The FY2025-26 headline: total merchandise imports $776 bn, exports $442 bn, deficit −$334 bn (PIB/DGCI&S provisional) — widened from −$184 bn in FY2018-19. The full 98-chapter decomposition is in this blog's Trade Balance — HSN-wise analysis; the commodity slice looks like this:

The composition is what matters. Oil sets the floor (~$200 bn/yr, price-driven — see the Mineral-Oil Ch.27 dashboard), but it actually fell $15 bn this year. The marginal pressure came from precious metals: chapter 71 supplied 37% of the year's entire import increase. And within it, the arbitrage channels visibly migrate as customs closes each one — the platinum-group line collapsed −92% ($5.4 bn → $0.4 bn) when the “platinum alloy” duty route was shut, and the flow snapped straight back into bare gold, while sourcing shifted to doré from mining countries (Ghana +449%, Chile +2,486% YoY) to capture the refining-duty differential. Silver imports rose 2.5× to $12.1 bn in a single year (from $4.8 bn).

The rupee slid ~12% over the past year (₹86 → ₹96.4) while India's inflation ran below the US's — purchasing-power parity alone would have predicted strengthening, as unpacked in Effective Buying Power & the China Comparison. The realized depreciation is essentially all real: a widening goods deficit whose marginal driver was bullion, and partly self-reinforcing — rupee weakness makes gold a better hedge, gold buying widens the deficit, the deficit weakens the rupee. The services surplus (~$189 bn) and still-large reserves explain why it is a slide and not a crisis; Where India's Forex Goes maps which country deficits the outflow lands in.

Bare metals — gold, silver, copper, aluminium and friends — now cost India roughly $120 billion a year in imports. The strangest pair of numbers inside that bill: copper scrap imports of $3.4 bn and aluminium scrap imports of $4.6 bn. That is $8 billion a year of other countries' waste, bought to feed India's secondary smelters — because India's own end-of-life metal isn't collected, sorted, or channelled to authorised recyclers at scale.

One metal broke the pattern. In FY2025-26, India's lead exports (chapter 78, $1.27 bn) exceeded its lead imports ($1.17 bn) — the only base metal where the circular loop has fully closed. Why lead? Lead-acid batteries are the one waste stream in India with mature, price-discovered reverse logistics: every battery shop, UPS replacement and truck depot feeds an established collection chain back to the smelters. The same collection-first logic is why Norway's deposit-return lottery works — the mechanics are dissected in The Recycling Lottery Playbook, which proposes the identical incentive layer for Indian waste streams.

The companies built on the lead loop show real financials (TTM, 3 Aug 2026):

Source: yfinance TTM data, 3 Aug 2026. Every profitable recycler at scale is lead-first — the trade data and the corporate landscape tell the same story from opposite ends. Operating margins are 3–10%; returns come from asset turnover (Jain earns ~29% ROE on a 3.4% margin). The dominant cost is buying scrap — feedstock access is the moat. P/E of 31–53 means the circular-economy tailwind is already priced in; the thesis being right and the stocks being buys are different claims. Nothing here is investment advice.

The growth runway is not new technology — it is that $8 bn of imported scrap, plus what India throws away. Two policy hooks aim squarely at the feedstock gap: the Ministry of Mines' ₹1,500 crore recycling incentive under the National Critical Mineral Mission (e-waste, Li-ion battery waste, end-of-life vehicles — all copper- and aluminium-dense), and DST's call for proposals on end-of-life solar-PV recycling. Silver and solar are the next lead: CEEW's Nov-2025 study shows PV recycling loses ₹10–12k/tonne today, but EPR certificate trading plus high-silver TOPCon modules entering the waste stream after 2028 flip it profitable — against 11,221 kilotonnes of cumulative PV waste projected by 2047.

Recycling attacks the metals leg. The rest of the recurring import bill responds to the older lever: building the capacity at home. This blog has mapped that terrain sector by sector, and the pattern is consistent — the wins come where a specific feedstock or intermediate is identified and priced, not where a headline target is announced:

• Chemicals — the Chemical Import Substitution executive summary and the Top-15 deep dive isolate a $67.9 bn deficit across 827 HSN-8 codes, with BPCL/RIL capex as the counterforce (tier base case ~$8.3 bn/yr by FY30).
• Energy — ethanol blending is the one substitution programme with a decade of audited results: ₹1.97 lakh crore of forex avoided and 316 lakh MT of crude displaced, traced in Annadata to Urjadata.
• Textiles — the man-made-fibre loop (ethanol chemistry → MEG → polyester → technical textiles) is mapped in From Fuel to Fibre.
• Fertilisers — the subsidised-import overlap (urea imports +148%, DAP +76% in FY25-26) and the NIPU-2026 response are in India's Subsidised Imports.
• The long series — eight years of chapter-level trend, including which chapters actually improved, in HSN-wise Historical Trends.

Gold is the honest exception. Recycling won't dent it and no factory substitutes it — it is a savings-behaviour and duty-arbitrage story, and the only levers are financial (sovereign gold instruments, duty design) and enforcement of the arbitrage channels documented above.

The commodity half of India's deficit is structural, but it is not monolithic. Sequence the response by what the data says already works: lead proved the recycling model → the ₹1,500 crore NCMM scheme and EPR rules extend it to copper and aluminium (substituting the $8 bn imported-scrap line first) → solar waste brings silver and silicon into the loop after 2028 — while chemicals, ethanol and fibre capex substitute the recurring industrial imports. The companies are already built. What the trade data says they need isn't smelters — it's India's own scrap.