China’s Institute of Modern Physics, part of the Chinese Academy of Sciences, is building what it calls the world’s first megawatt-level accelerator-driven subcritical system, a nuclear facility designed to turn thorium into a viable fuel source that could supply energy for centuries. The China initiative Accelerator Driven System, known as CiADS, is under intensive construction in Huizhou, Guangdong province, with a staged technical roadmap targeting multi-megawatt beam power within the next few years. If the facility performs as designed, it could reshape how nations think about nuclear fuel cycles and long-lived radioactive waste.
What CiADS Actually Is and Why It Matters
Most of the world’s nuclear reactors run on uranium, a finite resource concentrated in a handful of countries. Thorium, by contrast, is roughly three to four times more abundant in Earth’s crust and is distributed across dozens of nations. The catch is that thorium cannot sustain a nuclear chain reaction on its own. It needs an external source of neutrons to convert it into fissile uranium-233, which can then produce heat and electricity. That is exactly what an accelerator-driven system does: a powerful particle accelerator fires protons into a heavy-metal target, generating a shower of neutrons that drive the subcritical reactor core. Because the core cannot run without the accelerator beam, the system has an inherent safety advantage. Shut off the beam, and the reaction stops.
CiADS is not a commercial power plant. It is a principle-verification device, built to prove that this technology works at meaningful scale. The facility’s own institutional description labels it the first “megawatt-level ADS principle-verification device” internationally, a claim repeated on event pages hosted by the Institute of Modern Physics. That distinction matters because previous ADS experiments around the world have operated at far lower power levels, too small to demonstrate the neutron economics and materials performance needed for a real thorium fuel cycle. Reaching megawatt-class beam power would be the threshold where engineers can credibly test whether thorium breeding works efficiently enough to matter for energy policy.
A Staged Path to 2.5 Megawatts
The technical blueprint for CiADS follows a deliberate, step-by-step escalation. A peer-reviewed paper in Physical Review Accelerators and Beams describes the facility’s staged path toward 500 MeV operation. The linac, or linear accelerator, begins with lower-energy proton beams and scales upward through successive hardware upgrades. Each stage validates the beam dynamics, radio-frequency cavity performance, and cryogenic systems before the next energy jump. The roadmap outlined there expects the facility to reach approximately 2.5 MW of beam power by around 2027, placing CiADS in the multi-megawatt class that no other ADS facility has achieved.
That 2.5 MW target is not arbitrary. At that power level, the proton beam can generate enough spallation neutrons to sustain meaningful transmutation rates in the subcritical blanket. For a general reader, the practical translation is straightforward: below a certain beam power, a thorium ADS is a science experiment. Above it, the system can actually convert thorium into usable fuel and simultaneously destroy some of the most dangerous long-lived waste isotopes produced by conventional reactors. The difference between a few hundred kilowatts and multiple megawatts is the difference between a laboratory curiosity and a technology that could, in principle, extend nuclear fuel supplies by orders of magnitude.
Engineering a Safety Net for High-Power Beams
Running a continuous-wave proton beam at megawatt power introduces serious engineering hazards. If the beam drifts or the target system fails, the concentrated energy can damage components in milliseconds. A separate peer-reviewed study in Nuclear Instruments and Methods in Physics Research Section A details the fast protection system designed specifically for CiADS. The paper describes a timing and protection architecture for high-power continuous-wave beams, including rapid beam-abort mechanisms that can cut the proton stream before damage propagates. This kind of systems-engineering work is often invisible to the public but is the difference between a facility that operates safely for decades and one that suffers catastrophic downtime early in its life.
The existence of this protection-system study also serves as independent engineering evidence that CiADS is being built to megawatt-scale specifications, not merely proposed on paper. Designing and publishing a fast protection architecture implies that the accelerator hardware has advanced far enough to require real safety interlocks, a sign that construction has moved well past the conceptual phase. For countries watching China’s progress, this paper trail matters: it signals that the technical risks of high-power ADS operation are being addressed through peer-reviewed, internationally visible work, rather than behind closed doors.
Construction Progress and Expanding Ambitions
The Institute of Modern Physics has not been quiet about the facility’s status. A workshop announcement on muon source development for CiADS, scheduled for late January 2026 and hosted on the institute’s own conference system, describes the project as under intensive construction. The fact that researchers are already planning secondary science programs, such as muon source applications, around the facility suggests confidence that the primary accelerator infrastructure is on track. Muon beams are produced as a byproduct of high-energy proton collisions with targets, so planning a muon science program presupposes that the proton beam will be available at sufficient energy and intensity.
This expansion of the scientific mission beyond pure ADS verification is worth watching. If CiADS can serve double duty as both a thorium-cycle demonstrator and a muon source for materials science, condensed-matter physics, and other applications, the facility’s cost-benefit calculus improves considerably. Multi-use national laboratories, such as the Spallation Neutron Source at Oak Ridge in the United States or J-PARC in Japan, have shown that shared beam time across disciplines can attract broader funding and political support. By positioning CiADS as both an energy-research platform and a high-intensity particle source, China appears to be following that playbook to maximize scientific and strategic returns.
What Stands Between CiADS and a Thorium Future
Even if CiADS hits its performance targets, major hurdles stand between a successful demonstration and a thorium-powered grid. One challenge is fuel-cycle infrastructure: turning thorium into uranium-233 requires not only neutron-rich irradiation but also chemical processing to separate the bred fuel from other actinides and fission products. That reprocessing step raises its own safety, environmental, and proliferation questions, especially because uranium-233 can, in principle, be used in weapons if not properly denatured. A megawatt-class ADS can show that thorium breeding is technically feasible, but it does not by itself resolve how countries will regulate and secure the resulting materials over decades.
Another barrier is economics. Building and operating a high-power accelerator is expensive, and the added complexity of coupling it to a subcritical reactor core will not come cheaply. Proponents argue that the ability to consume long-lived nuclear waste and vastly extend fuel resources could offset those costs, particularly for countries with limited uranium deposits but ample thorium. Skeptics counter that rapidly falling costs for renewables and energy storage may erode the business case for such intricate nuclear systems. CiADS, as a research facility, is not meant to settle the economic debate, but the data it generates on reliability, maintenance, and operating efficiency will heavily influence whether future policymakers see ADS-based thorium reactors as a niche curiosity or a serious option.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
