Opening
China's controlled nuclear fusion research has reached a critical inflection point, combining scientific breakthroughs with sustained patient capital from state-owned investors. The country's EAST tokamak—China's "artificial sun" facility—achieved a dual milestone of 120 million degrees Celsius ion temperature and 160 million degrees Celsius electron temperature, with fusion parameters reaching the 10^20 scale; the device is scheduled to conduct its first combustion experiment in 2027. Simultaneously, in October 2024, the compact fusion energy experimental device (BEST) project in Hefei completed installation of its critical Dewar base component, with plans to demonstrate net energy output from fusion by 2030. This convergence of technical progress and capital commitment reflects a deliberate strategy to overcome the industry's "forever 50 years" perception—the long-standing joke that commercial fusion is perpetually decades away.
The acceleration stems from a shift in funding dynamics and institutional support. Shanghai's state-owned investment ecosystem, anchored by Shanghai State-owned Investment Co. and its Future Industries Fund, has adopted an explicit "sow the first seeds" approach, deploying capital at the earliest research stages rather than waiting for technology paths to converge. This patient capital model, combined with artificial intelligence accelerating experimental iteration, is reshaping the timeline for fusion commercialization from theoretical speculation to engineered reality.
Tokamak Technology: Engineering at Extremes
The tokamak represents the world's most mature magnetic confinement fusion approach. The device uses magnetic fields to confine plasma—ionized fuel gas—at temperatures between 100 million and 200 million degrees Celsius, a heat no physical container can withstand. The tokamak functions as a magnetic cage, holding this extreme plasma steady.
The engineering challenge is acute. Inside a single tokamak, conditions oscillate between two extremes: microwave and neutral beam heating maintain plasma at over 100 million degrees Celsius, while cryogenic systems keep high-temperature superconducting magnets at minus 200 degrees Celsius or colder. This temperature differential tests reliability, operational stability, and cost control at every component level.
The critical bottleneck lies in manufacturing precision. High-temperature superconducting tape—only approximately two microrons thick—must carry hundreds of amperes of current. Current designs achieve this baseline, but advancing to 1,000 or 5,000 amperes requires optimized tape formulations validated through AI-assisted design and extensive testing. Future fusion plants operating 24/7 will depend on highly intelligent control systems: standardized electrical engineering interfaces, modular architectures, advanced data analysis methods, and specialized AI models trained on high-quality experimental data. AI will analyze experimental data and extract decision-making logic to address core scientific challenges.
Deuterium-Helium-3: An Alternative Path
Fudan University's magnetic confinement fusion team, through startup Dawning Fusion (founded July 2025 in Shanghai), pursues a non-mainstream deuterium-helium-3 fuel route. This approach complements China's mainstream deuterium-tritium research and explores frontier territory in magnetic confinement fusion.
The deuterium-tritium reaction faces two critical obstacles tied to tritium: tritium's short half-life (12.33 years) means it decays into helium-3 naturally, requiring fusion plants to simultaneously burn tritium and breed replacement fuel while preventing leaks. Second, the 14-megaelectronvolt neutrons from deuterium-tritium fusion damage reactor structural materials, necessitating 1–1.5 meter-thick shielding layers.
Dawning Fusion chose the alternative: deuterium-helium-3 produces essentially no neutrons, eliminating the need for expensive, heavy shielding. This enables compact reactor designs based on high-temperature superconducting strong magnetic fields, allowing fusion plants to locate near cities or data centers without long-distance power transmission. If both fuel routes succeed, they will form a complementary integrated grid: large deuterium-tritium stations in remote areas and compact deuterium-helium-3 plants near urban centers. Since tritium decay produces helium-3, deuterium-tritium operations naturally supply helium-3 for the alternative route.
Dawning Fusion plans a 10-year timeline across three device generations. The first-generation device, "Chenguang," will validate high-temperature superconducting magnet reliability and stability under real operating conditions, while serving as a "fusion AI data factory" generating massive experimental datasets to verify and optimize physics models under strong magnetic fields and support intelligent device control development.
Shanghai's Capital Ecosystem
Shanghai has constructed a full-stack, systematic fusion industry ecology. The ecosystem encompasses multiple research teams (Dawning Fusion, Xinghuan Fusion Energy, Energy Singularity) and supply-chain enterprises (Shanghai Superconductor, Shang'ai Superconductor, Yixi Technology), creating a "up-and-down-the-stairs" integrated supply chain within the city.
Shanghai State-owned Investment Co., through its Future Industries Fund and Shanghai Sci-Tech Innovation Group, covers the entire capital chain from angel rounds through IPO, providing continuous support. Zhu Min, Chief Innovation Officer of Shanghai State-owned Investment Co. and Chairman of Shanghai Sci-Tech Innovation Group, frames the state capital mandate as "boldly sowing the first seeds." Traditional social capital waits for technology paths to clarify before entering; state capital cannot follow that model. "If state capital does not sow seeds first and bear risk, this ecosystem may disappear entirely, and this technology route will halt mid-development." State capital must "quickly fill critical gaps, stand firmly on the front line of technological innovation, dare to bet, dare to act, and dare to sow the first seeds."
Under state capital's "priming," the fusion industry has gained strong momentum. Various capital types now fear missing out, accelerating ecosystem growth. The state's strategic task, Zhu Min notes, is ecosystem and supply-chain construction; if difficulties arise, state-owned capital clusters will provide systematic, targeted support.
Fusion startups receiving this capital do not hoard it but redistribute through orders and technology to upstream and downstream supply-chain partners, driving coordinated development. Prof. Xu Min (Fudan University) emphasizes that supply-chain importance equals that of fusion devices themselves: long-term, supply-chain economics determine fusion value. Industry-wide health benefits individual enterprises. The path from Q>1 (net energy gain) to the first watt to ultimately one-cent-per-kilowatt-hour electricity requires long-term accumulation. Challenges include reducing high-temperature superconducting tape costs, advancing high-power gyrotron and neutral beam heating technology, and ensuring ion cyclotron technology meets future plant demands. These require full supply-chain coordination.
Timelines and Commercial Pathway
Prof. Xu Min projects that "the first electricity from a laboratory can be realized in about five years, and it is very likely to be realized in China." From that first watt to cost-competitive fusion power, a 20-year timeline is reasonable for supply-chain cost reduction and maturation.
Wei Fanjie, General Manager of Shanghai Future Industries Fund, notes that "although precise prediction of fusion commercialization timing remains impossible, the process has dramatically accelerated. In the past, people joked that fusion is always 50 years away, but this time may truly be different." Risk capital is now entering basic research at scale—2025 fusion startup financing alone may exceed 2 billion yuan, compared to roughly 2 billion yuan total across a decade of academic fusion research. Capital deployment efficiency has risen sharply.
Wei highlights that venture capital is moving forward while basic research moves backward, blurring the boundary between fundamental science and commercialization. AI intervention is dramatically accelerating research iteration. Talent influx is raising talent density; cultivating young researchers with "AI-native" thinking is especially critical, as they bring forward-looking perspectives and reshape iteration models.
Zhu Min frames the 50-year question in historical context: for an individual, 50 years demands near-lifetime effort; in human history, 50 years is a moment. Even if multiplied several times, it remains brief on civilization's scale. But distance should not prevent departure. The fusion sector must remain clear-eyed: science is inherently difficult, technology paths remain unconverged, and which route reaches engineering and application is unclear. Yet this sector is strategically vital to national positioning and strategic lifelines. With AI assistance, improved engineering validation, and technology-path convergence, humanity is narrowing the distance to fusion commercialization and away from the "forever 50 years" verdict.
"Today's mature industries were yesterday's future industries," Zhu concludes. "Without confidence and passion for future technology, progress halts. Because we believe, we see. Belief requires unwavering commitment. State capital, as the most steadfast and patient supporter and companion, stands ready with scientists and teams who see the future, hold confidence, and walk with resolve—together shortening the distance to innovation and reducing its difficulty."
