheet nor the Chinese consensus mentions the South China Sea, transshipment, or global imbalances — the last even though the G20 finance ministers' meeting in Asheville had ended without a communiqué three weeks earlier, a failure Washington blamed on Beijing's objection to imbalance language. An omission both governments chose is a negotiated silence, not a domestic edit. It belongs in the implementation audit, not in this one.
The record also runs the other way. Beijing's consensus includes items Washington's fact sheet omits: a point recalling that China and the United States were allies in the Second World War, and, in the readout, Xi's demand that Washington “handle the Taiwan question with prudence” and oppose Taiwan independence. Each side kept what served its audience.
The strongest counterargument is that none of this measures anything. Its best advocate would put it this way. Every government edits summit communications for domestic consumption, and Beijing's editing follows a fixed template — sovereign equality, reciprocity, no visible concession — that it applies whatever was agreed. Beijing has made real concessions while recasting them at home — the 2015 commitment to intervene in currency markets only against disorderly conditions was presented domestically as a sovereign reform choice, not a promise to Washington. If the template is constant, variation in the domestic record carries no information about implementation, and reading it closely risks mistaking propaganda architecture for policy intent.
That argument has real support. The template is visible here, and the 2020 Phase One deal cuts against this piece from the other side: Beijing published the purchase commitments in its own text of the agreement, and still fell well short of them. A commitment named at home is not a commitment kept. What the counterargument explains better than this piece does is the uniformity of the equality framing, which would appear whether the summit was a success or a failure. It does not explain the item-level differences — why the drug-enforcement item was kept and recast while the coal volumes were dropped. Those differences are what remains after the template is subtracted. The counterargument therefore narrows the claim rather than overturning it: the domestic record is weak evidence about intent, and modest evidence about which commitments Beijing is prepared to be held accountable for at home. Confidence is lowered accordingly, and the test below is designed so that it can fail.
The nationalist signal is where the evidence runs against a reassuring reading. One sensitive indicator of a summit's content is what happens to assertive messaging afterward. A leader who returns having conceded something sometimes sees the domestic information environment compensate — a claim restated, an exercise announced — to show that engagement did not mean weakness.
On Taiwan, there was no compensation signal. No exercise was announced in the Taiwan Strait, and Taiwan's defense ministry reported modest daily activity on September 28 and 29 — three PLA aircraft and five or six naval vessels a day.
In the South China Sea, there was. On September 24, while Xi was at the White House, Chinese coast guard and navy vessels blocked a Philippine resupply mission bound for Second Thomas Shoal — by Manila's account, the first blocked mission since the 2024 provisional arrangement. On September 27, the day after Xi landed in Beijing, the PLA Southern Theater Command announced a joint naval-air exercise around Scarborough Shoal (Huangyan Dao) to safeguard “territorial sovereignty and maritime rights and interests,” and China Daily called it “a stern warning to the Philippines.”
Two readings fit that sequence. The first is displacement: assertiveness was expressed, but redirected away from Washington toward a lower-cost target, preserving the summit's tone with the United States while still signalling strength at home. The second is that the summit had nothing to do with it: the exercise answered a Philippine-led multinational drill near Scarborough on September 23, and South China Sea incidents are frequent enough — three China-Philippines clashes in a single week in July — that one falling in any given 72-hour window is unremarkable. The base rate favours the second reading. What would separate them is repetition: if assertive action toward third parties clusters after future high-level engagements with Washington more than in matched windows before them, displacement gains weight. If it does not, the September sequence is coincidence.
The economic register adds a third channel, and it read the summit more coldly than the diplomats did. The framework's earlier “Two Xi Jinpings” piece separated the Xi that foreign audiences see from the Xi that Party audiences see. A further register is the signal the summit sent to markets. Mainland exchanges were closed on September 25 for the Mid-Autumn Festival. On September 24, the day of the talks, the Shanghai Composite fell 1.22 percent amid doubts about the trade truce. When trading resumed on September 28, the Shanghai Composite fell 1.67 percent and the CSI 300 fell 2.22 percent, with market reports citing fading summit optimism alongside a new US move against Chinese components in AI data centres and weak August industrial profits. Hong Kong's Hang Seng rose 0.54 percent that day but finished the week about 1.6 percent lower. Markets move on many inputs, and a holiday-shortened week is thin evidence. But no reading of those numbers supports a summit that reduced bilateral risk in investors' eyes.
Taken together, the three registers are consistent with a managed de-escalation in tone that did not reach substance. Beijing put on the domestic record what could be presented as mutual, and recast or dropped what looked like one-way compliance. The security services kept up pressure on a US ally while the leader was in Washington. Investors priced in limited relief.
Whether that is a useful outcome depends on the question. As near-term risk management, removing the summit itself as a flashpoint has value. As a test of whether the relationship has moved structurally, the domestic record offers no evidence that it has.
The test is registered here so it can be scored. The claim is that commitments Beijing placed on its own domestic record will be implemented more readily than commitments that appear only on Washington's. The horizon runs to December 31, 2026, and each item resolves against official Chinese and US government releases (MFA, Xinhua, Ministry of Public Security, the White House, State and Treasury), with major wire coverage as backup.
On the shared record: the China-US AI dialogue convenes by November 30, 2026, confirmed by both governments. Confidence: 75 percent. A postponement announced jointly with a new date before December 31 counts as ambiguous, not as a miss.
On Washington's record only: by December 31, 2026, China does not announce permanent scheduling of additional fentanyl-related precursor substances, and the Board of Investment holds no first meeting announced through Chinese official channels. Confidence: 60 percent for each. Either one happening counts against the claim.
The claim fails if the US-only items are implemented at least as fully as the shared items over the same period. In that case the domestic record adds nothing to the implementation audit, and this register will be dropped from the framework rather than reinterpreted.
On displacement: within seven days after the close of the APEC Leaders' Meeting in Shenzhen in November, the PLA or China Coast Guard announces an exercise or enforcement action against the Philippines, Japan, or Taiwan using sovereignty language. Confidence: 50 percent, which is to say no better than a coin flip on one observation. It counts toward displacement only if the seven days after exceed the seven days before. Routine “patrols” announced without sovereignty language are scored as ambiguous. A single result will not settle the question in either direction; it is the first data point in a series.
This entry is fixed as published. Results and a postmortem will run in January.
Washington saw a counterpart negotiating tariffs, AI incidents, and drug enforcement. China was shown a sovereign equal who conceded nothing that could not be called mutual, while its coast guard and navy kept pressure on a US ally. Whether either picture survives December depends on what the behavior shows — and the behavior is being tracked.
J. Kenneth Rhodes writes on geopolitical risk, institutional resilience, and economic frameworks for assessing system-level fragility. His work focuses on the gap between official narratives and behavioral evidence — and on the analytical infrastructure required to read one through the other. He is based in the United States. Contact: jkr@rhodesresearch.org | rhodesresearch.org