By Ramachandran Rajeev Kumar — 2026-05-10

Cartography and Concrete: India's 2040 China Doctrine

For two decades, India's strategic budget has been organised around a question that turns out to be the wrong one. The question was: how do we manage Pakistan? An answer has been quietly emerging across an unusually wide range of strategic perspectives — from the architects of multi-alignment to the most hawkish American critics of it. Pakistan is the symptom. The cause is China. Strip Rawalpindi out of the picture tomorrow and the file we have to deal with stays, item for item, on the same desk.

This is not a comfortable finding. It dissolves the working theory under which a great deal of Indian strategic energy has been deployed since the early 2000s — the theory that if Pakistan could somehow be managed, isolated, or coerced into normalcy, the regional file would settle. It will not settle. The friction between New Delhi and Beijing is not manufactured through Rawalpindi. It is hegemonic and geographic. Pakistan is the cheapest instrument Beijing has for the purpose Beijing always had: keeping us tethered below the Himalayas while contesting our maritime approaches.

If that diagnosis is right, then a 2040 doctrine has to be written against the China question first, with Pakistan handled as a derivative file. Everything else flows from getting that ordering correct.

What follows has three parts. First, what actually sits in the file once Pakistan is removed from the centre of the frame. Second, the diagnostic instrument the doctrine needs: a basket of six tests that any future Beijing engagement signal must clear before it qualifies as more than theatre. Third, the doctrine itself — including a section on the United States that the analytical register would not otherwise permit.

The file that survives Pakistan's removal

The structural drivers between us and China are specific and named. Strip Pakistan out and this is what remains.

We have an undemarcated 3,488-kilometre Line of Actual Control. No two cartographic editions of this border have ever been agreed. The 1993 Border Peace and Tranquillity Agreement and the 1996 Confidence-Building Measures held until they did not. The condition all along was that they could be revoked unilaterally, and were, at Galwan in June 2020. That came after the conspicuous warmth of Wuhan in 2018 and Mamallapuram in 2019. The line has been a managed ambiguity. It has never been a settlement.

The trajectory is asymmetric. Through the 2030s, China's GDP is forecast to remain roughly four-and-a-half times ours. The asymmetry will narrow but it will not invert. Our window to embed strategic infrastructure — continental, maritime, institutional — closes inside this decade.

The Brahmaputra and the Sutlej run from upstream control we do not hold. Hydrological leverage is the form of Chinese pressure least-discussed in Indian strategic conversation and most resistant to fast-track countermeasures. Major dam-cascade projects on the Yarlung Tsangpo continue. They are not announced as instruments of pressure. They do not need to be.

Tibet is now militarised at a scale unprecedented since 1962. The Western Theatre Command's force-disposition rotations that began after Galwan have not been rolled back. By independent estimates, more than sixty permanent People's Liberation Army villages have been built or upgraded on the Indian side of the LAC. These are not military bases. They are sovereignty markers cemented into the geography. They cannot be paid for in summits.

Beyond the Himalayas, China's dual-use port architecture across the Hambantota-Gwadar-Kyaukphyu arc is now sufficiently advanced that the distinction between civilian and military use is a matter of months, not years. The Indian Ocean Region's security architecture is being rewritten while we are asked to celebrate "stability" in the language of disengagement protocols at Pangong Tso.

And Belt and Road has matured. The China-Pakistan Economic Corridor, the Bangladesh-China-Myanmar economic corridor, and the Maldives leasing posture together form a continental envelope that does not depend on Pakistani agency to function. Pakistan is one piece. The architecture survives its removal.

That is the file. None of it disappears under a Pakistan-removed counterfactual. Several entries actually get harder to address, because the convenient cover argument that "China's behaviour is mediated through Pakistan and therefore reformable through Pakistan-pressure" is gone.

The doctrine has to address the file directly.

Why "reset" is the wrong word

A great deal of Indian and international commentary in the last decade has been organised around the question of whether a Sino-Indian reset is possible. The answer, after Galwan, is structural. A reset in the diplomatic sense — summits, joint declarations, language about strategic partnership — is always possible, because all that is required is the willingness of two foreign ministries to issue text. A reset in the operational sense — one that reverses the file above, or even meaningfully arrests its expansion — is a different category of event entirely, and it has to be tested differently.

The mistake that has cost us most over the last fifteen years is the conflation of these two. Wuhan was a diplomatic reset. Mamallapuram was a diplomatic reset. They were followed, within months, by a sharp deterioration on the ground. We lost strategic time, and we redirected attention away from the file at exactly the moment the file demanded it.

The right frame is this: every reset overture from Beijing must clear an evidentiary bar set against the file itself. Not against atmospherics. Not against text. Against the maps and the concrete.

What follows is a basket of six tests. Each is grounded in a specific marker we can verify by satellite, by ground inspection, or by independent observation. Their analytical value is in their parallelism. Any one test is a necessary signal. None is sufficient on its own. A reset that moves on three or more of these tests, simultaneously and over a multi-year window, is real. A reset that moves on one or two while the others remain frozen is theatre that buys Beijing time at our expense.

This is the operational core of the 2040 doctrine.

The six tests

Test 1. Cartography and dual-use infrastructure rollback in the Western Sector

A real reset begins with maps. Specifically, an exchange of clarified maps on the Western Sector of the LAC — the section running through Ladakh from the Karakoram Pass to Pangong Tso — combined with verifiable rollback of the dual-use infrastructure added since 2017. That category includes the Pangong bridges, the road network through the Shaksgam Valley, and the model villages constructed across the Arunachal frontier on the Eastern Sector.

Cartography is the first test because everything else is reversible. Buffer zones can be redrawn. Patrol-point access can be renegotiated. Force levels can be adjusted up or down within seasons. Maps and concrete are not. They are how Beijing has demonstrated the difference between text and substance for the last fifteen years. They are the only instruments that can demonstrate the reverse.

Test 2. PLA forward deployments rolled back to pre-2013 levels and held there for five years

Force structure is the easiest variable to perform and the easiest to reverse. The relevant marker is therefore not the count of PLA personnel within the Western Theatre Command in any single quarter. It is the stability of a rollback over a window long enough to defeat seasonal cycling and crisis-period repositioning.

The benchmark is pre-2013 levels — the period before the systemic forward deployment that has since become the new baseline. The required window is five consecutive years of held position. Five years is the threshold beyond which a "tactical posture" stops being plausibly explained as such, because by then it has cost Beijing more in opportunity than any single negotiation can plausibly recoup.

Test 3. Termination of Chinese support for Pakistan's nuclear and missile programmes, and unwinding of CPEC sovereignty-violating projects

This test addresses the instrument question directly. If Pakistan is the cheapest instrument Beijing uses to keep us tethered, then a reset in the operational sense requires Beijing to abandon the instrument. Not to restrain it temporarily. Not to pause specific transfers. To unwind the institutional and capability commitments that constitute the instrument.

There are two markers. First, termination of Chinese technology and component flows into Pakistan's nuclear and missile programmes — the precise list is well-documented in open-source proliferation literature. Second, formal unwinding of those CPEC projects that violate Indian sovereignty, primarily the routing through Pakistan-administered Kashmir. Both are costly signals. That is what makes them tests. A Beijing willing to issue summits but not willing to unwind these commitments has not changed its position on the file.

Test 4. Verified Chinese withdrawal to pre-April 2020 patrol points at Depsang

This is the most precise test in the basket. It targets the Depsang Plains specifically — the Bottleneck, the Y-Junction, and patrol points PP10 through PP13 — where Chinese ingress in 2020 was achieved without combat and where the field geography is now a matter of degradation rather than denial.

Depsang is the precision test because it is the location where Beijing has the strongest revealed preference not to withdraw. If Depsang comes back to pre-April 2020 patrol-point access, the calculus has changed. If it does not, no amount of disengagement language at other points along the LAC tells us anything new. Depsang is the load-bearing test, the one most resistant to atmospheric movement.

Test 5. Multi-year reduction of post-Galwan PLA force-disposition rotations into depth, satellite-verified

This test addresses what happened after Galwan. Beijing's response to the 2020 confrontation was not a return to status quo ante. It was the development of a new posture in which forward-deployed PLA formations rotate into operational depth on a schedule designed to defeat Indian intelligence cycles. The forward-deployment rotation pattern is itself the new baseline, and a reset that does not address it leaves the new baseline intact.

The verification standard for this test is independent satellite observation — third-party analysis from operators capable of providing imagery that neither government can edit. The window required is multi-year, for the same reason as Test 2: shorter windows are gameable through cycling.

Test 6. Dismantling of the 60 permanent PLA border villages on the Indian side of the LAC

The sixth test is the most demanding. The PLA border villages — by independent estimate, more than sixty — are not military infrastructure in the conventional sense. They are sovereignty markers cemented into geography. They are populated, supplied, and administered as civilian settlements. They are also monitored, networked, and integrated into the Western Theatre Command's intelligence picture. They cannot be removed by quiet adjustment of patrol protocols. They have to be demolished, physically, and the demolition has to be verified by satellite imagery and on-ground monitoring — by some equivalent of the United Nations Military Observer Group framework that has historically been politically untouchable in this theatre.

This test is included despite its demanding nature because without it, the other five tests are reversible. A Beijing willing to roll back forward deployments while holding sixty cemented villages on the Indian side of the LAC has reset nothing. It has issued a temporal pause. The villages are the test of irreversibility.

How to use the basket

These six tests are operationally complementary. They cover six distinct domains: cartography, force-structure persistence, the China-Pakistan instrument, the terrain at Depsang, post-Galwan rotation patterns, and the border villages. Movement on one without the others is insufficient. Movement on three or more, simultaneously, over a multi-year window, is the threshold for treating a reset overture as real.

This is what we should do every time Beijing makes a move: read it against the basket. Not against the diplomatic register. Not against the joint statement. Against the maps and the concrete.

The doctrine that follows — and what it does under reset

The basket is the diagnostic. The doctrine is the prescription. Seven pillars, calibrated against the China-hostility variable rather than treated as fixed points.

First, the 2040 horizon as the planning frame. India's demographic peak, the GDP threshold of seven to ten trillion dollars, and the post-current-phase realignment all land in the 2035-2040 window. The doctrine is built for the shape of the world that arrives in the 2030s, not for the world that exited in the 2020s.

Second, strategic autonomy under stress. This is not the autonomy of 1955, which was equidistance under conditions of weakness. The 2040 version is autonomy under conditions of weight. India in 2040 is a leading power, not a balancing power. Multi-alignment in this register means active engagement on overlapping coalitions whose terms we set. The Quad and the International North-South Transport Corridor are not contradictions. They are the geometry of autonomy at scale.

Third, Russia anchored as ballast. The distinction between ballast and counterweight matters. A counterweight, in the alliance sense, is expected to actively oppose the adversary. Russia will not perform that function for us, in 2025 or in 2040. What Russia performs is a different and more durable function. Russia prevents Indian envelopment in either a unipolar United States system or a United States-China condominium. Russia denies Beijing a free hand in Central Asia and the Arctic. Russia provides the technology, energy, and institutional cover that lets us avoid being priced by any single partner. The Russia anchor is a hedge against envelopment, not a bet on Russian strength.

Fourth, Iran preserved as a strategic asset. Chabahar, the Shahid Beheshti terminal, the IPGL operationalisation, the ten-year IPGL-Ports and Maritime Organization contract, and the wider International North-South Transport Corridor architecture together constitute the only land-bridge to Central Asia that does not transit Pakistan. The doctrine does not concede this under the Iran-United States-Israel inflection because the inflection is exactly the test the doctrine was built to absorb.

Fifth, Pakistan derailed via the China-proxy framing, with resource discipline. Four levers run in parallel. One: outsource the cost to China by accelerating CPEC's strategic and economic exposure. This carries the highest leverage. Two: diplomatic neutralisation via Gulf Cooperation Council normalisation. Three: amplification of internal Pakistani contradictions. Four: sustained economic strangulation through international financial-compliance forums. The win condition is not Beijing's surrender of Pakistan, because Beijing will not surrender Pakistan. The win condition is our resource reallocation — away from a derived problem, toward higher-stake vectors.

Sixth, the Gulf Cooperation Council anchored economically. The Saudi Public Investment Fund's announced India exposure, ADQ and Mubadala participation in National Investment and Infrastructure Fund tranches, the Aramco Ratnagiri thesis, and the India-UAE Comprehensive Economic Partnership Agreement together constitute an economic anchor that the China-United States rivalry cannot easily disrupt — because the Gulf states' Vision 2030 programmes depend on it as much as we do.

Seventh, the United States as transactional. This is the pillar that requires the polemic register, and the next section is where I drop the analytical voice.

A note before that section: the doctrine described above is calibrated for current conditions, and is built to bend, not break, under reset. Under reset conditions the Russia ballast pillar de-prioritises, because the envelopment risk eases. Iran retains its field value but loses its autonomy-signalling weight. The Quad downgrades from a balancing instrument to an issue-based coordination forum. The United States remains transactional regardless. The doctrine is built for a variable China-hostility level, not for a fixed one.

On the United States — the polemic section

I now drop the analytical register and say the part of this argument that the analytical register obscures.

The premise that India's strategic posture should be organised around tighter, sober alignment with the United States — the premise that informs most American writing on India and a great deal of Indian writing that takes the American writing seriously — is built on a category mistake. Treat the United States' technology stack, its capital markets, and its security architecture as permission slips revocable by any incoming administration at any Federal Register notice. That is what they are.

The evidence is not theoretical. CAATSA — the Countering America's Adversaries Through Sanctions Act — has been used as a brinkmanship instrument against our procurement choices that displeased Washington. The Trump-era trajectory on H-1B compression, on tariff theatrics, and on the open patronage of Khalistan-aligned activism, demonstrates that American interest in India is conditional on the political moods of constituencies that have nothing to do with our interests. The October 2022 and October 2023 export-control rules — the so-called "small yard, high fence" — were written without consulting Delhi. NVIDIA's Blackwell allocations are not commercial decisions; they track White House mood. Cadence and Synopsys, the electronic design automation tools without which serious semiconductor design is impossible, are American companies under American export-control jurisdiction. Our entire semiconductor ambition has, at the frontier nodes, no anchor that does not pass through American goodwill.

This is not a relationship in which we can plan a 2040 doctrine. It is a relationship in which we have to plan around the relationship. Use it where it is offered. Never depend on it. Never pay with sovereignty for its temporary availability.

There is one condition under which I would update this position. A Senate-ratified, administration-proof technology-transfer treaty — not a memorandum of understanding, not an executive-branch initiative — covering electronic design automation tooling, sub-7-nanometre fabrication, and jet propulsion. That is what Washington offered Tokyo in 1960. Until Washington offers Delhi the same, the language of "strategic partnership" with the United States is a permission slip in formal dress.

The United States goes first if the doctrine is forced to sacrifice an axis. Not because the United States is hostile — it is not — but because the United States is the axis on which we have the least durability of dependence. That is arithmetic, not nostalgia.

What we should do now

Returning to the analytical register, the consequences of the doctrine are five.

One. Adopt the verification basket as the formal protocol for reading every future Beijing engagement signal. Stand it up institutionally — in the National Security Council Secretariat, in the Ministry of External Affairs' East Asia division, in the relevant intelligence agencies. Make movement on three or more tests, simultaneously and over a multi-year window, the threshold for upgrading the relationship beyond transactional management.

Two. Verify the Russia anchor on a continuous quarterly cadence against the structural-absorption markers: Arctic basing rights to the People's Liberation Army Navy, nuclear command-code transfers to BeiDou, defence-industrial dependencies on Chinese servos and electronics. If three or more of these markers cross thresholds within an eighteen-month window, downgrade Russia from strategic ballast to "preferred non-aligned supplier" and reallocate accordingly.

Three. Run the four-track Pakistan derailment in parallel rather than sequentially. The win condition is our resource reallocation — measured in defence-budget percentage shifted from the western frontier to the China file, in diplomatic attention shifted from Pakistan-management to Indian Ocean Region posture, and in intelligence-cycle priority. Track quarterly.

Four. Hold the Iran pillar through the Iran-United States-Israel inflection. The pillar's value is unrelated to the inflection. Concede targeted costs — secondary sanctions exposure, slowed Israeli technology cooperation, slowed Phalcon-2 delivery cycles — as the price of the pillar.

Five. Treat the United States relationship as transactional until the Senate-ratified treaty exists. Use what Washington offers; do not pay for it with sovereignty. Build the 28-nanometre fabrication base at Tata-PSMC Dholera, the design-software hedge through European and Japanese tooling, and the propulsion programme through partnerships that are not cancellable by Federal Register notice. Be one node behind on frontier artificial intelligence compute and three nodes ahead on sustainment, sovereignty, and scale.

The doctrine, in one sentence: read every Beijing move against the basket, build the architecture for envelopment-prevention rather than counterweighting, hold the Iran pillar through the inflection, derail Pakistan via resource discipline, and treat the United States as a partner whose terms are revocable.

Cartography and concrete. Anything short of that is theatre we have already paid the tuition for.