Foreword

For decades, we have assessed China’s plutonium production for nuclear weapons.  For the last five years, the U.S. government has consistently claimed that China plans to increase its nuclear weapons arsenal four-to-fivefold from its 2000 to 2010-levels. This statement, as well as China’s ability to increase its arsenal so dramatically, begged our attention. 

Further, when we learned of China possibly using civilian breeder reactors to make plutonium for nuclear weapons, we were reminded of French efforts to do the same in the 1980s, an activity we opposed.  The use of civilian nuclear plants to make materials for nuclear weapons has always attracted our scrutiny and opposition. 

The world should fear another country with a large nuclear arsenal, at a time when the nuclear non-proliferation regime risks collapse.  Other countries are bound to take note of such a large, arsenal, particularly given China’s threats towards Taiwan.  This report is our way to engage in this debate and encourage counterinitiatives.  

 

Highlights

Over the last several years, the U.S. Department of Defense (DOD) has been assessing that China may need to produce new plutonium for nuclear weapons to grow its nuclear arsenal to include over 1000 nuclear weapons.  This assessment remains uncertain due to a lack of information about China’s existing stock of plutonium for nuclear weapons, the number of nuclear weapons China will actually build over the next five to ten years, and China’s plutonium requirements for its individual weapons. 

China has the means to make more plutonium for nuclear weapons, a type of plutonium typically referred to as “weapon-grade.” We evaluated two of the most likely potential sources of weapon-grade plutonium: the civilian China Fast Reactor 600 (CFR-600), and the old 821 Plant reactor that was built several decades ago as part of China’s military plutonium complex.  A CFR-600 can produce weapon-grade plutonium, estimated as 130 to 165 kilograms per year; the 821 Plant reactor could make 160 to 200 kilograms of weapon-grade plutonium per year. 

The military use of the weapon-grade plutonium produced in the CFR-600 would be an economical, efficient method to produce hundreds of nuclear weapons.  This plutonium is produced in the reactor’s “blanket,” a region surrounding the reactor core, as a byproduct of reactor operation.  The blanket fuel is composed of uranium oxide, comprised of almost exclusively uranium 238, of which a fraction becomes the desired plutonium 239 upon irradiation in the reactor.  The blanket can be removed from the reactor core when it is lightly irradiated, making it relatively easy to chemically process and separate plutonium. Making this option even more attractive for China, the bulk of the blanket can not only be processed at a new Chinese reprocessing plant under construction in Gansu, but also at already existing reprocessing plant(s), designed to separate “reactor-grade” plutonium from highly irradiated enriched uranium oxide fuel discharged from pressurized water reactors (PWRs).  

Some have questioned the efficiency of using the CFR-600 for producing weapon-grade plutonium for nuclear weapons.  However, the usefulness of such reactors for nuclear weapons programs was well documented in the debates about such reactors in the 1980s, when several countries were pursuing them (all but Russia abandoned those plans).^[1]  Moreover, Rosatom, the Russian state corporation that controls Russia's entire nuclear industry, appears aware of the proliferation risk posed by a fast reactor blanket, given that Rosatom’s Special Representative for International, Science, and Technology Projects stated that in order to reduce the proliferation risk, fast reactors in nonnuclear weapon states should have “processing technologies without PU [plutonium] production in the FNR [fast neutron reactor].”^[2]  

China has needed and received Russia’s assistance to build fast reactors.  Russia provided significant assistance to China’s experimental fast reactor and additional assistance to the CFR-600.  Russia has a mature fast reactor program, demonstrated over many years, but it has not managed to master the recycling of plutonium fuels in fast reactors.  As a result, Russia and China are currently cooperating in this area. 

China claims that the CFR-600 is a strictly civilian reactor, and Russia has imposed peaceful use constraints on at least the first facility, referred to as a demonstration reactor in the China/Russia agreement for cooperation.  However, China’s lack of nuclear transparency and the less-than-watertight agreement with Russia on the first CFR-600 reactor raises legitimate questions about its use to make weapon-grade plutonium for nuclear weapons.  Due to the China/Russia agreement only referencing the first, demonstration reactor, the concern of military use is heightened when considering the CFR-600 unit 2, already under construction, or successor units.  The peaceful use constraints as set out in the China/Russia agreements are less verified than the non-proliferation restrictions imposed by the United States on civilian nuclear cooperation with China.  Any changes may be subject to secret negotiations, and identifying violations and remedying them would depend on Russian enforcement.  Russia has extensive nuclear cooperation with China, far beyond the CFR-600 reactors.  Would Russia be willing to jeopardize all of its cooperation to confront China if it diverted blanket plutonium to nuclear weapons?  Would Russia even know this took place?  There is worry that the likely answer to both questions is no.  

Another potential source of plutonium for weapons is the 821 Plant reactor.  While it is widely accepted that the 821 reactor at Guangyuan stopped producing plutonium for weapons by 1990, there are indications that the reactor was shut down but could be restarted.  Contradicting this statement partially, some knowledgeable Western officials assessed that the reactor could have continued operating after 1990.  Activity at the reactor site observed in optical and thermal imagery over the years indicates on-going maintenance activities to allow a restart, or even recent operation.  Thermal activity indicates reactor operation, or another unexplained, nearly continuous heat source emanating from or near the reactor building, the former perhaps related to the irradiated graphite moderator. 

In fact, there are multiple indications that the reactor may have operated over the last decade, or may have been prepared for upcoming reactor operation.  It may have been converted to produce electricity, and changes around the reactor indicate an intention of electricity generation in more recent years. The status of the nearby reprocessing plant is also unclear, but thermal imagery suggests that it may be operational. 

Overall, the Plant 821 reactor is an enigma as a consequence of China’s lack of transparency.  The indications suggest that the reactor could operate in the future, but this conclusion is highly uncertain.  Weapon-grade plutonium production at Guangyuan may pose significant challenges, although comparatively fewer than building a new production reactor. 

DOD projections assess that China is aiming for an arsenal of 1500 nuclear weapons.  We tested whether China has enough weapon-grade plutonium in its existing stock, all produced prior to the early 1990s, when it ended such production, to make 1500 nuclear weapons.  The calculations are complicated by several uncertainties, such as how much weapon-grade plutonium did China produce for nuclear weapons purposes, and how much plutonium on average is in Chinese nuclear weapons?  Although our analysis shows that it is difficult to determine how far existing Chinese weapon-grade plutonium stocks can go in meeting China’s longer term needs to build a total of 1000 or 1500 nuclear weapons, it is likely that — if China wants to build more than 1000 nuclear weapons — it will need a new source of plutonium.  The alternative for China is to revert to weapon-grade uranium instead, or better, to rely only on its current plutonium stock and limit the size of its nuclear arsenal. 

Just one of these new sources of plutonium would be enough to satisfy China’s projected plutonium requirements.  One or two CFR-600s would be a better long-term (longer than a decade) source of plutonium for weapons from technical, economic, and convenience perspectives, and in fact allow an arsenal far greater than 1500 weapons.  The challenges of operating the 821 reactor are assessed as more severe than harvesting weapon-grade plutonium from the blankets of the CFR-600.  

Nonetheless, China should be discouraged from misusing the CFR-600 for military purposes.  China should be reminded that it is both a recipient and provider of civilian nuclear technology, with ambitions to export nuclear reactors all around the world.  If it does not take its own peaceful constraints seriously; how can China expect countries that receive Chinese civil nuclear assistance to take them seriously? 

China’s use of a CFR-600 reactor for military purposes may fundamentally depend on whether Russia enforces the peaceful use restrictions in its nuclear cooperation agreements with China.  But the Russian government, bogged down in its war in Ukraine and highly dependent on China for vital dual-use goods and oil sales, is currently unlikely to do so, although pressing them to do so would be advisable.  A potential ally could be Russia’s nuclear industry, which is unlikely to have an interest in seeing breeder reactors militarized.  It should be asked to use its leverage to dissuade China from doing so with the CFR-600 reactors. 

France and the United States may further be able to request that China provide assurances that it will not militarize a CFR-600 on the basis that the plutonium used to fuel the reactor in the future could have been produced in French-supplied pressurized water reactors (PWRs) or Chinese-built reactors containing French reactor technology.  Westinghouse also provided China with PWRs, and the United States should seek assurance that any plutonium produced in these reactors, or reactors subsequently built containing Westinghouse reactor or fuel technologies, will not be used to fuel a CFR-600 with military purpose.  

 

Introduction

China’s ambitious plan to vastly increase its nuclear weapons arsenal reportedly depends on establishing a new supply of fissile material, particularly weapon-grade plutonium.  According to the U.S. Department of Defense (DOD), China “probably will need to start producing new plutonium this decade to meet the needs of its expanding nuclear stockpile.”^[3] 

This DOD assessment is set against the backdrop that it believes China has accelerated its nuclear weapons expansion, a view shared, at least in part, by those often critical of DOD assessments on China’s nuclear arsenal, which categorizes China’s nuclear expansion as “among the largest and most rapid modernization campaigns” of the five acknowledged nuclear weapon states.^[4]  This view of a growing nuclear arsenal, requiring new capabilities, is strengthened by China’s expansion at the Pingtong Nuclear Facility, which is China’s only identified nuclear weapons components manufacturing and assembly/disassembly facility, indicating that China has improved its capacity to produce increased numbers of finished nuclear weapons, and recover old components in support of nuclear force modernization efforts.^[5] 

However, where China would produce additional weapon-grade plutonium is not immediately obvious.  There are several options.  The most widely publicly discussed source, including by the U.S. DOD, is a site containing two new civilian sodium-cooled, fast reactors, often referred to as “breeder” reactors, named the China Fast Reactor (CFR-600), units 1 and 2, each of which have a rated electrical power of 600 megawatt-electric (MWe) and a total power of 1500 megawatt-thermal (MWth).  China increased suspicion over the possible military use of these sites when describing the CFR-600 reactor as a “national defense investment project,” subject to military nuclear facility regulations, as reported by the DOD.^[6]  Unit 1 started test operations by the end of 2023 and has achieved or is nearing operational status; unit 2 is still under construction.  

China could also restart or use an old military production reactor at the 821 Plant in Guangyuan, codenamed the “Third-Line” reactor, that produced weapon-grade plutonium for nuclear weapons until about 1990. The actual date remains at issue. There are no official reports of the 821 reactor having been permanently shut down and dismantled, i.e., decommissioned. High levels of activity at the reactor site over the last decade indicate it may have been modified to produce electricity and may be operating today, or is being prepared for operation, allowing the production of weapon-grade plutonium in the future.  

Other potential plutonium sources for nuclear weapons include China’s large nuclear PWRs and several research reactors, including a small experimental fast reactor called the CFER.  But these all have more serious drawbacks to their use for military purposes and do not appear practical or economical.  For example, reactor-grade plutonium produced in PWRs and separated in its reprocessing plants could in theory be used in nuclear weapons, but China is unlikely to follow this path, because it would significantly affect its nuclear weapons designs and their reliability. Alternatively, a large PWR could be converted to make large amounts of weapon-grade plutonium, using a system of enriched uranium driver rods with an enrichment of 10-20 percent and natural uranium targets, in which the weapon-grade plutonium is produced. However, this conversion could be costly and generate domestic and international controversy, as the United States discovered in the late 1980s when it considered, and then rejected, using a large PWR to meet its own projected plutonium needs for weapons. 

 

The CFR-600

The CFR-600 is located on Changbiao Island in Xiapu County of China’s Fujian province.  Construction of the first, demonstration CFR-600 began in 2017.  The reactor reportedly started operation in mid-2023.^[7]  Satellite imagery confirms that the reactor is operating or at least nearing operation. Discharge of water from the cooling system is visible in imagery dated to August 8, 2024, September 10, 2024, and April 6, 2025, but it cannot be established whether the water flow indicates reactor operation or prolonged testing of the cooling system (Figure 1). 

A second, identical-looking fast reactor has been under construction since 2020 at the same site and is reportedly expected to begin operations in 2026. 

The responsible Chinese company is the China Institute of Atomic Energy (CIAE).  It was created by the China National Nuclear Corporation (CNNC), which has built a large number of pressurized water reactors (PWRs), based on Western PWR technology, that rely on enriched uranium fuels.  China presents the CFR-600 as a civilian facility, with the added twist that it is necessary to achieve carbon neutrality.^[8] 

The two reactors are reportedly designed with a shell, or “blanket,” of natural or depleted uranium oxide surrounding the entire core, located radially around the reactor core and above and below the fuel. The blanket is designed to capture excess neutrons leaving the core and to produce plutonium.  When the core is fueled with plutonium, the reactor would be designed to produce more plutonium than it consumes; this gain depends critically on the new plutonium produced in the blanket.  Because of this property, this type of reactor has historically been called a “breeder” reactor.  

Due to a lack of plutonium fuel fabrication capabilities in China and Russia, Russia has provided coreloads of highly enriched uranium instead of plutonium for the CFR reactor.  However, the Chinese plan calls for the regular use of plutonium to fuel its CFR-600 by the mid-2030s. 

Figure 1. Top image shows the reactor site as of December 2022, with reactor insets as of August 8, 2024.  Construction of the second CFR-600 has visibly progressed.  Bottom image shows cooling water discharge in September 2024. Coordinates: 26°48'12.59"N, 120° 9'18.66"E. 

Like Russia, China has announced its long-term plan to reach a “closed” nuclear fuel cycle, a term of art used to describe the combined recovery and recycling of plutonium in irradiated fuel in the fast reactor core and newly produced plutonium in the blanket.  The closed fuel cycle encompasses the fast reactor, breeder fuel reprocessing, and plutonium recycling capabilities, principally a plutonium fuel manufacturing facility making plutonium/uranium, aka mixed oxide (MOX) fuel. 

Breeder reactors, such as the CFR-600, are controversial due to their commercially unproven technologies, especially the operational safety problems created by using highly reactive sodium as reactor coolant and the difficulties of repeated plutonium fuel reprocessing and plutonium recycling campaigns necessary to establish a closed fuel cycle. While Russia has succeeded in building fast reactors, it has until relatively recently fueled them with highly enriched uranium (HEU).  Several countries have tried to build large-scale plutonium-fueled breeder reactors and failed, including France, Germany, Japan, and the United States, despite large expenditures.  

Both units 1 and 2 are expected to operate with uranium blankets likely designed for the production of super-grade plutonium (less than 3 percent plutonium 240), which is high-quality plutonium ideal for producing nuclear weapons.  While intrinsic to the notion of a breeder reactor, this high-quality plutonium is of great value to a nuclear weapon state seeking to build more nuclear weapons.  

 

Super-Grade Plutonium Production in the CFR-600

China has not published details about the amount of plutonium to be produced in the blanket of the demonstration CFR-600 and its nearby twin.  It has also not stated the expected isotopic content of the plutonium in the discharged blanket material.  Normally, the plutonium would rank as weapon-grade plutonium (less than 6 percent plutonium 240, or greater than 94 percent plutonium 239), and likely super-grade plutonium (less than 3 percent plutonium 240), a higher quality weapon-grade plutonium.  

The amount of plutonium produced annually in the blanket, assuming expected, full-power operation, can be estimated.  The French 1240 MWe Superphénix reactor, which had a thermal and electrical power twice as large as the CFR-600, was designed to produce 330 kilograms of super-grade plutonium per year.^[9]  Scaling this value to the power of the CFR-600 results in an annual production of 165 kilograms of super-grade plutonium.  A more sophisticated method, applied by an Asian nuclear expert, scales up the defunct, 280 MWe Japanese Monju breeder reactor, arriving at a value of 130 kilograms of super-grade plutonium per year.  Here, a range of 130 to 165 kilograms per year will be utilized, with an average of 148 kilograms per year. 

The plutonium will be treated in this report as weapon-grade plutonium, but if it is super-grade plutonium, it can be blended down with non-weapon-grade plutonium to yield more weapon-grade plutonium.  For example, if a feedstock of plutonium containing 90 percent plutonium 239 was blended with the output of the blanket, assumed to be super-grade containing 97 percent plutonium 239, then a resulting stock of weapon-grade plutonium containing 94 percent plutonium 239 would be 75 percent larger. So, the 130 to 165 kilograms of super-grade plutonium per year derived above could generate 228 to 289 kilograms of weapon-grade plutonium, rounded to 230 to 290 kilograms per year.  

The second CFR-600 would double all these numbers.  Ignoring gains from blending, even one of these two reactors would represent a formidable weapon-grade plutonium capability, a capability that is a low-cost by-product of a nominally civilian electricity production project. 

 

Cooperation with Russia

What Has Russia Provided to China for the Demonstration CFR-600?        China’s expertise on fast reactors grew out of its own experience and cooperation with Russia, the world’s leader in fast reactors with both the BN-600 and BN-800 fast reactors operational.  Its cooperation with Russia on fast reactors started in earnest with the 65 MWth experimental fast reactor, the CEFR.  

The 2000 agreement between Russia and China for the CEFR specified cooperation on(1) the design of the CEFR, including the core and its fuel assemblies, equipment and instrumentation, monitoring, and control systems; (2) development and research on the CEFR core and its fuel elements and assemblies, main circulation pump,  and emergency equipment; (3) the supply of key equipment, such as main circulation pumps and instrumentation; (4) participation of Russian specialists in installation and operation of the supplied Russian equipment as well as Chinese equipment; (5) highly enriched uranium (HEU) fuel supply; (6) technology for the fabrication and use of uranium and uranium/plutonium fuels; (7) personnel training in reactor operation at Russian fast reactors; (8) nuclear safety; and (9) technical assistance in the design, equipment development, and construction of the CEFR. 

The cooperation was so extensive that China gained enough technology and experience that it reportedly designed the first unit of the CFR-600 itself.  China had initially intended to buy BN-600 fast reactors from Russia, but chose a more domestic path instead. 

Nonetheless, Russian assistance continued.  In addition to an across-the-board intention to cooperate, as exemplified in their November 2016 joint statements on the “development of strategic cooperation in the field of the use of nuclear energy for peaceful purposes,”^[10] China and Russia continued to cooperate on breeder reactors. 

In December 2018, China and Russia reached a cooperation agreement on the CFR-600, which was referred to as a demonstration reactor in the agreement. According to the agreement, Russian enterprises have made significant contributions to the demonstration CFR-600 reactor, including project expertise; key reactor plant equipment and nuclear fuel handling equipment; calculation codes; technical documentation; participation in installation, supervision, and commissioning; and staff training.  In addition, China pledged to procure fuel assemblies containing highly enriched uranium or MOX fuel throughout the lifetime of the demonstration reactor, in accordance with the needs of the Chinese reactor operator and the capabilities of Russia.  

However, neither country can produce plutonium fuels.  As a result, as mentioned above, Russia is supplying highly enriched uranium fuel for the first unit of the CFR-600.  Russia’s TVEL Fuel Company, a subsidiary of Rosatom, plays a key role in Russia’s supply of HEU fuel for unit 1 of the demonstration CFR-600; a contract signed between TVEL and China’s CNNC in January 2019 committed TVEL to provide HEU fuel for the CFR-600’s first seven years of operation.^[11]  TVEL’s Elemeash Machine Building Plant in Elektrostal began producing HEU fuel for the CFR-600 in 2021.  This facility delivered three batches of fuel by 2022, for the first core loading as well as the bundles for the first refueling of the reactor.^[12] 

Later, the plan calls for the use of plutonium fuels in the CFR-600 and any subsequent breeder reactors.  The creation of a closed fuel cycle for the CFR-600 reactors, namely reprocessing breeder fuel and recycling plutonium, requires both a plutonium fuel fabrication plant and a plutonium fuel reprocessing plant. 

China is building a demonstration MOX fuel fabrication plant at the CNNC Gansu Nuclear Technology Industrial Park in Jinta, Gansu Province.^[13]  In 2018, the China Nuclear Power Engineering Co of the CNNC published a short post on its website, labeling this facility as the first MOX fuel production line in China, with an annual production capacity of 20 tonnes of fuel per year.^[14]  It explicitly states that the facility will provide fuel for fast reactors.  Hui Zhang at Harvard University’s Belfer Center used this publication to locate the construction site at the Gansu site.  It is unclear when this facility will operate, but a 20 tonne per year capacity should be adequate for both CFR-600 units.^[15] 

MOX fuel fabrication is part of China’s cooperation with Russia.  Rosatom publications give a date of 2034 for when their collaborative effort will produce pilot batches of MOX fuel for the fast reactors, with an eye toward large-scale production of licensed MOX fuel for the CFR-600 reactors. 

The reprocessing of fast reactor MOX fuel has proven difficult in practice around the world, and China is not thought to have started construction of such a reprocessing plant.  China has been building reprocessing plants to separate plutonium from fuel irradiated in PWRs at Jiuquan (Plant 404) and the CNNC Gansu Nuclear Technology Industrial Park in Jinta, Gansu Province.  This type of reprocessing is widely viewed as a way to produce enough separated plutonium, albeit reactor-grade plutonium rather than weapon-grade, for the initial cores of fast reactors or for recycling into PWRs.^[16] It also serves as a preliminary step on the path to fast reactor fuel reprocessing.^[17] 

Collaboration between the two states on the closed fuel cycle was further expanded in March 2023 when Rosatom and China’s Atomic Energy Authority signed a Comprehensive Programme for Long-Term Cooperation, which covers a wide range of areas “expanding cooperation in current projects, as well as implementing new projects related to fast neutron reactors such as production of uranium-plutonium fuel and handling of used nuclear fuel.”^[18]  The latter may be referring to fast reactor fuel reprocessing. 

Reprocessing blanket fuel to separate weapon-grade plutonium is much less challenging, since the fuel is uranium oxide containing almost exclusively uranium 238 and is only lightly irradiated, meaning it contains a low concentration of plutonium and is relatively easy to chemically process.^[19]  The irradiated fuel in the bulk of the blanket, e.g. the radial blanket uranium oxide fuel, can be processed at China’s already existing reprocessing plant(s), designed to separate “reactor-grade” plutonium from highly irradiated enriched uranium oxide fuel discharged from pressurized water reactors (PWRs).  

However, the radial blanket contains a significant amount of fuel that requires processing.  The CFR-600 is estimated to have over 30 tonnes of uranium oxide fuel in its radial blanket, and typically all blanket fuel would be discharged and processed each year.  It can be left in longer if more, but lesser quality, weapon-grade plutonium is desired. As a result, blanket fuel from one CFR-600 could be processed in separate campaigns at the 50 tonne per year Jiuquan reprocessing plant, or blanket fuel from both units could be processed in one of the reprocessing plants being built at Gansu.  A downside would be that the blanket fuel would lower throughputs of PWR spent fuel through the reprocessing plants, but this is a minor consideration if the high-quality plutonium is needed for nuclear weapons. Moreover, China has relatively modest requirements for separated, reactor-grade plutonium. 

There is much less information about Russia’s supply of reactor components, HEU, or technology to the second CFR-600 reactor.  A possibility is that China is building this reactor, or a successor one, using parts manufactured domestically. 

Because Russia’s fast reactor program is mature and has performed far better than Western or Japanese commercial fast reactors, the CFR-600’s successful operation is more likely, and Russia’s continued cooperation can aid in solving start-up and operational problems.  However, as was the case of PWRs originally imported from France and the United States, China can be expected to further indigenize the production of breeder reactors. 

Russian Requirements for Peaceful Use on the Demonstration CFR-600   China and Russia’s nuclear energy cooperation has raised concerns about the potential military applications of resulting fissile material, namely the blanket plutonium.  However, there are several restrictions in place, such as Russian export laws and bilateral agreements with China that explicitly prohibit the use of Russian supplied materials, technology, or byproduct materials, such as plutonium, for the production of nuclear weapons.  The 1996 Agreement on Cooperation in the Areas of Peaceful Use of Nuclear Energy, signed by the two nations, explicitly “prohibits the use of exported nuclear items for nuclear weapons or any military purposes.”^[20]  Additionally, Russia’s 2000 Regulations on the Export and Import of Nuclear Materials includes a specific provision banning the use of exported Russian nuclear materials, or their byproducts, such as plutonium, for military purposes. The 2000 Russian/Chinese agreement on cooperation on the CEFR also contains a clause that nuclear materials, technology, equipment, facilities, and byproduct material cannot be used by China for the production of nuclear weapons and other nuclear explosive devices. This language is repeated in the 2018 agreement signed between Russia and China on cooperation on the design, construction, commissioning, and operation of the demonstration CFR 600 MWe reactor (singular, not plural).  The relevant language in Article 8 states that:

Nuclear materials, equipment, special non-nuclear materials and relevant technologies obtained by the People’s Republic of China in accordance with this Agreement, as well as nuclear and special non-nuclear materials, facilities and equipment produced on their basis or as a result of their use will not be used for the production of nuclear weapons and other explosive devices or for any military purpose. 

It should be pointed out that the 2018 agreement concerns a single fast reactor, called the demonstration fast reactor, and does not reference the other unit. 

Comparing Russia’s Restrictions on China with the United States’             China’s nuclear agreements with Russia, particularly Russia’s restrictions on the use of its supplied nuclear material for military purposes, contain less restrictive language than those with the United States.  The U.S. has a long history of peaceful nuclear cooperation with China, beginning with the 1985 agreement between the two countries on nuclear cooperation.  When the 30-year agreement was set to expire, it was updated and renewed in 2015.^[21] 

Similar to Russian restrictions, the 1985 U.S.-China agreement and its successor were designed to ensure that nuclear equipment, materials, and technology exported to China were strictly for peaceful purposes.  It also explicitly stated that U.S. exports were “subject to U.S. laws and regulations in effect at the time of export.”^[22]  U.S. export controls further reinforced these restrictions, preventing the transfer of nuclear technology and materials that could potentially be used for military purposes. 

The updated 2015 agreement reflected changes since 1985, such as China’s accession to the Nuclear Non-Proliferation Treaty (NPT) in 1992 and ongoing concerns surrounding China’s commitment to non-proliferation, particularly regarding transfers to North Korea and Pakistan by Chinese entities.^[23]  To address these concerns, a Nuclear Proliferation Assessment was submitted to Congress alongside the 2015 agreement for review. 

The recent Russian agreements are facility or nuclear material specific, unlike the U.S./China agreements which cover all peaceful nuclear cooperation between the two countries.  This raises questions about constraints on successor plants.  Moreover, restraints on technology are hard to enforce. 

The 1985 and 2015 agreements between the United States and China were also far more detailed and restrictive than either Russia and China’s 1996 Agreement on Cooperation in the Areas of Peaceful Use of Nuclear Energy or their 2018 agreement on the construction of the demonstration CFR-600.  Similar to the China-Russia agreements, the 2015 U.S.-China agreement explicitly prohibited the use of transferred material and technology for military purposes.  However, the U.S. agreement went beyond this basic statement of restriction by introducing a supplementary Agreed Minute, which laid out additional safeguards to prevent military misuse of provided material. 

The Agreed Minute addressed several key concerns related to the potential use of supplied materials and technologies for military purposes.  It specifically discussed byproduct materials of nuclear energy and went beyond Russia’s agreements, which merely included clauses of peaceful use of byproducts.  The U.S. agreement required annual information exchanges on byproducts of nuclear material.^[24]  

Although many U.S. proliferation concerns centered on transfers to other states, such as North Korea and Pakistan, the agreement also contained special restrictions on the storage of any plutonium produced as a result of U.S.-supplied material or technology.  This condition would allow the United States to track any produced plutonium if it was separated from irradiated material:

Plutonium and uranium 233 (except as contained in irradiated fuel elements) and high enriched uranium transferred pursuant to this Agreement or used in or produced through the use of material or equipment so transferred shall only be stored in facilities to which the Parties agree [emphasis added]. 

In light of all the provisions included in the Agreed Minute, the 2015 U.S.-China agreement imposed far more rigid restrictions and oversight mechanisms than Russia’s agreements with China, which were comparatively vaguer, and more broadly stated that supplied materials could not be used for military purposes.  The U.S. restrictions would make it significantly more difficult for China to secretly divert nuclear materials for unauthorized uses and make detection of any violation quicker. 

 

The Military Use of the CFR-600 Reactors

The publicly available nuclear cooperation agreements on the surface appear to prevent China from using the CFR-600 for military purposes, but they may not in practice.  China’s lack of nuclear transparency and the less-than-watertight agreement with Russia on the demonstration CFR-600 raises legitimate questions about its use to make weapon-grade plutonium for nuclear weapons. The concern about applicable restrictions and Russian enforcement thereof is heightened when considering the CFR-600 unit 2 or successor units. 

There are long-standing concerns about China’s ambiguous statements and lack of transparency about the CFR-600.  As stated above, China has described the CFR-600 as a “national defense investment project,” directly contradicting the statements of Chinese officials and public claims that the reactor is not intended for any military use.[25]  The uncertainty surrounding the CFR-600’s potential military use is further heightened by China’s general lack of nuclear transparency.  For years, China has rejected calls at the Conference on Disarmament for a moratorium on the production of fissile material for use in nuclear weapons, despite supporting a Fissile Material Cutoff Treaty.  Another instance is China’s discontinuation of annual reporting of its civilian unirradiated plutonium stocks under the international arrangement contained in INFCIRC/549.^[26]  If the plutonium produced by the CFR-600 was genuinely intended for nuclear energy purposes, it would be expected that China would readily disclose this information, but it has not. 

Lastly, while the publicly available cooperation agreements with Russia do include restrictions against military use, and the restrictions on the demonstration CFR are clearer, they also are less verified than the non-proliferation restrictions imposed by the United States on civilian nuclear cooperation with China.  Moreover, any changes may be subject to secret negotiations, and China’s adherence depends on Russian oversight and enforcement.  Russia has extensive nuclear cooperation with China, far beyond the CFR-600 reactors.  Would it be willing to jeopardize all of its cooperation to confront China if China diverted blanket plutonium to nuclear weapons?  Would Russia even know it took place?  It is legitimate to worry that in the current climate the likely answer to both questions is no.  

 China’s Third Line Reactor: Another potential source of plutonium?

While the CFR-600 is an important source of plutonium, it may not be the only one. Another potential source of additional plutonium is an old production reactor located at the 821 Plant in Guangyuan, Sichuan Province.  The 821 reactor, aka “Third Line reactor,” began development in the 1960s, with the reactor becoming operational in 1974.  It is a graphite-moderated, water-cooled reactor.  It is located at 32°29'44.75"N, 105°35'25.04"E. Figure 2 shows a ground image of the reactor, and Figure 3 shows its placement at the Guangyuan site.  

The annual plutonium production at the 821 Plant reactor depends on the power of the reactor, which China has not revealed. Estimates of the 821 reactor power have varied widely, from 250 MWth to over 1200 MWth.^[27]  More recent research suggests that both the 821 reactor and China’s earlier reactor at the 404 Plant (also known as 801 reactor, here called 404 reactor) had the same initial nominal power of 600 MWth, and both reactors’ power was subsequently increased, with the 821 reactor’s power slightly higher than the 404 reactor’s power.^[28]  However, China has never provided official information about the power of these reactors. Instead, available information from open-source means relies on Chinese blog sites or other types of reminiscences of officials or former workers at the two sites.  Their comments are often shrouded and incomplete, requiring interpretation.^[29] 

Institute assessments of the reactor’s peak power have varied between 500 and 1000 MWth, with initial power as low as 250 MWth, reflecting the uncertainty of the available sources.  The current Institute assessment puts the peak power between 600 and 750 MWth, based on the new information discussed above.  Using a simple weapon-grade plutonium production formula gives an annual value of 200 to 250 kilograms per year at 100 percent capacity factor, where capacity factor means full-power operation at peak power for that fraction of the year.^[30]  Capacity factors are hard to estimate, but, assuming a nominal rate of 80 percent, it would yield a yearly production of 160 to 200 kilograms.  However, despite claims to the contrary, such a high capacity factor is often not obtained in practice.  A capacity factor of 60 percent would lead to 120 to 150 kilograms of weapon-grade plutonium per year.  Although the lower capacity factor may better represent long-term operation, the larger rate is used in this report to represent a nominal plutonium production value. 

Figure 2. An undated ground image sourced from https://kknews.cc/zh-my/history/mn92al6.html. While undated, the structures present indicate it was taken in the early 2000s or earlier. 

Figure 3. A March 2023 image of the Guangyuan Plant complex, highlighting the reactor building, the reprocessing building, and a solid waste storage area that was constructed in the early 2010s and has expanded since. 

The 821 reactor at Guangyuan is China’s second plutonium production reactor.  The first reactor is called the 404 Plant at Jiuquan, located at 40.22317° N, 97.35605° E. 

The Jiuquan reactor has reportedly undergone at least the first stages of dismantlement,^[31] meaning it is no longer available to produce weapon-grade plutonium.  In contrast, most agree that China’s production of weapon-grade plutonium stopped by about 1990 or earlier,^[32] however, there are no official reports that the reactor was being dismantled or prepared for decommissioning.  There are ambiguous statements in blog posts or personal histories about the 821 Plant that independent researchers have interpreted as the reactor having closed permanently.  However, reviewing these sources did not reveal any statements that the reactor was shut down in an irreversible manner.  

Available open-source records mostly discuss the reactor switching to civilian products, providing no details about the status of the reactor.  Some cite a decision in August 1987 by China’s State Council that the 821 factory would "stop production of military products and switch to civilian products."^[33] 

Complicating this issue, China continues to treat weapon-grade plutonium production as classified.  In light of the 50th anniversary of the 821 Plant, celebrated in 2019, China’s Atomic Energy Agency released a news article, unveiling the 821 Plant as “a secret place in the nuclear industry that spans half a century.”  The clearest statement provided on the reactor follows a discussion of what happened in the 1980s, where it explains that both weapon-grade plutonium and a conversion to electricity production were suspended: “The construction of the reactor waste heat power generation project has been suspended along with the cessation of military product production.”^[34]  But that was in the 1980s.  Noticeably, there were no discussions on the current status of the reactor, to the point that its omission is noteworthy.  

Western officials have stated that the 821 reactor could have continued operation after conversion from weapon-grade plutonium production to civilian production, or was maintained to allow restart operations later.^[35]  Institute staff were told in the 2000s by knowledgeable U.S. officials that the reactor was kept on “a standby status” after ending its production of weapon-grade plutonium for military purposes. 

But did it remain ready for future operation, or was it being maintained for later decommissioning, which would eventually likely entail removing the irradiated graphite moderator?  Entombing the graphite on-site does not appear likely for an above-ground reactor. 

China’s ongoing secrecy appears excessive if the reactor is being decommissioned.  Changes visible at the site over the last decade, but as recent as 2024, show that certain other facilities at the 821 Plant were razed and nuclear waste was being processed.  While some of the activity at the reactor site observed in optical and thermal imagery over the years could support the management of the core, including the irradiated graphite for future dismantlement, it could also be related to reactor operation.  In fact, there are multiple indications that the reactor may have been operating during the last decade, or that it was undergoing preparations for upcoming reactor operation and generation of electricity, despite the earlier decision in the 1980s not to do so. 

 

Read the full report in PDF. 

 

 

*William Goodhind is a researcher with Contested Ground.