What India's own government data says about the distance between how much active pharmaceutical ingredient the country consumes and how much it makes — and why, five years and ₹25,000+ crore of incentive schemes into "Atmanirbhar" bulk drugs, China's share of India's API imports sits at 73.71%, roughly where it stood before the schemes began.
Pharmaceuticals · Import Substitution · China Dependency
Five Years and ₹25,000 Crore Later, China Still Supplies Three-Quarters of India's Drug-Making Ingredients
Revised
· v1.0.1 · what changed
India does not publish a single "domestic API demand" figure, so any clean "we consume X, we make Y, the gap is Z" claim about the country's active-pharmaceutical-ingredient dependence is, by construction, an estimate rather than a reconciled government number. What the government does publish, across dozens of PIB releases and Parliament replies since 2020, is enough to show the shape of the problem clearly: China supplies not quite three-quarters of everything India imports to make its own medicines, that share has barely moved since PLI incentives for bulk drugs began, and the gap between "capacity exists" and "China dependence has actually fallen" shows up inside the government's own data, molecule by molecule.
The measurement problem
Before comparing consumption to production: three separate government data series get routinely conflated in casual writing about this topic. Each has a different scope, comes from a different source, and updates on a different schedule.
| Series | Value | What it actually measures |
|---|---|---|
| Finished-formulation market | US$60bn | Total Indian Pharmaceutical Market — all sold formulations, domestic + export. Not upstream API demand. |
| of which domestic consumption | US$23.5bn | FY2023-24 split — domestic sale of finished drugs, still not API-level. |
| API/bulk-drug/DI imports | US$4.35bn | 200 HSN categories, FY2024-25 — closest proxy to "how much API India brings in," but an import figure, not a consumption figure. |
| Domestic API production, total | not found | No PIB/DoP/Economic Survey document states an aggregate annual API production value or tonnage for the industry as a whole. |
| Domestic API demand, total | not found | Production + imports − exports = apparent consumption is never published as a single government estimate. |
The closest available whole-industry figure: National Accounts Statistics 2024 puts the entire "Pharmaceuticals, medicinal and botanical products" industry's output at ₹4,56,246 crore for FY2022-23 (value added ₹1,75,583cr, 9.26 lakh persons employed) — but this bundles formulations and bulk drugs together, with no API-only breakout. The Economic Survey 2025-26 separately states India has roughly 500 API manufacturers, accounting for about 8% of the global API industry — a capacity-share claim, not a production-value one.
Where capacity exists
The PLI Scheme for Bulk Drugs is the government's primary import-substitution lever for APIs specifically — distinct from the much larger PLI–Pharmaceuticals scheme, which targets formulation scale rather than API import substitution. Progress is real but incremental, and running well behind the physical infrastructure meant to house it.
PLI–Bulk Drugs: three snapshots
| As of | Companies / projects | Investment | Cumulative sales | Imports avoided |
|---|---|---|---|---|
| Jun 2025 | 32 cos / 48 projects | 4,709 | 1,962 | 1,483 |
| Sep 2025 | — / 48 projects | 4,763 | 2,315 | 1,807 |
| Dec 2025 | 33 products / 48 projects | 4,814 | 2,720 | 2,192 |
Of 41 notified critical products, 33 have a subscribed manufacturer, 48 greenfield projects were approved against those 33, and 38 projects covering 28 products have actually been commissioned — roughly 56,800 MT/year of new domestic capacity. Against the ₹6,940cr outlay, and the import bill this is meant to displace (US$4.35bn ≈ ₹36,700cr per year), ₹2,192cr of cumulative avoidance since the scheme began is a modest dent so far.
Bulk Drug Parks: still under construction, five years in
| Park | Land (acres) | Total cost (₹cr) | Central grant | Status |
|---|---|---|---|---|
| Gujarat (Jambusar) | 2,015 | 2,507 | ₹1,000cr cap | Environmental clearance Feb 2024; civil works mostly complete; effluent-treatment tenders awarded, only partly done; parking tender Nov 2025 |
| Himachal Pradesh (Haroli) | 1,405 | 1,923 | ₹1,000cr cap | Road/drainage/water tenders awarded and in progress; zero-liquid-discharge tender floated; site grading underway |
| Andhra Pradesh (Nakkapalli) | 2,002 | 1,877 | ₹1,000cr cap | Fencing complete; external water pipeline installing; additional land's environmental clearance granted Jan 2026 |
Scheme tenure runs FY2020-21 to FY2026-27 — with one year left on the clock and all three parks still building common infrastructure (roads, effluent treatment, water, power), not yet leasing plots to API manufacturers.
Two named flagship projects
The import dependency
China is not one supplier among several for India's APIs — it is, for practical purposes, the only significant external source. No other single country exceeds 2.5% of the import bill.
| Origin | Share | Value (US$mn) |
|---|---|---|
| China | 73.71% | 3,204.70 |
| EU (bloc) | 13.64% | 593.03 |
| Singapore | 2.49% | 108.26 |
| United States | 1.96% | 85.22 |
| Japan | 1.82% | 79.13 |
| Switzerland | 1.03% | 44.78 |
| Mexico | 0.80% | 34.78 |
| UK | 0.76% | 33.04 |
| Hong Kong | 0.53% | 23.04 |
| Malaysia | 0.51% | 22.17 |
| Remainder | 2.75% | 119.56 |
Source: DGCIS via PIB (same release as the China/total figures above). Only China's dollar value was published directly; every other origin's value here is calculated as share × the US$4,347.75mn total, since PIB's release gave shares for these origins but not individual dollar amounts. "Remainder" is PIB's own catch-all for origins below Malaysia's 0.51%; the eleven rows sum to 100.00% of the published total. No government breakdown of the EU's figure by member state was found; trade literature names Italy and Germany as likely leaders, but that attribution is not government-sourced.
By therapeutic category
The government's own framing: "Therapeutic segments with high import dependence include antibiotics, anti-fungal, anti-amoebic, gastrointestinal disorders, antidiabetic, endocrine and hormonal disorders, cardiovascular, oncology, female infertility, contraception, neurology/substance use disorders and essential amino acid deficiency." Grouping the 90-line Chinese-dependence Annexure by value (the grouping is this post's own; the underlying figures are government-published):
| Category | HS lines | China value (US$mn, approx.) |
|---|---|---|
| Antibiotics & beta-lactams (penicillins, cephalosporins, tetracyclines, macrolides, quinolones) | 15 | 1,608 |
| Other organic-chemical KSMs, not elsewhere classified | 49 | 932 |
| Steroids / hormones | 3 | 112 |
| Vitamins (B1, B6, B12, C, H) | 6 | 80 |
| Antipyretic / analgesic intermediates | 9 | 54 |
| Sulfa drugs & nitroimidazoles | 8 | 15 |
By molecule: a working sample of the 90-line list
| Molecule | FY23-24 | FY24-25 | Trend | PLI-localised? |
|---|---|---|---|---|
| 6-APA (penicillin intermediate) | 94.08% | 95.92% | worse | no |
| Erythromycin & salts | 93.04% | 97.65% | worse | no |
| Penicillins & salts | 77.03% | 92.87% | worse | yes |
| Norfloxacin | 99.99% | 99.90% | flat | yes |
| Rifampicin | 89.17% | 89.01% | flat | yes |
| Vitamin B12 (cyanocobalamin) | 92.62% | 93.82% | worse | no |
| Vitamin C (ascorbic acid) | 82.45% | 87.76% | worse | no |
| Vitamin B6 | 63.50% | 88.28% | worse | yes |
| Ciprofloxacin & salts | 97.98% | 98.02% | flat | no |
| Ibuprofen | 99.34% | 80.73% | improved | no |
| Clavulanic Acid | off the ≥70% list entirely | substituted | yes | |
| Atorvastatin | off the ≥70% list entirely | substituted | yes | |
| Metformin | no PIB figure found | — | no | |
Full 90-row Annexure available from the underlying PIB release; this is a representative sample.
Value understates it — volume is worse
A rupee-value view of dependency is the flattering one. China's share of India's bulk-drug imports rose from 64% to 71% by value between FY14 and FY23 — but from 62% to 75% by volume over the same period (CARE Ratings, secondary), meaning Chinese APIs are priced below the rest of the import basket and a rupee-only KPI understates physical dependence. The same pattern holds year-on-year: FY24→FY25 bulk-drug import value rose 1.8% while volume rose 9.3% (pharmabiz/DGCIS, secondary) — landed unit prices are falling, which is what you'd expect if a dominant low-cost supplier is gaining share.
Trend assembled from multiple point-in-time releases (no single government document covers all years): China's import share rose through FY14–FY23 (roughly 40% → 71–75%), and the most recent primary snapshot (Feb 2026) puts it at 73.71% for FY24-25 — near the FY23 peak, not meaningfully reduced by five years of PLI. No PIB or Parliament document states a target percentage or year for closing this gap.
Capacity ≠ substitution
The single most important pattern in the government's own data: being on the official list of 28 "localised" PLI products does not reliably predict falling China dependence. Some molecules genuinely substituted; others have domestic capacity and unchanged or worsening import dependence, in the same dataset, from the same press release.
This is not a reconciliation problem introduced by this post — it is a first-hand inconsistency inside the government's own release (the same PIB release's Annexure-I and Annexure-II, published the same day). It is the strongest evidence available that the policy needs an outcome metric (import tonnage displaced), not only an input metric (capacity commissioned).
Why the gap persists
Government explanations exist, but thin out fast from narrative to number.
"Despite a very strong base, due to low-profit margins and non-lucrative industry, domestic pharmaceutical companies have gradually stopped manufacturing APIs and started importing APIs, which was a cheaper option with increased profit margins on drugs. With the availability of cheaper APIs from China, the pharmaceutical industry relies heavily on imports."TIFAC (Dept. of Science & Technology), 15 Jul 2020
"Potential risks related to import of APIs arising from geopolitical uncertainties include Single-Source vulnerability, Price volatility and predatory pricing. Such single source vulnerabilities threaten self-reliance and pharmaceutical security as experienced during the COVID-19 period."PIB, 3 Feb 2026 — the closest any release comes to alleging predatory pricing
Between these two, most releases since 2020 repeat a single unexplained phrase — imports persist "mainly due to economic considerations" — without ever decomposing what that means: labour cost, power tariff, land, capital cost, or Chinese state subsidy are each plausible components, and none is quantified in any PIB or Department of Pharmaceuticals document located for this post. That is a checkable gap: the Medical Devices PLI scheme's own PIB language explicitly names "high cost of finance, inadequate availability of power" as a quantified manufacturing disability versus competing economies — the API/bulk-drug scheme has never received the equivalent analysis in public.
Ways to compete
Ten recommendations, each tied directly to a gap surfaced above.
| # | Recommendation | Evidence |
|---|---|---|
| 1 | Pay PLI incentives against import tonnage displaced, not capacity commissioned. Vitamin B6, Rifampicin and Norfloxacin show capacity can exist for years without moving the import-dependence needle. | Annexure-I vs II, same release |
| 2 | Track and publish dependency by volume (MT), not value alone — China's volume share already exceeds its value share, so a value-only KPI flatters the headline. | CARE Ratings FY14-23 series |
| 3 | Commission a quantified India-vs-China cost-disability study for APIs specifically: power tariff, effluent-treatment compliance cost, capital cost per MT, cost of finance. Never decomposed for bulk drugs the way it has for medical devices. | Absence confirmed against the Medical Devices PLI precedent |
| 4 | Investigate trade remedies on the flattest-dependency molecules. Rifampicin, Norfloxacin and Ciprofloxacin sit at 89–99% dependence with commissioned domestic capacity that isn't gaining share. | Molecule table above; the government's own "predatory pricing" risk language |
| 5 | Publicly reconcile the "53 vs 41 critical bulk drugs" list. If 12 molecules are no longer critical, say why; if they still are, notify them. | 2020 Cabinet approval vs every PIB release since Feb 2021 |
| 6 | Accelerate common infrastructure at all three Bulk Drug Parks. Import substitution cannot come from fenced, graded land, and the scheme's FY2026-27 sunset is one year away. | Park-status table above |
| 7 | Give small-molecule APIs a programme at Biopharma-SHAKTI scale. The new biologics-focused Biopharma SHAKTI initiative (₹10,000cr/5yr) is larger than the entire PLI–Bulk Drugs outlay (₹6,940cr) — despite small-molecule import dependency being the larger, more immediate problem today. | Budget 2026-27 vs PLI-Bulk-Drugs outlay |
| 8 | Fund a Strategic API Reserve, formally. No confirmed, funded stockpiling mechanism exists for the ~90 China-concentrated molecules, unlike petroleum or foodgrain strategic reserves. | Referenced only in secondary sources, unconfirmed in any PIB release found |
| 9 | Court non-China API investment, not just domestic-from-scratch manufacturing. The EU supplies 13.64% and every other country under 2.5% — there is no meaningful second supplier to fall back on. | Country table above |
| 10 | Publish an explicit target — a percentage and a year. Every current release reports achievement to date; none states where China's import share should be, or by when. | Confirmed absent across every release reviewed |
Sources: Press Information Bureau releases PRID 1606725 (17 Mar 2020), 1607483 (21 Mar 2020, Cabinet approval), 1638810 (15 Jul 2020, TIFAC), 1701048 (26 Feb 2021), 1847446 (2 Aug 2022), 2010924 (2 Mar 2024), 2085345 (17 Dec 2024, pharma market), 2121425 (Apr 2025), 2158120 (19 Aug 2025), 2200939 (9 Dec 2025), 2221492/2221766 (1 Feb 2026, Budget/Biopharma SHAKTI), 2222528 (3 Feb 2026, dependence country table), 2224376 (6 Feb 2026, park costs), 2237414 (10 Mar 2026, China-import Annexures), 2243248 (21 Mar 2026, Economic Survey backgrounder), 2244474 (24 Mar 2026, Gujarat park status), 2246026 (27 Mar 2026, PharmaMed). Underlying trade data via DGCIS as cited within PIB 2222528 and 2237414. Two figures are explicitly marked secondary rather than government-sourced: the FY14-23 value-vs-volume trend (CARE Ratings, via Business Standard/PTI, 6 Aug 2023) and the FY24-25 volume-growth figure (pharmabiz.com, citing Ministry of Commerce/DGCIS). Figures marked "not found" in this post were genuinely searched for and not located in a primary government source as of compilation, not omitted for convenience. This post's headline dependence figure (73.71%) and the Bulk Drug Park land/status figures were independently spot-checked against separate search results before publication and found consistent.
- v1.0.1 — 16 September 2026 — filled in the country-origin import table's dollar values (calculated as share × the published US$4,347.75mn total, since PIB gave shares but not per-origin dollar amounts beyond China's) and restored Malaysia (0.51%) as its own row after finding the published shares summed to 99.49% without it. Attempted to source real per-origin values directly from TRADESTAT/DGCIS instead of calculating them; found only chapter-broad HS29/HS30 data (a 6×-larger, differently-scoped series covering all organic chemicals or all pharmaceuticals, not the specific ~200-line API/bulk-drug/intermediate basket this table uses) accessible through this sandbox's network, so calculation from the already-correctly-scoped percentages was used instead.
- v1.0.0 — 16 September 2026 — first published.
About this article: Researched, written and edited by Umashankar Triplicane Dwarakanathan, with AI research assistance; every figure is meant to trace to the primary source cited. See the Editorial Policy for how sourcing, AI use and corrections work.
