Chinese Submarine Warfare – A Natural Evolution or Game Changing Revolution?

Recent media reports have highlighted the increasing sophistication and numbers of submarines being produced by the People’s Republic of China, describing a submarine arms race. Although the US retains a technological edge, the People’s Liberation Army Navy (PLAN) is apparently closing the distance. The PLAN is currently on track to field both a new SSN (the Type 95) and a new SSBN (the Type 96) and what is known about the designs of both would seem to indicate improvements in Chinese nuclear submarine design. Moreover, China’s submarine detection capabilities appear to be developing in ways that suggest that they may pose a robust threat to American SSNs operating in the first island chain. Should qualitative improvements be combined with a formidable manufacturing capacity, China will likely start to challenge US military maritime dominance under the waves.

Furthermore, the recently retired former Chief of Defence Staff, Admiral Radakin, raised the ‘sobering point’ that the lack of a formal arms control agreement between the US and China is a gap in the nuclear arms control regime. Although this does not imply imminent risk of conflict, there are concerns that Chinese SSBNs and Anti-Submarine Warfare (ASW) capability could, in tandem with the broader growth of the Chinese nuclear deterrent, become leverage with the US as China continues to pursue reunification with Taiwan.

Across domains, Chinese weaponry is becoming more sophisticated which, at the very least, is closing the gap on Western systems. China’s defence policy is moving from a ‘model of quantity and scale to a model of quality and efficiency’. However unverified claims of technological advancement are regularly championed in PRC sympathetic publications. It is therefore a legitimate question to ask just how concerning these undersea advancements are for Western military planners and strategists.

The Gradual Improvement of Chinese SSNs

To date, China has seen more success in developing diesel electric submarines than nuclear submarines. The PLAN has approximately 48 Diesel Electric attack submarines (SSKs) of which nearly half boast an ‘Air-independent Propulsion’ system that improves survivability by extending the periods for which the boats can remain submerged. Notably, in fielding the AIP equipped Type 039 Yuan Class (which uses a Stirling Engine) China has stolen a march on Russia, which does not have an AIP equipped conventional submarine. It is, however, of note that the PLAN does not yet field the fuel cells found on the most modern western conventional submarines. Even so, these submarines are well suited to contesting areas at or just beyond the first island chain, where their relative lack of endurance is less of a handicap and they enjoy an advantage in quietness over SSNs.

China’s fleet of nuclear submarines, by contrast, has typically been an area of relative weakness. For example, the Type 093A submarine is comparable to a Russian Victor III class submarine designed in the 1970s, a point illustrated by the ease with which the JMSDF was able to detect its narrowband tonals in 2018 when the submarine loitered at (likely) low speeds near the Senkaku islands.

It might be inferred that many of the weaknesses of the Type 093 are a reflection of its reliance on an inefficient powerplant based on two 75MW reactors, which constrained Chinese designers in terms of the size of the vessel they could build while still achieving a speed of 30 knots (necessary to track targets like aircraft carriers). By virtue of its size, it has been inferred that the 093A SSN lacks the horizontal floating isolation raft found on most western and Russian submarines instead relying on less effective compound isolation rafts. Additionally, limited additional space precluded the placement of its towed array (which the submarine was not initially designed with) between the pressure hull and outer hull, requiring the addition of an outer ‘hump’ which likely generates hydrodynamic noise and increases the risk of detection due to the formation of horseshoe vortices which contribute to cavitation.

China has gradually made iterative improvements to the Type 093 design over the last two decades. In the early 2000’s China took receipt of Russian Kilo class submarines from which, per some analysts, they reverse engineered pneumatic mounts for newer versions of the Type 093A. From the 2010s onwards discussions of work on airbag based isolation emerged in both public fora and academic journals and it appears likely that an airbag isolation raft was incorporated into the Type 093B, reducing its narrowband signature relative to the Type 093A. Other improvements included a redesigned sail to improve hydrodynamic drag and the incorporation of pump jet propulsion.

Towards the end of the last decade, a slide show leaked by a Chinese Rear Admiral provided insights into the next generation Chinese SSN, the Type 095. Per the Admiral’s slides the Type 095 will be roughly 1300 tonnes heavier than the Type 093, indicating a vessel closer to the Virginia and Yasen classes in size. This is likely a function of China taking receipt of the KLT-40S, a Russian reactor which closely resembles the 150MW OK-650 reactor used on the Akula Class. The ability to build smaller more powerful reactors based on Russian designs opened up a vista of opportunities for Chinese designers who could now design larger boats without compromising on qualities such as speed. Most importantly, a larger boat can carry more effective sound isolation mechanisms. Illustrating this, the slides revealed that unlike its predecessors the vessel would be single hulled – indicating a degree of confidence in its sound isolation systems and suggesting that the larger Type 095 can fit a horizontal isolation raft. The size of the Type 095 also enables the incorporation of a VLS and (presumably) the ability to host sensors such as a towed array without the design compromises made for the Type 093. China is also making strides in propulsion. In a recent interview, Rear Admiral Ma Weiming discussed the maturation of a turboelectric propulsion system which would be incorporated on the Type 095. Turboelectric propulsion allows for the elimination of the reduction gears that a drive shaft would connect to, further contributing to quieting.

Despite these notable improvements, it is worth noting that even this design falls short of what might be considered state of the art. The Russians, for example, have replaced the OK-650 with the KTP-6 monoblock reactor which eliminates separate steam generators, resulting in a reactor that is both more compact and quiet. As such, China is still a generation behind the state of the art in areas such as powerplant design. Type 95 design as envisioned would nonetheless represent a submarine that is much more competitive than its predecessors and comparable in quality to the improved Akula.

China’s Progress in Fielding Modern SSBNs

As alluded to by Admiral Radakin, it is perhaps the nuclear question that requires particular consideration. China has been building and fielding SSBN since the 1980s, in the form of the relatively noisy Type 92 Xia class. These were gradually replaced by the quieter Jin class Type 94, which are still known to be noisier than both the UK and US equivalents. The latest two additions to the fleet have benefitted from further improvements in reducing sound emission, leading some commentators to differentiate these as a Type 94A. Although the Type 94 has 12 vertical Ship Launched Ballistic Missile (SLBM) tubes that could carry the JL-2 missile (NATO designation CSS-N-14) potentially with a 700 kiloton nuclear warhead, there are a number of reason why new developments are being viewed as indicative of a step change in capability.

Chinese SSBN activity hitherto has been expected to operate within these Chinese dominated waters. The JL-2 missile range of approximately 7000km posed limited direct threat to US territory when constrained to the South China sea. An upgrade to the 10000+ km range JL-3 missile, reportedly carrying multiple warheads, would extend that threat to include Hawaii and Alaska, moving further towards the Bohai Sea (between China and Korea) would bring much of the western seaboard of the US into range without the submarine having to leave its bastions.

Naturally the new Type 96 Tang Class specifications are likely to remain secret. However, the boat is expected to displace 20,000 tonnes and is to boast a 24 vertical launch (JL-3) missile capability. As well as improved acoustic stealth due to the incorporation of features such as the shaftless pump mentioned by Admiral Ma, China continues to champion advances in sonar technology which are also likely reflected in the Type 096. The incorporation of these latest technologies combined with the Tang’s greater size and missile capacity means it will likely be able to sustain longer range and longer duration deployments. The PLAN is also growing in confidence and experience in projecting further afield, beyond the South East Asia sea. In this context, longer range submarine patrols could well become more likely, particularly if the PLAN can control chokepoints such as the Bashi channel early in a conflict. Interestingly, there is some indication, based on academic research conducted within the Chinese shipbuilding sector, that the Type 096 will have an ice-rated hull and some academics have indicated that China has expressed an interest in access to Russian Arctic bases. Given that a northern ballistic trajectory to the US is efficient and would not require transiting chokepoints such as the Bashi channel, it might also be the case that the Type 096 will be deployed in an area where the Euro-Atlantic and Indo-Pacific overlap. Despite improvements, however, the Type 096 retains certain unusual features such as the hump characteristic of previous SSBNs (which likely houses it’s VLS) which would impede certain aspects of the submarine’s performance.

Chinese declaratory nuclear policy is very clear, in that it maintains a no-first strike stance and also claims to maintain stockpiles only at ‘the minimum level required for national security’. However, it continues to grow its nuclear arsenal and its delivery capability. Different inferences can be drawn regarding whether this represents a hedge against the US’ perceived counterforce advantages over China or a shift in real (if not declaratory) policy that indicates a greater reliance on nuclear blackmail. In either instance, however the ability and willingness of the PLAN to project and sustain SSBNs into the wider Pacific or the Arctic should be of considerable concern. Just as Russia is able to sustain Borei SSBN patrols in the Pacific, threatening to split US ASW efforts between two theatres, a similar scenario could emerge on the US west coast. SSBNs close to the US mainland obviously significantly reduce the time available for missile intercept should any launch occur. The ability to create such dilemmas for the US will become increasingly significant. The implicit threat posed by one or more Tang submarines in the Pacific may well prove useful Chinese leverage in encouraging the US not to become involved in Taiwan reunification issues.

Improving Defensive Capabilities in the First Island Chain

Added to this, recent articles coming from China (and presumably regime authorised) claim that advances in AI, will soon hugely reduce the survivability of enemy submarines operating against the PLAN. The report claims computer modelling shows that AI will facilitate the detection and tracking of even the most quiet of submarines even in the face of active decoy and evasion techniques. This will reduce enemy submarine survivability ‘to 5 percent’. However, the simulation relied on the presence of multiple, layered, active and passive and interlinked sensors and would also, presumably, require significant data on target vessel signatures. The simulations should, at this stage, be seen as a validation of AI supported layered ASW sensor-based defence rather than a declaration of existent capability. The simulations acknowledge that considerable hardware would need to be deployed in target areas, which implies its efficacy primarily in areas already under PLAN control and/or in constrained areas, particularly chokepoints.

Furthermore, it cannot be inferred from the reports that the necessary sensors, processing nodes and undersea communication networks have been developed and tested to the degree necessary to make such detection probabilities a reality yet, although the development of an ‘undersea great wall’ comprised of layers of overlapping sensors in areas such as the South China Sea has been a long term (and whole of state) effort. Of note, the principle modelled is not dissimilar to that of the UK’s Atlantic Bastion, which is a priority within the Strategic Defence Review.

Despite the above caveats, the messaging of such articles appears to be clear: the PLAN’s ability to defend against submarines in its own back yard will eventually be insurmountable. Of course, this does not prevent submarine launched strikes from waters not so intensively monitored, or sensor saturated. However, in principle, it would extend the Chinese subsea Anti-Access/Area Denial (A2AD) range and further restrict their adversaries’ planning freedoms. Taken together, China’s sensor networks and fleet of quiet diesel electric submarines could pose a real challenge to the submarines of the US and other allies that would look to help defend Taiwan in case militarily supported reunification attempt by the PRC, should these submarines have to operate in the first island chain.

The focus on overlapping layers of sensors close to the Chinese mainland in the east and south China Sea could also be interpreted as evidence of the importance that the PLAN attaches to keeping American SSNs equipped with long range missiles at arm’s length. Beyond the conventional threat they pose, American Virginia class SSNs equipped with the long-range hypersonic weapon which will be fitted in the Virginia payload modules can, if employed in tandem with the US nuclear arsenal, pose a critical risk China’s capacity for a second-strike capability. To an extent, then, development of an undersea great wall and the growth of China’s nuclear arsenal could be seen as two halves of the same coin.

Conclusion - Close but no Cigar

To return to the opening question of how concerning these advancements should be, it is clear that, militarily they are of considerable concern. Although media reporting may be overstating current PLAN capability, the direction of travel in both numbers and sophistication is clear.

This being said, there are real limitations to consider. Even newer Chinese SSNs may well fall short of the performance standards achieved by their quieter Russian and Western counterparts, something likely to be true of SSBNs as well. Furthermore, China’s constrained geography, which means that submarines can only leave the first island chain through chokepoints like the Bashi Channel (other routes of egress being too shallow) further limits Chinese submarine operations.

Within the first island chain, the sheer number of Chinese assets and sensors available coupled with the capacity to process and disseminate data could pose a significant risk to US SSN operations. However, this level of sensor saturation will be difficult to achieve beyond well-defined areas.

It should be noted however that even modest improvements are noteworthy given the qualitative advantage that the US needs to offset China’s industrial heft. Moreover, given China’s penchant for building multiple variants of any class of vessel, the emerging generation of Chinese SSNs is likely to represent the final stage of the PLAN’s trajectory.

© RUSI, 2025.

The views expressed in this Commentary are the authors', and do not represent those of RUSI or any other institution.

For terms of use, see Website Terms and Conditions of Use.

Have an idea for a Commentary you'd like to write for us? Send a short pitch to commentaries@rusi.org and we'll get back to you if it fits into our research interests. View full guidelines for contributors.