China accelerates nuclear reactor and fusion push as AI power demand rises
China is stepping up investment in nuclear fission and fusion as fast-rising electricity demand from AI data centers adds pressure to expand power supply and strengthen energy security. The push also supports Beijing’s broader goals in clean energy, technology self-sufficiency and export-oriented industrial development as tensions with the U.S. remain high.
Highlights
- China accounts for over half of the 58 new global nuclear plants under construction and aims to quadruple data center power use to 774 terawatt hours by 2030.
- Shanghai-based Junhe Atomic raised Rmb300mn and targets deployment of small modular reactors at coal plant sites, with demonstration operations planned for around 2031.
- Fusion startup NovaFusionX, valued at over Rmb10bn and backed by Rmb2.4bn investment, aims for first plasma in 2024 and net energy gain by 2029.
AI-driven power demand reshapes nuclear strategy
As reported by Financial Times, China is intensifying its nuclear build-out as the boom in energy-hungry AI infrastructure increases long-term demand for reliable electricity. More than half of the 58 new conventional nuclear plants under construction worldwide are in China, while the country is also investing in small modular reactors and expanding research into nuclear fusion.China’s electricity demand is reaching record levels as manufacturing and exports stay strong, even as other parts of the economy slow. Forecasts from consultancy Wood Mackenzie show the country’s data centers are expected to quadruple electricity consumption to 774 terawatt hours by 2030, pointing to a structural rise in computing-related power demand driven by AI workloads.
Damien Ma, a non-resident expert at Beijing-based research group Gavekal, says China is set to have the world’s most dynamic and significant nuclear industry over the next decade by a wide margin. He says China is seeking to deepen supply-chain self-sufficiency, turn domestic capability into an export industry and pursue fusion power, while the U.S. faces capital and regulatory constraints in sustaining a similar build-out.
Private ventures expand reactor and fusion ambitions
Private companies are also moving into new nuclear technologies, often with ties to state-backed capital and research networks. Junhe Atomic, a Shanghai-based small modular reactor developer founded in 2025 by former state nuclear executive Tian Jiashu, has raised Rmb300mn, including backing from HongShan Capital, and is in the design phase for its first demonstration reactor.An executive at Junhe says the company aims to have its demonstration project operating around 2031, although engineering issues remain in structural components, modularisation and process efficiency. The company sees potential to deploy smaller reactors at sites now occupied by ageing coal-fired power plants, linking the technology to the eventual replacement of thermal power assets and infrastructure.
In fusion, Shanghai-based NovaFusionX is developing a field-reversed configuration system under founder Guo Houyang, whose past work includes research roles in the U.S. at the University of Washington and TAE Technologies. Guo says the company aims to achieve first plasma by the end of this year and targets a net energy gain by 2029, with plans for a first demonstration reactor in the early 2030s costing about Rmb3bn.
NovaFusionX has attracted Rmb2.4bn from investors and has a valuation of more than Rmb10bn. Research from the Special Competitive Studies Project, a U.S. think-tank, says China has mobilized roughly three times as much investment into commercialization-relevant fusion projects as comparable efforts under the U.S. energy department, and fusion is one of eight priority technologies in the Communist party’s latest five-year economic plan.
In our earlier report, we covered Morgan Stanley’s warning that U.S. data centres could face a widening electricity shortfall as AI chip production accelerates faster than power capacity. The analysis highlighted how the gap could become a major constraint on data-centre expansion, underscoring the need for significant new generation and grid investment to meet fast-rising AI-driven demand.