Introduction

The current growth in China’s nuclear arsenal is deliberate, reflecting a carefully assessed strategy to attain parity, or near parity, with the United States (US) and Russia. The consequences of this expansion will cascade across the region, increasing India’s vulnerability and compelling New Delhi to contend with a dual nuclear challenge from China and Pakistan. China’s atomic growth is not merely quantitative but also qualitative, marked by visible shifts in its nuclear posture and, critically, its delivery capabilities.

China has increased its nuclear arsenal year on year. Latest data from the Stockholm International Peace Research Institute (SIPRI) shows that between 2024 and 2025, China’s stockpile increased by at least 100 warheads—from 500 to 600—representing a 20-percent rise.^[1] This rate of expansion is greater than any nuclear weapons state, whether de jure or de facto.^[2] At least partially, this is unsurprising. Long before its current build-up, China had strongly opposed a tripartite nuclear arms control agreement with the US and Russia aimed at reducing nuclear stockpiles. As People’s Liberation Army (PLA) Senior Colonel Bo Zhou observed in 2019, “For such an agreement to work, either the US and Russia would need to bring their nuclear arsenals down to China’s level, or China would need to increase its arsenal drastically. Neither scenario is realistic.”^[3]

Yet the current pace of China’s atomic arsenal distention clearly indicates that this assessment is only partially accurate. While China is unlikely to match US or Russian stockpile levels, Sino-Russian nuclear cooperation may be intended to narrow the gap with Washington and provide Moscow and Beijing greater leverage in future arms control negotiations, including the possibility of a tripartite agreement and a Fissile Material Cut-off Treaty (FMCT), even if the latter remains remote over the next decade.

India faces critical choices in response to China’s growing atomic strength. New Delhi confronts some trade-offs that can be framed as two tests: a maximal and a minimal response. At the maximum end, as argued below, India would need to conduct more nuclear tests to meet key qualitative thresholds—improving yield-to-weight ratios, performance, and reliability—while also expanding the production rate of its atomic arsenal. However, if these qualitative and quantitative requirements cannot be met, a minimal response would still require substantial investments in ground-, air-, sea-, and space-based sensors, alongside a defensive missile shield capable of countering both ballistic and hypersonic missiles already in the inventory of the People’s Liberation Army Rocket Force (PLARF). A nuclear-tipped hypersonic missile threat from China would necessitate not only the development of comparable offensive capabilities, as India is currently pursuing, but also a credible hypersonic interception capability integrated with a robust sensor network. This effort would need to be supplemented by the accelerated expansion of India’s subsurface nuclear fleet, particularly nuclear-powered attack submarines or Submersible Ship Nuclear (SSNs), which would extend the operational reach of the Indian Navy’s (IN) Ship Submersible Ballistic Nuclear (SSBN). In parallel, India’s SSBNs would require longer-range ballistic missiles, some of which have already undergone partial testing.  These measures constitute the minimum requirements for India.

The first part of the analysis evaluates the trajectory of China’s atomic capabilities and its efforts to pair nuclear strength with advanced delivery capabilities. The second part evaluates the development of Beijing’s space-borne sensor, missile delivery capabilities, and Missile Defence (MD) architecture geared for Damage Limitation (DL). The third and fourth parts evaluate, respectively, India’s maximalist and minimalist responses to Beijing’s growing strength in nuclear weapons and delivery capabilities.

China’s Nuclear Expansion: A Comparison with India

These developments must be viewed against the backdrop of 2023-24, when China finished the development and operationalisation of its new Fast Breeder Reactors (FBRs), the CFR-600, which receive Russian fuel support. This assistance will substantially augment the country’s weapons-grade plutonium inventory.^[4] Russia started shipping out the initial fuel load in late December 2022.^[5] FBRs are crucial because they produce more plutonium and fissile material than they consume during power generation.^[6] As fast-neutron reactors, they are particularly efficient in generating fissionable material.^[7] Super-grade plutonium, or plutonium-239, which is highly suited for atomic weapons, can be extracted through the reprocessing of spent FBR fuel.^[8] China has reprocessing plants at its Gansu facility that are capable of reprocessing spent fuel. Two plants are already in operation at the site, and a third is under construction. These can be potentially used for reprocessing spent fuel from its FBRs.^[9]

Reinforcing these concerns is Beijing’s decision, since 2017, to cease reporting its tritium, uranium, and separated plutonium (Pu) stocks.^[10] Unlike other designated nuclear-weapons states (NWS), China has not reported its plutonium holdings under the International Atomic Energy Agency (IAEA) Guidelines for the Management of Plutonium (INFCIRC-549). Its holdings are estimated at roughly 2.9 metric tonnes (MT) of military-grade plutonium, as shown in Table 1.^[11] China’s warhead stockpile surged by over 71 percent from 350 warheads in 2022 to an estimated 600 in 2025.

Table 1: China’s Plutonium and Warhead Stockpile

Start of Year	Plutonium (Pu) Stocks (Metric Tonnes)	Number of Nuclear Warheads
2022	2.9	350
2023	2.9	410
2024	2.9±0.6	500
2025	NA – Data unclear	600

Sources: International Panel on Fissile Materials (IPFM), SIPRI Yearbook-2025 and Sasakawa Peace Foundation (SPF).^[12]

With existing plutonium stocks of 2.9 tonnes, China could potentially accumulate some 965 nuclear warheads by 2030.^[13] This figure is consistent with the latest Pentagon estimates that put the figure at 1,000 warheads.^[14] Anything above 1,000 warheads will require diversion of plutonium from China’s two under construction FBRs in Fujian and the two Heavy Water Reactors (HWRs) in operation since 2002 at its Qinshan Nuclear Power Plant.^[15] Both FBRs and HWRs are most suited for producing weapons-grade plutonium, as shown in Table 2. Any further increase in China’s warhead stockpile using plutonium beyond that produced by these facilities would constitute a serious violation of Beijing’s pledge under INFCIRC-549 and would be alarming, particularly if it involved the use of Russian fuel to produce super-grade plutonium-239. Fundamentally, these factors explain the distention in China’s nuclear weapons stockpile.

Table 2: China’s Weapons Grade Plutonium Sources

Type of Plutonium	Pu 238	Pu-239	Pu-240	Pu-241	Pu-242
Weapon Grade (Pu produced from FBRs and HWRs)	0	93-97	3-7	<0.5	0
Reactor Grade Plutonium sourced from Light Water Reactors (LWRs)	2.4	53.8	22.0	15.5	6.3

Source: Yuki Kobayashi, Sasakawa Peace Foundation^[16]

More consequentially, China may be compelled for other reasons to boost its plutonium stocks through diversion from its FBRs and HWRs (Table 2) to give it the strength for strategic warheads that match the upper limits of the New START Treaty concluded between the US and Russia, which stands at 1,550.^[17] Reaching such levels would leave Beijing with little alternative but to divert plutonium from these facilities (see Table 2). This may become even more pressing if China were to emulate the US’s practice of rotational deployment, which requires a large reserve of non-operational warheads.^[18] Rotational deployment is pursued because maintaining full-scale deployment is very difficult: warheads require periodic maintenance, including tritium replenishment roughly every four years, to preserve their explosive strength.

Table 3: Nuclear Warhead of Select Nuclear States (2025)

Country	Deployed Warheads	Stored Warheads	Military Stockpile	Retired Warheads	Total Inventory
United States	1 770	1 930	3 700	1 477	5 177
Russia	1 718	2 591	4 309	1 150	5 459
China	24	576	600		600
India	180	180	180		180
Pakistan	170	170			170

Source: SIPRI Yearbook, 2025^[19]

As Table 3 indicates, both the US and Russia maintain more than 3,000 non-operational warheads.^[20] Although China is a designated NWS and is therefore not required to place nuclear material under IAEA safeguards, it pledged in a 1988 agreement with the IAEA to use imported nuclear material only for “peaceful purposes.”^[21] The fuel for China’s FBRs is sourced from Russia, while its HWRs are of Canadian origin.^[22] However, it is important to note here that China has 14+/-3 tonnes of weapons-grade Highly Enriched Uranium (HEU).^[23] China’s Lanzhou gaseous diffusion plant (Plant 504) and the Heping gaseous diffusion plant produce HEU for weapons.^[24] This is above the roughly 3 metric tonnes of military-grade plutonium stocks in its inventory (see Table 1). To be sure, the Lanzhou and Heping enrichment facilities are also used for the production of Low Enriched Uranium (LEU) for China’s naval reactors, and the HEU is used for research reactors.^[25]

In addition to these developments, China has undertaken significant construction activity at its Lop Nur nuclear test site.^[26] This is a portent of future intent, as China could pursue a renewed round of testing to improve the reliability and performance of its nuclear warhead designs. Lop Nur has witnessed extensive test-related preparations, which Beijing has sought to portray as benign industrial activity. However, geospatial intelligence derived from Earth Observation (EO) and Synthetic Aperture Radar (SAR) satellites indicates activity consistent with preparations for testing new warhead designs.^[27] Beyond the prospect of China resuming larger-yield atomic tests, there is also evidence of subcritical testing. Such tests use small quantities of fissile material, and China is assessed to have conducted them. Indeed, in 2020, evidence revealed that Beijing had carried out small-yield atomic explosions at Lop Nur’s Northern Tunnel Test Area (NTTA), which has since been decontaminated.^[28] These tests can be conducted in confined explosive chambers, making them difficult to detect.^[29]  Consequently, in 2020, the NTTA is believed to have been used for low-yield testing.^[30]

As a result of this extensive subcritical testing, Beijing has acquired the capacity to miniaturise warheads, thereby strengthening its multiple independently targetable re-entry vehicle (MIRV) capabilities.^[31] All these advances point not only to a vast increase in the size and quality of China’s nuclear forces but also, as discussed below, to rapid progress in its delivery systems. Beijing’s subcritical nuclear tests have certainly given it the capacity to miniaturise warheads that are necessary not just for integration with advanced missile delivery systems but equally indispensable to support the requirements for a larger arsenal, reliability, and performance.^[32]

Hypersonic missiles equipped with low-weight, low-yield nuclear warheads—enabled by advanced computational modelling and simulation—are therefore likely to emerge as a major challenge in the coming years.^[33] However, subcritical tests that use very small amounts of plutonium that do not reach criticality are essential for computational modelling.^[34] France, for example, has already started developing a Hypersonic Cruise Missile (HCM) capability as part of its current nuclear modernisation.^[35] The US, Russia, and China remain the principal leaders in hypersonic missile systems.^[36] China has already tested low-yield devices, which could potentially be integrated into the hypersonic missile segment of its nuclear arsenal.

At the declaratory level, China pursues a No First Use (NFU) policy, though its posture is shifting. While China has repeatedly maintained that it keeps its warheads and delivery capabilities de-mated, integrating them only during a crisis, growing evidence suggests this may no longer be the case. Some analyses indicate that it could relax its NFU commitment during a crisis or war, especially if its nuclear forces were struck by conventional weapons.^[37] More troubling for India is China’s possession of the MQ-19 and the more advanced MQ-26 Ballistic Missile Defence (BMD) system, designed to intercept intermediate-range ballistic missiles (IRBMs) such as the Agni-V. If China reverses its NFU and strikes India first, whatever surviving retaliatory capabilities India uses will be intercepted, limiting the damage for the Chinese mainland.^[38]

At present, China is moving away from a posture that separates its warheads from its delivery systems—a shift that is either under way or already complete. Further, Chinese analysts argue that a move towards a launch-on-warning (LoW) posture, evident since 2023, remains compatible with the declared NFU policy.^[39] Evidently in the last few years, China has launched several early-warning satellites, which are indispensable for detecting and tracking missile launches and for preparing retaliatory forces, missile defence, evasive action, and ‘early warning’ for counter-attack.^[40] This should seriously concern India’s strategic managers, compelling New Delhi to do the same to counter Beijing.

China’s Delivery Systems and Space-Based Sensor Capabilities

Although the People’s Liberation Army Rocket Force (PLARF) is equipped with a wide variety of missile capabilities, some of its most advanced are hypersonic projectiles crucial for both conventional and nuclear missions. These delivery systems include the deployment of the Dong Feng-17 (DF-17), an intermediate-range, road-mobile ballistic missile armed with a Hypersonic Glide Vehicle (HGV).^[41] HGVs are designed to evade radar detection and missile defences owing to their extreme velocities. Although Chinese sources claim that the DF-17 and its HGV are geared primarily for conventional strikes, the system is fully capable of executing nuclear strikes, with more advanced variants offering greater penetrability.^[42] During its ascent phase, the DF-17 follows a predictable flight path but becomes unpredictable during the glide phase.^[43] Another hypersonic missile, the YJ-19 cruise missile, uses a scramjet engine to sustain high speeds and mobility.^[44] These dual-capable (conventional and nuclear) missiles were among the hypersonic weapons displayed by China during its annual parade in early September 2025, alongside the YJ-21 and the DF-26D.^[45] The YJ-21 has also undergone upgrades that enhance its penetration capabilities. More recently, China successfully tested the DF-27A HGV, which reportedly has a range of approximately 8,000-9,000 km and achieved speeds of Mach 8.6 over a distance of 2,100 km.^[46]

A further capability is the DF-100, or CJ-100, supersonic cruise missile, which reportedly travels at a speed of Mach 4 with a range of 4,000 km, geared for land attack missions.^[47] It can be launched from a road-mobile Transporter Erector Launcher (TEL) as well as from an H-6N bomber, effectively extending the range of the CJ-100 from 4,000 to 6,000 km.^[48] This allows the H-6N airborne platform to stay well within Chinese territory and out of the striking range of the Indian Air Force’s (IAF) air combat fighters.

In addition, China has deployed the Dong Feng – 26 (DF-26) IRBM in Inner Mongolia, allowing its crews to train for battlefield missions that permit them to “swap between nuclear and conventional warheads.”^[49] Second, China has also unveiled a dual-use air-launched ballistic missile (ALBM) delivered from its H6-KN bomber jets in 2025. Finally, China’s Submarine-Launched Ballistic Missile (SLBM), the JL-3, which is capable of hitting targets at intercontinental ranges, is deployed aboard Type 094 SSBNs and is also intended for deployment on the next-generation Type 096 submarines.^[50] Taken together, these delivery systems provide China with a diverse set of options to conduct nuclear strikes against India.

Beyond these capabilities, China operates a spacecraft constellation known as the China High-resolution Earth Observation System (CHEOS), as shown in Table 2. These are a combination of satellites with IR/EO and radar sensor payloads that India will need to develop and launch by 2029 as part of its future multi-orbit Space-based Surveillance System-3 (SBS-3).

Table 3: China’s High-definition Earth Observation System (CHEOS)

Satellite Type and Numbers	Functions and Applications
5 Gaofen (GF) - 11	5 GF-11 high-resolution remote sensing optical satellites. Details on satellite series 1-7 primarily civilian (potentially dual use), but numerical values from 8 and above have military applications. 1.5 diameter plus apertures providing optical imagery of up to 10 centimeters (cm). Multispectral, hyperspectral and SAR features. Believed to be military satellites under the name JianBing 16, despite being labelled civilian.
GF5-02, GF5-01A, ZiYuan (ZY) and 1-02D ZY1-02E	Payload includes Hyperspectral imager dubbed the Advanced Hyperspectral Imager (ASHI) for “wide swathe observations”, ZY1-02E has an extra thermal infrared camera.

Source: Author’s compilation from multiple sources^[51]

Maximum Requirements for India Against China’s Nuclear Expansion

For India, the challenge is more complex: it cannot match China pound for pound, but it will need to cushion itself with a nuclear arsenal that runs into the high hundreds in the long run. The current Indian rate of production is at a modest ten weapons per year, which is insufficient. But this glacial accumulation is driven not only by production constraints but also by the absence of robust delivery systems, some of which remain under development while others are yet to materialise—an issue the author addresses in greater detail when discussing India’s minimum requirements.^[52] Reverting to the capabilities required under the maximum test, New Delhi will need an entire panoply of capabilities listed from 1 to 6. These cover the following:

1) Conducting a series of fresh atomic tests to validate thermonuclear capability and new designs. If tests reaching criticality are not feasible, New Delhi would need to seriously consider a series of subcritical tests, as China does at the Lop Nur facility. Subcritical tests use small amounts of plutonium but do not produce a self-sustaining fission reaction, unlike a nuclear detonation. They are vital for ensuring the safety and reliability of a nuclear weapon.^[53] Russia conducts subcritical tests at Novaya Zemlya, the United States at the Nevada Test Site,^[54] and China at Lop Nur; however, there is no evidence of India conducting subcritical tests. Although India is not a signatory to the Comprehensive Nuclear-Test-Ban Treaty (CTBT), subcritical tests are permitted under the treaty.^[55]

2) An arsenal exceeding 800 nuclear weapons to match China;

3) Reversing or relaxing India’s NFU Policy enshrined in nuclear doctrine;

4) Comprehensive testing and deployment of offensive missile capabilities with extended ranges and multiple launch platforms;

5) Missile defence that includes ballistic and hypersonic interceptors supported by an extensive sensor architecture;

6) Comprehensive development and deployment of nuclear subsurface platforms.

The first requirement—resuming atomic testing—would come at a price. International sanctions are likely if India were to conduct nuclear tests, potentially ending international cooperation in its civilian nuclear power programme, which is vital to meeting India’s energy needs. Nuclear energy has assumed greater importance following the Modi government’s decision to open the Indian civilian nuclear sector to private investment and participation.^[56] Consequently, liberalisation of the nuclear energy sector constrains India’s ability to resume atomic testing. The sector remains heavily dependent on foreign cooperation, particularly for fuel supplies and access to new nuclear technologies, which are central to meeting the ambitious energy targets of current and future governments.

However, forgoing testing to build-up India’s nuclear energy programme risks undermining the quality, performance, and reliability of its nuclear arsenal. It would also entail surrendering—or at least diluting—the ability to validate new weapon designs, including a thermonuclear design, which reportedly failed during testing in 1998.^[57] Even if India forsakes “hot” testing, it cannot overlook sub-critical testing, which allows India to stay within the bounds of its nuclear test moratorium and will not breach the CTBT.^[58]

The second requirement under the maximalist test concerns the quantitative balance between India and China. While numerical strength may not currently preoccupy India’s leadership, it is likely to be a more pressing challenge in the years ahead. A brief comparison of the respective nuclear arsenals of India and China illustrates this growing imbalance.

Table 4: Plutonium Stocks of Select Nuclear Armed States (2024-2025)

Country	Total Plutonium Stock in Metric Tonnes (MT)	Military Grade Plutonium MT
United States	87.6	38.4
Russia	193	88+/-8
China	3	2.9+/-0.6
Pakistan	0.58	0.58+/-0.2
India	11	0.7+/-0.16

Source: International Panel on Fissile Materials^[59]

Although India’s overall Pu stocks are larger than China’s, its holdings of military-grade plutonium are smaller, as shown in Table 4. There is also no indication that India is accumulating nuclear warheads at the pace China is pursuing. In contrast to India, China’s warhead numbers increased by an average of nearly 100 warheads annually over the same period (Table 1). If anything, India’s warhead strength is at roughly 10 per year over the same time span (Table 5).

Table 5: India’s Total Plutonium and Warhead Stockpile

Start of Year	Plutonium (Pu) Stocks (Metric Tonnes)	Number of Nuclear Warheads
2022	11	160
2023	11	172
2024	11	180
2025	NA

Source: International Panel for Fissile Materials, and SIPRI.^[60]

Further, China’s weapons grade HEU programme stands in sharp contrast with India. The latter does not possess a weapons-grade HEU programme, and its production of HEU is primarily geared for its nuclear submarine reactor programme.^[61] India has a pilot enrichment plant at the Bhabha Atomic Research Center (BARC), and a larger enrichment plant has been in operation under the Rare Materials Project at Rattehalli in Karnataka.^[62]

Table 6: Rate of Increase of India’s Nuclear Warheads (2026-2036)

Start of Year	Rate of Increase per year	Total Number of Warheads by 2036
2026 – 2036	10-20	280-400
2026 – 2036	20-30	380-500
2026 – 2036	30-40	480-600
2026 – 2036	40-50	580-700
2026 – 2036	50-60	680-800

Source: Based on the author’s calculations. Data derived from SIPRI Yearbook 2025 and IPFM^[63]

Table 6 projects India’s warhead acquisition at approximate low-end rates from 2026 to higher-end levels by 2036, assuming that New Delhi draws on at least part of its existing 11 MT plutonium stocks (see Table 4). According to the author’s calculations, a prescription for 680-800 warheads would fall within the limit of India’s current plutonium stocks. However, it is important to note that India’s military-grade plutonium stocks, as shown in Table 4 is roughly 700 kilograms (kgs) plus or minus 160 kgs.^[64] In theory, India should be able to produce anywhere between 130 and 210 warheads at an assumed rate of four kgs per warhead with this stock, of military-grade plutonium.^[65] Four kgs is a ballpark figure. The amount of plutonium used is likely to vary depending on the type of warhead, which include boosted fission weapons, low yield weapons and thermonuclear weapons.^[66] The latter category of weapons are a capability India is not confirmed to possess. The higher the yield of the nuclear weapon, the more plutonium it is likely to consume.

India is also likely to use a sizeable amount of its plutonium stock to generate electricity by blending it with uranium oxide, creating a Mixed Oxide (MOX) fuel, limiting the need for enriched uranium and an enlarged nuclear arsenal. MOX fuel will power India’s conventional reactors and PFBR.^[67]

In addition, it is unlikely that all of these warheads would survive a nuclear first strike by India’s two primary adversaries, but it is worth taking, especially if India’s leadership does not want a very expansive arsenal that matches China and remains committed to an NFU policy. India also still has an opportunity to rev-up its nuclear weapons stockpile starting in 2026 at a rate of 50-60 warheads per year at the high-end (see Table 6), to reach a target of 680-800 weapons by the time China potentially fields an arsenal of 1,500 warheads or above by 2035-36. To be sure, the accumulation of 680-800 warheads by 2036 will be built on the basis of the current Indian stockpile of 180-200 warheads (see Table 5).

India’s 500 MWe Prototype FBR, expected to become fully operational in 2026 (as shown in Table 7), is capable of producing an estimated 140 kilograms of weapons grade plutonium per year. If 5 kilograms of weapons-grade plutonium are used per warhead, the PFBR could theoretically support the production of about 28 weapons every year on the condition the PFBR functions optimally.^[68] On this basis, plutonium from the PFBR alone could enable India to reach roughly half of the author’s proposed target of 680-800 weapons by 2036. Furthermore, only about half of this arsenal is likely to be operationally available at any given time, especially if India adopts a rotational system similar to those used by the US and Russia, and potentially China, as discussed in the previous part. Further, if a portion of India’s arsenal is to survive a nuclear first strike, India would need to build delivery systems, sensors, and missile defences.

On the other hand, if India’s defence and nuclear establishment opt for an arsenal larger than that recommended here, they would face a substantial burden in preserving the safety of the country’s nuclear weapons and in preventing unauthorised use, theft, or diversion of the stockpile. In fact, India will need to make steeper investments in human, financial, and technological resources to secure an arsenal that exceeds this author’s prescription. Conversely, if the Indian decision-makers and strategic establishment settles for an arsenal less than 680-800 weapons as prescribed by the author, they will be compelled to meet all the minimum requirements analysed in the next part.

Table 7: Fast Breeder and Research Reactors

Reactor Type	Operational Status
Prototype Fast Breeder Reactor (PFBR)	Expected to be Operationalised by mid-2026 and capable of producing Pu 239
Dhruva Research Reactor	Active and produces isotopes including tritium
Cirus Research Reactor	Decommissioned in 2010

Source: BARC^[69]

India does have the option of expanding its military-grade plutonium stocks from its Prototype Fast Breeder Reactor (PFBR) at Kalpakkam, which is expected to reach criticality within the next few months.^[70] Like China, India does not declare its civilian plutonium inventory under INFCIRC/549,^[71] unlike several other countries. In addition, some of its Pressurised Heavy Water Reactors (PHWRs) remain outside IAEA safeguards, giving India the option of increasing fissile material production, as shown in Table 8. Indeed, Anil Kakodkar, Atomic Energy Commission (AEC) Chairman, conceded the same in 2006: “Yes, very clearly. Not from civilian reactors, but from power reactors.” [PHWRs listed in Table 8].^[72]

India could also use its native reserves of natural uranium to build up its weapons-grade plutonium stocks, as China could do from its PHWRs. PHWRs are also necessary for the production of tritium, which is an isotope vital to the performance of nuclear weapons such as boosted fission devices and hydrogen bombs. India is known to produce tritium from PHWRs.^[73] Atomic weapons need to be replenished with tritium that decays in 12.3 years and more rapidly than other nuclear materials such as uranium-238 and plutonium-239.^[74]

Today, pressure to negotiate an FMCT—or arms control more broadly to restrain fissile material production—is not on the agenda of major powers such as the US, China, and Russia, though this could change over the next decade or so. Thus, the key issue and question for Indian decision-makers is determining the size of India’s nuclear arsenal. As long as the size of an Indian arsenal remains undefined or indeterminate, adverse consequences down the line could ensue, especially if India’s strategic managers are not confident about the quantitative strength of its arsenal by the time a consensus potentially crystallises among the designated NWS for a Fissile Material Cut-off Treaty (FMCT) in the next 10-15 years.

While nuclear test preparations can be undertaken within weeks, fissile material for weapons use cannot be accumulated at short notice. Building fissile stocks instead requires years, depending on the arsenal size India seeks. Facing two nuclear-armed adversaries—China and Pakistan—each possessing arsenals in the range of 700-800 weapons at a minimum, and in China’s case, it would be slightly above this figure were it to accumulate an arsenal within the ambit of existing stocks of plutonium (Table 4), achieving a comparable scale is likely to take a decade and should be an Indian objective, as shown in Table 6. This would not give India parity with China but would provide a buffer that still carries risks and vulnerabilities in the face of China’s growing nuclear strength. India will have far fewer operational warheads than China does by the mid-2030s, assuming Beijing diverts plutonium from its two HWRs and FBRs analysed in Part-I.

Table 8: India’s Unsafeguarded PHWR Reactors

Reactor	Power (Mwe)	Start-up Date	Safeguards Type
Tarapur-3	540	8/18/2006	Unsafeguarded
Tarapur-4	540	9/12/2005	Unsafeguarded
Madras-1	170	1/27/1984	Unsafeguarded
Madras-2	220	3/21/1986	Unsafeguarded
Kaiga-1	220	11/16/2000	Unsafeguarded
Kaiga-2	220	3/16/2000	Unsafeguarded
Kaiga-3	220	5/6/2007	Unsafeguarded
Kaiga-4	220	1/20/2011	Unsafeguarded

Source: IFPM^[75]

If the size of the Indian arsenal does not expand at a significant pace, while India continues to adhere to its nuclear test moratorium that precludes ‘hot’ testing of new warhead designs, should New Delhi reconsider the NFU principle enshrined in its nuclear doctrine? This leads to the third requirement under the maximum test: reversing India’s NFU. To date, no Indian government has shown any inclination to reverse or substantially modify the NFU pledge. This has been a longstanding commitment across successive administrations, irrespective of party affiliation, and indicates that India’s nuclear arsenal is conceived primarily as a retributive force. Although, former National Security Advisor (NSA) Shivshankar Menon did qualify that the NFU was not as watertight as enshrined in India’s nuclear doctrine.^[76] Indeed, he has stated that if India discovers that its adversaries are preparing for a nuclear attack, New Delhi might not be fettered from taking pre-emptive action by striking first.^[77] Thus, while New Delhi at a declaratory level, evident from its doctrine, remains wedded to NFU, it may be compelled to mimic China operationally by veering close to a LoW capability.

Indeed, there is evidence of missile canisterisation by India that enables better storability and transportation of pre-mated warheads with missiles.^[78] It will enable rapid nuclear-armed missile launches such as the Agni-5 Intermediate Range Ballistic Missile (IRBM).^[79] Further, India has deployed the Arihant SSBN on deterrent patrols, especially against Pakistan, indicating that India may be potentially deploying mated warheads in peacetime.^[80] It will not be long before India is compelled to do the same against China.

Given the premium India places on assured second-strike capability and a retaliation-only posture, New Delhi would therefore need to accelerate efforts to improve and accumulate key delivery systems and platforms, sensors, and missile defences that meet the minimum requirements. These cover requirements ranging from four to six laid out under the maximum requirements.

Minimum Requirements for India: Hypersonic Missiles, Sensors, and Nuclear Submarines Are Imperative

Hypersonic Missiles

India remains in the testing phase of its hypersonic missile capabilities. The HCM has completed a 1,000-second test but is still awaiting government approval.^[81] More crucially, even after clearance, it is likely to take seven years before being fully operationalised.^[82] An Indian HGV capability is at a more advanced stage, having undergone an initial developmental trial, and is expected to be inducted in two or three years following additional trials.^[83] Despite this progress, the HGV will still require extensive user trials across the three services.

Beyond hypersonic missiles already under development, India will need—at a minimum—to place a higher premium on three sets of capabilities: sensors, missile defence against hypersonic missiles, and the subsurface leg of its nuclear forces. Together with hypersonic systems, investments across these five areas will be necessary to offset constraints on the size and quality of the arsenal. The first of these is space-based sensors capable of detecting missile launches, including incoming ballistic and hypersonic projectiles. At present, several space-related “blindspots” persist against conventional military threats and against nuclear-tipped ballistic and hypersonic missiles.^[84]

A core component of this architecture comprises ground-based sensors, which India does possess as part of its ballistic missile defence (BMD) system. However, these are largely limited to detecting and tracking hypersonic projectiles in their terminal phase because of horizon constraints. For instance, India is in possession of the Swordfish Long-Range Tracking Radar (LRTR), developed by the Defence Research and Development Organisation (DRDO) and the Bharat Electronics Limited (BEL).^[85] An Active Electronically Scanned Array (AESA) endows the Swordfish LRTR with the capacity to detect missiles at a range of 1,500 km or more and forms the core element of India’s BMD network.^[86] Yet the Swordfish is primarily effective and geared to detecting and tracking ballistic missile threats, such as Intermediate-Range Ballistic Missiles (IRBMs). As a ground-based radar system, it is constrained in its interception capabilities against hypersonic missile threats.^[87] Similarly, sea-based sensors have limitations in that they are terrestrially based, like their ground-based counterparts, and are constrained by the horizon and the Earth’s curvature.

Consequently, surface-based radars and sensors can, at best, detect and track hypersonic missiles very late in flight, often only during the terminal phase, leaving little time for interception.^[88] India operates a single surface vessel, INS Dhruv, to track ballistic missile launches by China and Pakistan, but it is not configured to track hypersonic missiles.^[89] India also maintains static, ground-based missile tracking stations at Sriharikota, Chandipur, and the Thumba Equatorial Rocket launching centre; however, these are primarily designed to track Indian missile and space vehicle launches. India will need to build more of them along the coast and supplement them with more ground-based mobile tracking stations.

India also lags in sea-based interception capabilities, even against ballistic missiles, let alone hypersonic threats. At most, India so far has conducted its flight trial of a sea-based endo-atmospheric ballistic missile interceptor from the INS Anvesh in April 2023.^[90] The latter is also capable of tracking incoming ballistic missiles. More such surface vessels are required for tracking and intercepting ballistic and hypersonic missiles. Further, the future Project 18-Class Guided Missile Destroyers (GMDs) of the Indian Navy (IN) will be armed with sea-based BMD capabilities for the first time.^[91] None of India’s past or in-service GMDs with the IN have such capabilities. The missiles aboard the next-generation Project 18-class destroyers will have interceptors derived from Phase II of India’s land-based BMD architecture.^[92]

Phase I of India’s BMD system was operationalised in 2024. Known as the Prithvi Air Defence (PAD) system, it comprises two layers: an exo-atmospheric layer engaging targets around 80 km, and an endo-atmospheric layer—the Advanced Air Defence (AAD) system—engaging targets at 15-25 km.^[93] Deployed at sea, these systems would provide a limited interception envelope against incoming ballistic missiles, helping protect high-value coastal military installations and critical infrastructure.^[94] While the Project 18-class destroyers are expected to be armed with hypersonic missiles for surface warfare roles, they are not intended for hypersonic missile interception.^[95]

Sensors

A space-based sensor network, integrated with robust ground- and sea-based sensor architecture, is crucial to augment India’s defence against ballistic missile threats and, even more so, hypersonic missiles—capabilities India currently lacks. Operation Sindoor, conducted in May 2025, exposed these limitations, particularly the absence of space-borne sensors. As a space technology expert averred: “You look at the war [Op Sindoor] that we had with Pakistan. The Chinese were providing real-time, space-based intelligence on our attack vectors, on our aircraft, on our movements, and that was for a skirmish. Imagine what happens when you are fighting the Chinese.”^[96] V. Narayanan, Chairman of the Indian Space Research Organisation (ISRO), who observed that India’s limited constellation—adequate against Pakistan—would be insufficient against China, given India’s 7,000 km coastline and extensive, partly contested land frontiers requiring persistent surveillance.^[97]

The nature of this space-based sensor network is therefore as important as its existence. While a Low Earth Orbit (LEO) constellation is essential, it is not sufficient on its own. In September 2024, the Modi government approved nearly INR 27,000 crore for the Space-Based Surveillance (SBS)-3 programme, catalysed by lessons from Operation Sindoor.^[98] SBS-3 envisages a 52-satellites in Low Earth Orbit (LEO) constellation, representing belated but progress of an LEO-based Small Satellite (SmSat) network, it is also being driven with urgency due to the inadequacies of India’s space-borne sensor network evident during Op Sindoor.^[99] It is again Pakistan that is shaping or at least playing a catalytic role in determining the acquisition priorities of the Modi government and not China. The latter’s capabilities have to be a core metric and benchmark for determining nuclear-related defence procurements for India.

An LEO-based SmSat network is only the first step, albeit indispensable for detecting and tracking China’s nuclear-tipped ballistic and hypersonic missiles. Hypersonic missiles are particularly difficult to intercept because of their extreme speeds, unpredictable trajectories, and low flight paths. Despite reports indicating that India will develop a multi-orbit space capability under SBS-3,^[100] an LEO-based SmSat network remains a vital prerequisite for hypersonic interception, albeit with clear limitations.

Spacecraft in a prospective LEO constellation, owing to their wide field of view, can detect the entire flight of a hypersonic missile from launch to impact. Space-borne infrared sensors in LEO are crucial in this role, as they can monitor the Earth’s surface across the visible, near-infrared, and shortwave bands of the Electromagnetic Spectrum (EMS). These will have to be paired to space-borne electro-optical sensors, which will be vital.^[101] However, despite the capacity of space-borne radars and IR sensors to detect heat emissions of freshly launched missiles, the heat signature of a hypersonic projectile is a tenth less than that of a rocket lifting off.^[102] That apart, space-based sensors will encounter problems with detection and tracking if there is a simultaneous launch of multiple hypersonic missiles or even a combination of missiles.^[103] In addition, LEO-based spacecraft are susceptible to jamming caused by EW and space debris.^[104] If China’s HGVs and scramjet-powered cruise missiles are to be effectively intercepted, they will have to go beyond dependence on LEO-based spacecraft.

Thus, replicating the American approach—if not in the number of spacecraft, then in terms of a multi-orbit capability, which includes spacecraft in LEO, Medium Earth Orbit (MEO), Highly Elliptical Orbit (HEO), and Geosynchronous Orbit (GEO)—is indispensable for India in addressing hypersonic missile threats. For hypersonic missiles, continuous tracking capability is an absolute imperative, necessitating a multi-orbit, space-borne sensor capability.^[105] There is evidence to indicate that India, under the aegis of the tri-service Defence Space Agency (DSA), has been pursuing near-continuous tracking through a multi-orbit sensor capability that faced a budgetary shortfall, which is now catalysed in the wake of Operation Sindoor.^[106] Given the confirmed hypersonic missile advances in China’s hypersonic missile capabilities, the development of a multi-orbit capability becomes absolutely imperative and urgent.

The US could also serve as a model. For instance, the American Space-Based Infrared System (SBIS) and Defence Support Programme (DSP), which is being bolstered pivotally in part by the LEO-based Resilient Missile Warning/Missile Tracking (MMWT), Hypersonic and Ballistic Tracking Space Sensor (HBTSS), and Next Generation Overhead Persistent Infrared (NGOPI), are vital to detecting and tracking, especially hypersonic missile launches. Northrop Grumman’s HBTSS, in particular, is among the constellation models India could consider for early missile warning and missile tracking. HBTSS represents a layered capability, with a higher concentration of satellites in LEO complemented by a smaller number in MEO and HEO, strengthening resilience and improving coverage over polar regions.^[107]

More specifically, as shown in part two of this paper, India lags considerably in hyperspectral satellite capabilities against its primary foe, China. Although India launched a HyperSpectral Imaging Satellite (HySiS) in 2018 equipped with an infrared sensor,^[108] this capability remains insufficient; additional HySiS satellites will be necessary as part of the SBS-3. Detection of HGVs will require electro-optical and infrared sensors because the glide phase generates considerable heat, touching thousands of Kelvins with strong infrared signatures.^[109] In IR imagery, however, HGVs appear much smaller and less distinct than ballistic missiles, often occupying less than 0.1 percent of the image area.^[110]

Artificial Intelligence (AI) in the form of deep learning techniques and small detection algorithms to augment the reliability of detection will be necessary for optimal sensor performance.^[111] Indeed, AI is vital if a vast amount of sensor data is to be processed in real-time. Supplementing these, as noted earlier, will be a necessity for radar satellites. Radar satellites are indispensable because of their capacity to determine range and velocity.^[112] Detection of hypersonic cruise missiles requires demanding detection and tracking requirements.

Ballistic and Hypersonic Missile Defence

Third, India’s hypersonic missile and hypersonic missile defence capabilities remain a work in progress. Missile defences—against both ballistic and hypersonic missiles—are critical because they provide Damage Limitation (DL) against a nuclear first strike against counterforce and countervalue targets.^[113] Given India’s commitment to NFU, its indefinite moratorium on nuclear testing, and limited nuclear force expansion, MD becomes vital. Basing strategy exclusively on Mutual Assured Destruction (MAD) or mutual annihilation in the face of India’s unilateral restraints in the aforementioned areas is unnecessarily risky. Therefore, it becomes imperative for Indian decision-makers to build a robust missile defence architecture.

Countries on the Chinese periphery will need not only BMD capabilities against the PLARF’s long-range precision missile attacks, as one United States Army analysis aptly points out,^[114] but equally defences against Chinese hypersonic missiles. This is indispensable for India in the early stages of war with China. Notwithstanding evidence that India is pursuing Project Kusha, which is a three-tiered ballistic and hypersonic missile interception system that includes M1 (150 km range), M2 (250 km range) and M3 (400 km range) interceptors, its effectiveness can only be sustained through a multi-tiered sensor network. Further, Project Kusha is expected to start user trials only from 2027. Even then, the space-based segment of the interception network will remain indispensable, and India has yet to fully develop this capability.

Although ground-based, airborne, and space-based sensors will help, the absence of sea-based sensors capable of detecting hypersonic missile threats will leave gaps in sensor coverage and fusion. Sea-based sensors will be as crucial to detecting hypersonic strikes against land-based targets as they are for threats at sea. Indeed, the purpose should be to ensure that the Project 18 Class destroyers are also equipped with advanced sensors and projectiles geared not just for ballistic, but also for hypersonic interception. As of today, the United States Navy (USN) is the only service among navies across the world that is closest to acquiring a hypersonic interception capability, with Raytheon modifying its SM-6 missile through the redesign of the steering control section of the missile to augment its speed and range.^[115] This is an area where the IN, along with the Indian government and the defence industrial establishment, will be required to invest resources.

For starters, hypersonic missile defence investments should prioritise point-defence systems for critical military installations and nuclear facilities, with gradual expansion to wider area defence, including cities. Taken together, integrated sensors across surface platforms, airborne assets, and space systems—coupled with effective interceptor capabilities—will be crucial to meeting the minimum requirements for credible deterrence against Chinese nuclear use.

SSBNs and SSNs

Supplementing these efforts is another condition to satisfy the minimum required against China’s accelerating nuclear strength and delivery capabilities: a strong Indian SSN fleet to escort the IN’s SSBNs. Indeed, India’s SSBNs would not constitute an effective or credible deterrent without SSNs. Yet SSNs have long been underprioritised, making trade-offs in resource allocation unavoidable.

Since the sea-based leg of India’s nuclear deterrent receives funding under a classified or highly secret budgetary head, then more monetary resources will need to be allocated for the accelerated development of SSNs, especially if the current and any future governments want to avoid financial cutbacks to the conventional fleet of the IN. On the other hand, the development of SSNs will consume resources that could also be drawn from technical and financial resources, allocated for the conventional fleet. The question then becomes which part of the IN’s conventional fleet suffers diversion of its capital outlays—the surface fleet or the submarine arm. As of today, the IN plans to acquire 17 new warships and nine new submarines.^[116] A Request for Proposal (RFP) to the tune of INR70,000 crore is under consideration as part of Project 17 B, the next- generation frigates, and two additional multi-purpose vessels.^[117] Another project for submarines that is in an advanced stage of cost negotiation are the Project 75-India (I) and Project 75 (add-ons).^[118] Project 75-I is for the cost of INR70,000 crore and the add-ons will cost an additional INR36,000 crore.^[119] A further INR36,000 crore is being allocated for eight Next Generation Corvettes (NGCs). Cumulatively these projects will cost roughly INR240,000 crore.^[120]

The government, in consultation with the IN’s force planners, may have to consider limiting the investment in the conventional fleet’s expansion in order to service the need to develop and expand the SSN fleet. If this cannot be done, raising the budget for the nuclear segment of the IN’s fleet becomes indispensable. Consequently, if the IN’s SSBNs are to deploy into the Western Pacific or far seas and develop a survivable capability that can launch a retaliatory nuclear strike from beyond the Indian Ocean Region (IOR), they will have no choice but to invest in a potent SSN escort force. India is pursuing such a capability under Project 77.^[121] India’s decision-makers face a trade-off; especially if they are unprepared to allocate more financial resources between building SSNs under Project 77 and conventional submarines under Project 75-I, then allowing the SSN development programme to supersede the development of conventional submarines must become a priority.

An alternative would be to limit the INR70,000-crore allocation for the future conventional submarine programmes or reduce capital outlays for surface combatants. Even more tellingly, India is capable of building SSNs indigenously, but not conventional submarines.^[122] The development of conventional submarines has only occurred by way of licenced manufacturing agreements with foreign Original Equipment Manufacturers (OEMs) such as Naval Group of France and the impending agreement with Germany’s ThyssenKrupp Marine Systems (TKMS) and Mazagon Dockyard Limited (MDL). Negotiating these agreements and creating sustained developmental strength for conventional submarine construction has been protracted, preventing the entrenchment of native design and industrial strength. This is not the case with nuclear submarines such as SSBNs and SSNs. As Indian Navy Chief Admiral Dinesh Tripathi has observed, SSNs are difficult to build, which also explains why their arduous development is making slow progress.^[123] The first of these submarines will be commissioned and ready only in the mid-2030s and the second by 2040 on the assumption that there are no additional delays.

Consequently, prioritising and accelerating the SSN leg of the subsurface fleet should acquire urgency, especially given the availability of domestic design and developmental capabilities. That apart, Project 77 is expected to build a total of six SSNs for an allocated US$14 billion.^[124] Indeed, conventional submarines to be developed under Project 75-I are likely to take the same amount of time as the first two SSNs, at best, and at worst longer, on the condition that the TKMS-MDL deal is successfully negotiated. That apart, the IN has tended to prioritise the development of surface vessels such as frigates and destroyers, for which it has gained substantial proficiency. This is not the case with submarines and aircraft carriers, which require sustained capital infusion for development and force projection beyond the IOR.^[125]

Notwithstanding the capital-intensive nature and protracted development cycles involved with all variants of submarines, the SSNs should be on top of the priority list. India may have to live with diminished strength in conventional submarines against Pakistan.

Existing and Future SSBNs and SLBM Delivery Systems

The SSBN fleet’s first submarine, INS Arihant, was commissioned in 2016 but suffered a major operational and technical setback within a year when significant water ingress affected its propulsion system, rendering it inactive for nearly a year.^[126] Arihant’s follow-on, INS Arighat, launched in 2017, was plagued by delays and commissioned in late 2024—an unreasonably long development cycle.^[127] Two additional SSBNs, Aridhaman (S4) and an unnamed vessel (S4), are under development and are likely to be more advanced than the in-service SSBNs, but they are still based on the Arihant design. Aridhaman, launched in 2021, is now expected to be commissioned in early 2026 following sea trials.

Despite the deployment of two SSBNs—Arihant and Arighat—and the prospect of two additional vessels of the same class expected to follow in the coming years, the subsurface leg of its nuclear triad requires vast qualitative improvement along three metrics: stealth, operating range, and missile payload relative to its Chinese counterparts. Let us consider missile range: the deployed missiles belonging to the sea leg of India’s nuclear capabilities consist of 750 km-range SLBMs, which are woefully inadequate for striking targets deep inside China, as shown in Table 9.

Table 9: Range of Indian SLBMs

SLBM	Range	Operational Status
K-15	750 km	Operational or In-service
K-4	3,500 km	Developmental/testing
K-5	5,000 km	Developmental
K-6	6,000-8,000 km	Developmental

Source: SIPRI,^[128] Hindustan Times^[129]

To be sure, the DRDO and the IN are developing and testing longer-range SLBMs, including the 3,500 km K-4, the 5,000 km K-5, and the 6,000-8,000 km K-6, as shown in Table 9. While these systems remain some distance from operational maturity, they are essential for holding China’s economic and military-industrial centres of gravity at risk.^[130] Operating range is restricted for India’s SSBNs because they are subjected to a bastion strategy, whose chief advantages are maintenance and protection.^[131]

India’s operational SSBNs are therefore likely to remain confined to the Bay of Bengal, where favourable sea depths support subsurface operations. Although this may ensure survivability against anti-submarine warfare (ASW) in the short term, it severely limits flexibility and the ability to operate beyond waters close to India’s coast.^[132] Such geographic restriction enables adversaries to plan countervailing measures—an approach the PLA Navy is already pursuing.^[133] Stealth constitutes another serious limitation: India’s SSBNs reportedly exhibit relatively high noise levels and remain vulnerable to low-frequency acoustic detection.^[134]

Most critically, the IN lacks an SSN fleet dedicated to escorting and protecting its SSBNs, a capability essential for sustained operations at extended ranges, particularly in the Pacific. This absence represents a fundamental weakness. SSNs are vital for projecting naval power over vast distances, and without them, the credibility of India’s sea-based nuclear deterrent will remain limited. The development and deployment of an SSN force are therefore unavoidable.  

Ultimately, SSBNs armed with long-range SLBMs and paired with SSNs will constitute the most critical component of India’s retaliatory nuclear strike capability. Given that India’s nuclear weapons are intended solely for retaliation—reflecting its commitment to NFU and a restrained posture that includes the separation of warheads from their delivery systems—comprehensive investment in the sea-based leg of the nuclear arsenal is indispensable. The ability to hold targets deep within the Chinese heartland at risk and to operate effectively in far seas is therefore central to the credibility of India’s deterrent.

Conclusion

When push comes to shove, India’s choices are unenviable but unavoidable. Comprehensively meeting the requirements of the maximalist test would entail prohibitively costly actions by the Indian government. If India chooses to forgo renewed nuclear testing and remains committed to NFU, it will have little alternative but to satisfy the requirements of the minimalist test outlined in the forgoing analysis. Under the minimalist requirement, India will need to reconsider its glacial accumulation of atomic weapons to reach a target of 680-800 weapons by 2036. It will need to accelerate the expansion as this author has prescribed in this paper.

In this context, the government and the defence and strategic establishment should not be overly constrained by the so-called “Goldilocks Dilemma”, which assumes that India, locked in a triangular competition with China and Pakistan, risks exacerbating instability with Pakistan if it acquires capabilities necessary to deter China.^[135] This line of reasoning presumes that India must get deterrence “just right” against both adversaries simultaneously. In practice, such calibration is unattainable in a triangular nuclear contest.^[136] An inability to prioritise clearly only weakens India’s capacity for effective military planning against either threat.

Accordingly, privileging investments in countervailing capabilities directed primarily at China—such as a comprehensive sensor network, nuclear platforms including SSNs, and missile defence—becomes necessary, even if this generates some degree of friction with Pakistan. Pursuing the minimalist requirements identified in the preceding analysis, would lend greater coherence and clarity to India’s nuclear posture. Many of the capabilities developed to deter China are fungible and applicable to Pakistan. Space-, air-, and surface-based sensor networks, for instance, would enhance detection, tracking, and identification of missile launches from both adversaries. Similarly, hypersonic missiles and defences would have relevance in the Pakistan context. Taken as a whole, these capabilities will bolster India’s deterrence against China.

---

Kartik Bommakanti is Senior Fellow, Defence and National Security, Strategic Studies Programme, ORF.

---

All views expressed in this publication are solely those of the author, and do not represent the Observer Research Foundation, either in its entirety or its officials and personnel.

Endnotes

• International Affairs
• Nuclear Security
• China
• India
• china nuclear arsenal
• china nuclear capabilities
• china nuclear expansion
• china nuclear policy
• china nuclear power program
• china nuclear stockpile
• china nuclear weapons
• china weapons of mass destruction

The views expressed above belong to the author(s). ORF research and analyses now available on Telegram! Click here to access our curated content — blogs, longforms and interviews.