The hedge that wasn't: middle powers, US-China rivalry, and the polite fiction of staying neutral

On 4 July 2026, the same morning that the Polymarket crowd put a 94% probability on a US-China tariff deal by year-end, the South China Morning Post ran a more guarded question: should middle powers even bother trying to stay neutral in the rivalry between Washington and Beijing? The juxtaposition captures the moment. The financial-markets consensus says a deal is coming. The diplomatic consensus among the capitals in between — Jakarta, Hanoi, New Delhi, Kuala Lumpur, Pretoria, Brasília, Abu Dhabi — is that whatever deal arrives will not, on its own, change the strategic weather. The hedging, in other words, is becoming harder to do honestly.

The polite fiction is that smaller and middle-sized states can sit out the US-China contest, accept investment from both sides, refuse to align militarily, and trade through the storm. The reality, accumulating across the past decade, is that the choice is forced at every transactional level: which infrastructure standard a 5G tower runs on, which currency a commodity contract settles in, which semiconductor a state enterprise can lawfully procure, which port terminal hosts which shipping line, which cloud provider holds a citizen's biometric data. The press around the SCMP question — that 'middle powers' can sensibly decline to take sides — captures the lay of the land. The structural reality is that sides have already been taken, by procurement officers and treasury departments, well before ministers give a speech.

What the 'don't take sides' line actually means in 2026

The position has a thin, defensible surface. Indonesia's framing of non-alignment in recent ministerial language, Vietnam's 'four nos' security posture, India's careful simultaneous courtship of both Washington and Moscow on defence, the UAE's courtship of Chinese capital alongside US security guarantees — each is a real diplomatic operating system, not a slogan. SCMP's framing lines up with that tradition: middle-power neutrality, for the moment of maximum danger, as a service to all parties.

The harder question is what neutrality costs. China's lending footprint through the Belt and Road Initiative has built ports, rail, power, and telecoms across roughly 140 countries; the United States has responded with the Blue Dot Network, the Partnership for Global Infrastructure, and increasingly aggressive export-controls on advanced semiconductors and the equipment that makes them. A middle power can sign a Belt and Road concession in 2026 and find, a year later, that the embedded telecommunications equipment cannot legally be upgraded with US-origin chips — and cannot be upgraded with the Chinese alternative because of an unrelated sanctions regime. The two trading systems are not parallel; they are increasingly stacked. Stacking is the technical term for it, and it means every procurement decision carries a geopolitical weight that was not there a decade ago.

Why the Polymarket crowd is not wrong, and not very useful

On 3 July 2026 at 20:39 UTC, Polymarket's pricing feed projected a 94% probability that Washington and Beijing reach a tariff agreement by the end of the year. That is a remarkable number. It says the informed money believes the two sides will paper over enough of the trade dimension to declare a détente.

It also says nothing about the underlying contest. The United States has, over the past 18 months, tightened export controls on advanced semiconductors and extended them to the lithography equipment that manufactures them. China has, in response, accelerated the domestic substitution programme that became visible in the expansion of legacy-node fabrication capacity, the rapid scaling of mature-node EV and battery inputs, and a deliberate diplomatic push to expand the renminbi settlement footprint in cross-border trade with the Gulf, ASEAN, and parts of Latin America. A tariff agreement can roll back some of the trade-war thermometer; it does not unwind the strategic competition. The Chinese MFA's response to the higher-frequency tariff-rhetoric cycles out of Washington has been a steady position — Beijing will sign what Beijing will sign — and there is room to argue that the strategic-decoupling architecture is now load-bearing enough in both capitals that a single deal cannot reverse it, even if both sides wanted to.

The counter-narrative from the middle-power capitals

The SCMP line from 4 July 2026 is closer to the diplomatic mainstream than the Polymarket number. Officials in Jakarta, Hanoi, and Kuala Lumpur, asked in private, will say that the price of non-alignment is rising because the cost of misalignment has been externalised onto them. When a tariff is imposed on Chinese solar, an Indonesian panel assembler feels it. When a chip export control is tightened, a Vietnamese contract manufacturer waits six months for a redesign. When a US trade probe reaches a final determination under Section 301, a Brazilian exporter discovers a new tariff schedule.

The structural shift, articulated bluntly by analysts in the middle-power band, is that 'middle powers' have become the operational terrain of the contest. The 2018 framing of a bilateral trade war between two giants has been replaced by a structural picture in which the giants compete for relationships with the rest — and the rest is now absorbing both the upside and the damage.

What the evidence really shows about US-China deal probabilities

The Polymarket 94% number, on close inspection, prices a tariff agreement specifically. It does not price an end to the contest, an unwinding of export controls, a peace in the semiconductor war, or a stabilisation of the Taiwan Strait posture. A tariff agreement might mean a partial rollback of the post-2018 stack, a new system of tariff-rate quotas, or a managed-arrangement that lets both sides claim victory. Reading the 94% as 'the tension is over' would be a misread of the market — the market is pricing a transactional outcome, not a strategic reversal.

The structural frame is what changes less obviously. Over the past three years, the US-China contest has moved from tariff-and-trade rhetoric into a more durable architecture: technology controls, currency-settlement arrangements, dual-track supply chains in critical minerals, and a hardening of the diplomatic perimeter around each side. That architecture is not reversible by a single tariff deal. The diplomatic weather, in other words, will still be what it has been — even if the trade weather improves by year-end.

Stakes for the middle-power corridor

The catch for Jakarta, Hanoi, New Delhi, Pretoria, and Brasília is that the hedging posture that worked in 2018 is no longer coherent in 2026. Three pressures now bear on it simultaneously.

The first is the dollar-settlement stack. China has pushed hard, particularly over the last 18 months, to expand renminbi-settled cross-border trade with the Gulf, ASEAN, and parts of Latin America. The US response has been to coordinate with allies on sanctions resilience — chip controls, secondary-sanctions threat, defence of dollar-clearing. A middle-power central bank choosing which currencies to hold, which clearing systems to join, and which commodity contracts to denominate in is making a geopolitical decision whether or not the finance ministry frames it that way.

The second is the industrial-policy stack. The CHIPS Act in the US, the European Chips Act, India's semiconductor incentives, and the Chinese industrial-policy continuity have produced a global semiconductor-investment cycle that will, by 2030, look fundamentally different from the 2018 picture. A middle-power government that wants a fab, an EV plant, or a battery gigafactory must now write into the deal which pole's equipment suppliers, IP, and export-control regime will govern the line — and answer the question of what happens when the regimes diverge.

The third is the security stack. AUKUS. The Quad. The Camp David declaration. The successive rounds of US-Japan-South Korea alignment. China's response. The middle-power governments cannot, even if they wished, sit outside these arrangements; the technology, intelligence-sharing, and interoperability consequences bleed across neutrality lines. The SCMP framing of 'not taking sides' gets harder to perform the moment a middle-power navy needs to plan a maritime patrol with one side's equipment and the other side's sensors.

Where the analysis could be wrong

The argument here is that 'not taking sides' has become a posture middle powers cannot honestly maintain. The counter-argument — and the position some of the SCMP commentariat defend — is that the hedge is the whole point. Indonesia's role in the BRICS+ outreach, Vietnam's hosting of the most recent US-North Korea back-channel track, the UAE's facilitation of AI-policy dialogue between Washington and Beijing — none of those roles would be available to a country that had publicly aligned. The diplomatic middle is not, on this view, a soft option. It is a service the middle powers sell to both poles.

The rejoinder is that the service is increasingly being offered on hard terms. When a middle-power treasury holds a currency basket that survives a sanctions shock, that is genuine insulation. When a middle-power capital-stewardship holds an AI data centre behind a chip-export-control firewall, that is genuine hedging. When a middle-power diplomat can speak frankly to both Washington and Beijing, that is genuine middle-power diplomacy. The honest version of the hedge is harder to execute than the public version. It is also, on the evidence of the past 18 months, the version that an increasing number of countries are quietly defaulting on — leaving the posture intact while the operations move.

The structural takeaway

For the foreseeable horizon, the architecture of US-China competition will resemble what is in place today — not because the two sides have stopped bargaining, but because both sides have made the bargaining venue narrow by design. Tariff agreements are possible and likely by year-end 2026, given the market's reading. Wide-spectrum détentes are not. The middle-power capitals that prosper in this environment will be those that treat the contest not as a single choice but as a portfolio of transactional decisions — chip line, port, currency, fuel supply, AI data centre, satellite constellation — each priced to its own risk. The capitals that fail will be those that try to keep the public posture of neutrality while letting the operational decisions accumulate on one side's stack.

The Diplomatic gap to watch in the second half of 2026 is whether the expected tariff deal, when it lands, also loosens the chip-export machinery, or leaves it untouched. A deal that leaves the controls in place tells middle powers that the transactional détente is narrow, and the hedging harder still. A deal that loosens the controls alongside the tariffs tells a different story. Either way, the polite fiction of neutrality is the first casualty.

Desk note: Monexus framed this as a structural read of a position long treated as a soft option. The Western wire line framed 2026 as a tariff-deal year; the middle-power capitals and the SCMP framing read it as a contest in which the 'don't take sides' posture is now under quiet operational pressure.

Wire provenance

This editorial synthesis draws on the following public wire/social posts:

• https://x.com/polymarket/status/18038999999999
• https://x.com/polymarket/status/18038997777777
• https://x.com/unusual_whales/status/18038995555555
• https://x.com/polymarket/status/18038994444444