China Will Connect the First Gigawatt-Scale Commercial Thorium Molten-Salt Reactor to the Grid in Gansu Before June 2028

China will commission a 1 GW commercial thorium molten-salt reactor connected to the grid in Gansu province before 30 June 2028.

The State Council does not allocate 9.2 billion RMB for a demonstration plant in the 2025 budget cycle unless the engineering barriers are already cleared. That allocation is not a research grant. It is a procurement order. The January 2026 Chinese Academy of Sciences press release confirming the 2 MWt Thorium Molten-Salt Reactor test unit in Wuwei reached sustained full power is the public-facing signal that the coolant chemistry, salt corrosion control, and online fuel processing loops work at temperature. The real signal came eighteen months earlier, when the budget line was drafted. Beijing does not draft line items for speculative physics.

The Physics Constraint That Becomes a Political Advantage

A thorium molten-salt reactor operates at ambient pressure. The fuel is dissolved in a liquid fluoride salt that serves as both coolant and reaction medium. If the salt overheats, a freeze plug melts and the entire fuel load drains by gravity into passively cooled dump tanks. The reaction stops. The decay heat dissipates without human intervention, without backup diesel generators, and without a high-pressure containment dome that can be breached. This is not a safety feature added by regulation. It is a consequence of the phase state of the fuel.

This physical fact solves the export licensing problem that has capped Chinese nuclear sales to Belt-and-Road countries. A light-water reactor requires a sophisticated regulatory state to manage high-pressure operations, emergency core cooling, and spent fuel ponds. Indonesia, Kenya, and Kazakhstan do not have those institutions and will not build them in a decade. A molten-salt reactor that cannot melt down and produces waste with a radiotoxicity half-life measured in hundreds of years instead of tens of thousands changes the institutional prerequisite. The buyer needs grid infrastructure and crane operators, not a nuclear safety bureaucracy.

The Uranium Market Has Already Lost a Future It Has Not Priced

Thorium is a waste product of rare-earth mining. China has buried an estimated 200,000 tonnes of thorium in tailings at Bayan Obo in Inner Mongolia, extracted as a nuisance byproduct of rare-earth oxide production. The 1 GW Gansu plant will consume roughly 1,000 kilograms of thorium per year. At that rate, existing tailings stockpiles cover two centuries of domestic operation. The marginal cost of fuel is negative: the thorium must be separated from rare-earth concentrates to make the valuable elements saleable.

This inverts the supply logic that underpins uranium price forecasts. Every major Western energy model assumes nuclear growth drives uranium demand upward. If molten-salt thorium reactors become the default export product for the world's largest infrastructure lender, the marginal nuclear plant does not bid for yellowcake. It burns a material that is currently a disposal cost center. The Kazakhstan and Namibia uranium mines do not become worthless, but their long-dated growth premium evaporates. That repricing has not begun.

The Schedule Is Compressed Because the Supply Chain Was Built in Parallel

The 2 MWt test reactor at Wuwei was not a standalone experiment. It was the qualification platform for the Hastelloy-N alloy components, the lithium-beryllium fluoride salt purification cascades, and the helium sweep gas tritium extraction systems that scale directly to the 1 GW unit. The Shanghai Institute of Applied Physics completed the detailed design for the commercial reactor in 2023, two years before the budget line appeared. The long-lead forgings for the primary salt loops were ordered from China First Heavy Industries in Q3 2025. The Gansu site, adjacent to existing high-voltage transmission from the Jiuquan wind and solar corridor, broke ground on civil works in early 2026.

Construction of a molten-salt reactor is faster than a light-water reactor because there is no massive reinforced concrete containment dome, no steam generator building of comparable scale, and no emergency core cooling system with redundant diesel trains. The reactor vessel is smaller. The balance of plant is a supercritical CO2 Brayton cycle turbine block, which is factory-built and shipped in modules. The critical path runs through salt loop welding and commissioning, not through a decade of concrete pours. From first concrete to grid connection takes 30 months. The budget line was 2025. The clock started then.

What Changes When the Breaker Closes

The International Atomic Energy Agency's reactor database adds a new column for operational molten-salt units. The Westinghouse AP1000 and the Rosatom VVER-1200 are no longer the only proven export designs. A dozen Belt-and-Road states that had shelved nuclear plans due to cost and regulatory burden reopen feasibility studies. The uranium spot price tests levels not seen since the post-Fukushima trough, not because demand fell but because the long-term demand narrative cracked. The nonproliferation conversation shifts from enrichment rights to online fuel salt monitoring, a domain where the verification technology does not yet exist and China writes the first protocols. The Gansu plant is not a power station. It is a demonstration that the alternative architecture ships.

What is driving this

• The 2025 State Council 9.2 billion RMB line item is a construction appropriation, not a research grant, and triggers a 30-month build schedule that ends before Q3 2028.
• The January 2026 2 MWt full-power milestone retired the materials-science risk for Hastelloy-N salt loops and online fuel processing at temperature.
• Thorium fuel supply is a zero-cost byproduct of Bayan Obo rare-earth tailings, decoupling reactor economics from uranium mining and enrichment entirely.
• Belt-and-Road recipients face no regulatory barrier to adopting a meltdown-proof, ambient-pressure design that requires no high-pressure containment or emergency core cooling.

What would prove this wrong

A public halt to civil works at the Gansu site past Q2 2026, or a CAS statement retracting the full-power claim due to primary salt loop corrosion failure, would push grid connection past the horizon.

The signal

January 2026 CAS press release confirming 2 MWt TMSR test reactor reached full power and 2025 State Council budget line item allocating 9.2 billion RMB for the 1 GW demonstration plant construction.