China’s 15th Five-Year Plan, covering 2026–2030 and released in draft this March, runs to sixty-two chapters. Most Western analysis will catalogue the individual targets and move on. But the document rewards a different kind of reading. Modern Five-Year Plans are less policy manifestos than signalling documents, strategic frameworks that guide local governments and the private sector toward national priorities. Approached that way, the plan becomes considerably more interesting. It reveals what Beijing fears, what it is betting on, and what it has decided to leave out.

Two premises run through the text, stated more directly than Party documents usually permit. The first is domestic: China has a structural demand problem. The plan names effective demand as insufficient, flags a prominent contradiction between strong supply and weak demand, and admits to blockages in the domestic circulation. Expanding internal demand is elevated to a strategic base-point for the entire planning period. Western macroeconomists, the IMF, and critics like Michael Pettis have been making essentially this argument for fifteen years: that China’s growth model systematically suppresses household consumption in favour of investment and exports, producing imbalances that are unsustainable in the long run. For most of that period Beijing treated the critique as external noise, subordinate to supply-side upgrading. That the 15th Plan puts demand deficiency into its diagnosis of the development environment, using language that could have come from a Brookings working paper, suggests the problem has finally been absorbed at the planning level. The second premise is external: great-power competition has intensified, the trade order is under severe strain, coercive politics are ascendant.

What the document does not address matters too. The housing correction, borderline deflation, a current-account surplus approaching $1.2 trillion: the macroeconomic problems that dominate short-run commentary get remarkably thin treatment. Real estate appears in several sections but only as a risk to be “resolved” and expectations to be “steadied,” without concrete mechanisms. Deflation and the surplus go unaddressed entirely. A Five-Year Plan is pitched at a five-to-ten-year horizon, and the short-term firefighting is handled elsewhere. Still, the omission is telling. The surplus is what generates the trade frictions that dominate China’s relationships with its major partners. Deflation signals the demand failure that prevents the internal circuit from functioning. The housing overhang keeps households saving instead of spending. The omission leaves the impression that Beijing either considers these problems too politically sensitive to address in a public planning document or expects them to resolve on their own. Given that each has been worsening for years, the second interpretation is hard to credit.

The pattern of what the plan invests in and what it ignores points to two fears. The first is that the global network on which China depends for technology, energy inputs, and demand absorption will be weaponised against it. The second is that the domestic economy cannot generate enough demand to reduce that dependence. The two compound: an economy that cannot consume what it produces remains tethered to foreign markets, and tethered means exposed.

The plan’s response to these two fears runs in opposite directions simultaneously. One half reads as though China is building an economic fortress: technology sovereignty, energy autarky, supply-chain redundancy. The other reads as though it wants to be an economic forum: renminbi internationalisation, new multilateral institutions, trade-architecture expansion, and Global South coalition-building. Even though the two are not inherently incompatible (capital controls and currency internationalisation can coexist through deliberate sequencing, letting in stable investors first and flightier capital later), the domestic conditions the plan reveals may undermine the strategy from within. Yet, some of the tension in the document may simply be an artefact of the drafting process. Five-Year Plans are negotiated documents, shaped by bureaucratic compromise and the need to signal different things to different audiences. But the specific tensions that matter here, between defensive industrial investment and the macroeconomic conditions required for international integration, operate through mechanisms that do not depend on whether anyone in Beijing intended them. They are consequences of the plan’s commitments, not of its rhetoric.

Beijing has its own name for this architecture. Since 2020 the governing concept has been shuāng xúnhuán, dual circulation. The domestic circuit (production, distribution, consumption within China) was to become the primary engine of growth. The international circuit (trade, investment, technology exchange) would play a supporting role. The 15th Plan reveals that the two circuits are pulling apart, and contains no instrument for bringing them back into alignment.

The fortress

Six fields are designated for breakthrough under what the plan describes as exceptional policy intensity: integrated circuits, precision machine tools, high-end scientific instruments, base software, advanced materials, and biological manufacturing. None of these are new priorities. China has been targeting semiconductors, advanced materials, and precision manufacturing since the late 1970s. What has changed is the urgency, not the list. The target list is a map of perceived vulnerability, each item corresponding to a chokepoint where coercive leverage could be applied against China.

The defensive posture dates specifically to 2018-2019, when American entity-list actions against Huawei and ZTE demonstrated how chokepoint leverage could cripple a national champion almost overnight. The 15th Plan retains the 14th Plan’s self-reliance agenda but adds a second layer: arriving first in the domains that will structure the next network, a shift explored further below. R&D spending is targeted to grow above seven percent annually. A new immigration system for high-technology talent is announced.

The defensive logic shapes physical infrastructure too. Over 100 gigawatts of offshore wind, roughly 110 gigawatts of nuclear, west-to-east transmission capacity exceeding 420 gigawatts, a commitment to double non-fossil energy within a decade. Desert wind-solar bases across the Kubuqi, the Tengger, the Badain Jaran, converting barren terrain into the backbone of the national power system. Grain capacity rising to 725 million tonnes. Strategic mineral reserves expanding. These are the investments of an economy that expects the network around it to become hostile.

Two kinds of leverage

Not all of these defensive investments work against the plan’s international ambitions. The green transition is the one domain where the fortress logic and the international logic converge, and it draws out a distinction in the nature of economic power that the rest of the plan obscures.

Coercive leverage works by worsening a counterpart’s position if they refuse to cooperate: threatening exclusion, withholding access. American semiconductor export controls are the paradigm. So are SWIFT-based financial sanctions and Chinese rare-earth restrictions. Coercive leverage is effective in the short run but self-undermining over time, because the threat itself makes dependence dangerous. The target has every incentive to invest in substitution and alternative networks, and exercise coercive leverage often enough and the counterpart builds the capacity to walk away. The network degrades precisely because it has been weaponised.

Attractive leverage works the other way. It makes the shared arrangement so valuable that counterparts do not want to leave, even when political relations deteriorate. No threat is needed. Others rely on you not because you coerce them but because the alternatives are too costly or do not yet exist.

China’s position in the green transition is approaching attractive leverage. China now produces the overwhelming majority of the world’s solar panels, a dominant share of its batteries, and a rapidly growing share of wind turbines and grid infrastructure. Every other country’s decarbonisation depends, to a substantial degree, on Chinese manufacturing capacity of roughly 1,000 gigawatts of new solar per year. That is not chokepoint power. It is the structural leverage of indispensability. In climate, the fortress is the forum. In nearly every other domain the plan defaults to coercive logic. The exception that has not yet settled is artificial intelligence. If China achieves frontier AI capability and builds out the compute infrastructure, training-data pipelines, and model ecosystems the plan envisions, the question is whether that becomes attractive leverage (the world wants access to Chinese AI the way it wants access to Chinese solar panels) or coercive leverage (chokepoint control over compute and standards that can be weaponised). The conditions that would push it one way or the other are already visible: if Chinese AI tools are open-weight, cheaper, and widely accessible, they become a global public good that draws others in. If compute access is gated through state-controlled infrastructure and model use is conditional on political alignment, they become a weapon. The plan does not resolve that question, and probably cannot. But AI is the domain where the distinction will matter most over the coming decade, and the plan’s silence on which logic will govern it is conspicuous.

The forum

The fortress half of the plan is about insulation. The forum half is about reshaping the international architecture on different terms.

The renminbi sections envision an expanded role for the currency in trade invoicing and investment settlement, a cross-border payments architecture independent of Western-controlled infrastructure, and development of offshore markets. The dollar’s centrality rests less on American economic weight than on network effects: SWIFT messaging, correspondent banking chains, dollar-denominated commodity contracts. Everyone invoices in dollars because everyone else does, and switching costs are high. That Beijing frames its new payments system as needing to be self-controlled tells you which form of leverage it circumvents. Washington’s increasingly coercive use of existing payments infrastructure (the pattern that converts attractive leverage into the chokepoint variety) is what makes a Chinese alternative plausible to third parties who might otherwise see no reason to switch.

In trade architecture the plan pushes for expansion of RCEP (the fifteen-country Asia-Pacific trade bloc) and accession to the CPTPP (the high-standard Pacific trade pact originally designed without China). The CPTPP bid is telling: the agreement was built to set trade rules that excluded Beijing, and Beijing’s pursuit of membership is an effort to sit inside the architecture built to keep it out.

The plan also proposes a world organisation for AI cooperation and a Belt and Road AI governance platform, a bid only credible because the 15th Plan simultaneously makes China a technological leader. AI receives by far the most operational detail of any frontier technology: compute infrastructure, data governance, an “AI Plus” programme deploying the technology across agriculture, healthcare, education, governance, and industry. The plan treats AI not as one frontier among many but as an organising logic that cuts across the rest.

The broader coalition strategy is explicit. China as champion of a self-strengthening Global South. BRICS and the Shanghai Cooperation Organisation expanding. Not a rival bloc but an alternative institutional ecology where Chinese participation incentives are high and American structural leverage is low.

The bind

Both the fortress and the forum depend on the same thing: a domestic economy strong enough to absorb what it produces. The problem is that the internal circuit cannot carry the weight assigned to it. The demand-side reforms that would make it self-sustaining have not been undertaken. Without adequate domestic demand, the fortress programme does not feed the internal circuit. It overflows into the external one, as surplus production that must be exported, generating the trade frictions the forum is supposed to manage.

Subsidised capital costs and soft budget constraints for state-linked firms, combined with local-government incentives that reward output over profitability, push investment well beyond the point where marginal returns would justify it. The surplus goes abroad as cheap solar panels, electric vehicles, batteries, steel. The plan calls for rectifying this ruinous competition, borrowing the popular term neijuan (involution), a passage unthinkable in earlier iterations. But the proposed remedies (capacity monitoring, early-warning systems) work against the plan’s own imperatives and assume a degree of central control over local governments that the problem itself calls into question. The fortress requires overinvestment by design, and overinvestment produces overcapacity by arithmetic. A sufficiently large economy cannot invest defensively without projecting offensively. Beijing sees the surplus capacity as a side effect. Brussels and Bangkok see something indistinguishable from mercantilist aggression. The tariff response is the same.

The renminbi faces a related difficulty, though the standard framing may be slightly off. The usual critique is that reserve-currency status requires open capital accounts, deep financial markets, and predictable institutions. But the plan may not be aiming at reserve-currency status in the way the dollar holds it. What the payments-infrastructure sections describe is something narrower: a parallel transactional system for trade settlement, one that allows participating countries to conduct commerce without routing through dollar-denominated plumbing. That does not require full convertibility. It requires a functioning payments rail, a critical mass of willing counterparties, and enough coercive overreach by the incumbent to make switching costs tolerable. China may not need the renminbi to be the dollar. It may only need it to be an alternative to SWIFT. A transactional currency can operate with capital controls in a way that a reserve currency cannot.

There is some evidence that the sequencing can work. During the 2015-16 market turmoil, foreign investors liquidated over 20 percent of their Chinese bond holdings. Beijing restricted domestic capital outflows aggressively but did not restrict foreign investors from exiting. Foreign capital returned and exceeded pre-crisis levels. Whether that restraint holds in a more severe crisis, with flightier private capital now in the market, is the open question.

The missing piece

Whether the sequencing succeeds or fails, the deeper problem remains: why the internal circuit cannot absorb what the fortress produces. The mechanisms that suppress Chinese household consumption are well documented: thin social insurance (particularly for the nearly 300 million internal migrant workers denied urban public services by the household registration system), high out-of-pocket healthcare costs, modest pensions, state-owned enterprises that retain earnings rather than paying dividends, and financial repression that effectively taxes savers to subsidise borrowers. The plan acknowledges the problem and tinkers at the edges (improved elderly care, expanded childcare, modest insurance reforms) without touching any of these structural pillars.

The reason it does not is political, not technical. The reforms that would raise household consumption (fiscal transfers, registration-system liberalisation, state-enterprise dividend reform, relaxation of financial repression) would each shift the economy’s primary resource-allocation mechanism from state-directed investment toward decentralised consumer spending. For a system whose political legitimacy rests in part on the party’s ability to steer capital toward strategic ends, that is not a policy adjustment but a threat to its organising logic. The household registration system is the clearest case: it simultaneously suppresses consumption and fragments the internal market, and reforming it would address both. The plan mentions it in neither context.

The national-accounts identity makes the political constraint a geoeconomic one. If household consumption stays suppressed and investment is sustained by the fortress programme, excess savings flow abroad as a trade surplus, generating persistent pressure on partner economies and persistent fuel for the protectionist coalitions the forum is supposed to defuse.

The 15th Five-Year Plan is not a blueprint for a new order. It may not even be a document for an interregnum, since that word implies a regnum to follow, and there is no particular reason to assume one will. What we may be looking at instead is a world of overlapping ordering attempts that never cohere into a single architecture. The 15th Plan fits that description uncomfortably well: a programme that optimises for two incompatible outcomes, a world in which the American-anchored network holds and one in which it fractures, without committing to either.

Fortress and forum are not impossible to reconcile in principle, but the sequencing required demands time, demonstrated restraint, and a domestic economy that can absorb its own production. The 15th Plan has the first, has one instance of the second, and lacks the third entirely. The fortress generates the externalities (overcapacity, trade friction, capital-account opacity) that compress the timeline the forum needs, while the macroeconomic rebalancing that might buy time would require the state to transfer resources to households and trust them to spend, which is the one thing the plan’s political logic will not permit. Dual circulation was the conceptual solution. A dual circuit whose internal loop cannot generate adequate demand is just a single circuit with a nationalist paint job.

The rest of the world is beginning to draw its own conclusions. Europe has already imposed tariffs on Chinese electric vehicles. The Global South countries that Beijing courts in the forum are often the same ones whose nascent industries are undercut by the fortress’s surplus output, and they are not passive about it. Brazil, India, and Indonesia have all erected tariff walls and launched anti-dumping investigations against Chinese EVs and steel. The forum is failing not only because of Chinese overcapacity but because the countries Beijing needs as coalition partners are actively prioritising their own industrialisation over cheap Chinese imports. Washington, meanwhile, is converting its own attractive leverage into the coercive variety at an accelerating pace. The optimal response for an incumbent in a reserve-currency competition is the opposite: oversupply safe assets, flood the market, shrink the prize a challenger could capture. Instead, Washington is weaponising the very infrastructure whose openness made it dominant, which paradoxically makes the Chinese alternative more credible even as China’s own contradictions mount. The incumbent and the challenger are both degrading the networks they need, and neither has a domestic political path to the reforms that would stop it. The interregnum, if that is what this is, may last a long time, not because the contradictions are manageable but because no one can afford to resolve them.