Tearing apart the silver thrifting illusion

Silver.

The sell-side is having a field day with this one. Total silver demand from photovoltaic solar manufacturing fell 19% in 2026. Photovoltaic demand is set to drop from 186.6 million ounces last year to roughly 151 million ounces this year. Every analyst note I've read this month reaches the same conclusion: silver's biggest growth engine is fading, the thesis is broken, take profits.

They're reading the dashboard again. The engine is somewhere else.

Read that again. The market's single biggest demand headwind — a 35-million-ounce cut from solar — is fully baked into the 2026 forecast. The deficit still widened from 40.3 million ounces to 46.3 million. That's not a demand story wearing demand's clothes. That's a supply story that nobody wants to read because it doesn't fit the neat narrative about thrifting.

I've been wrong about silver before — badly wrong in 2021, when I thought the Reddit squeeze had structural legs and it didn't. This is different. The Reddit crowd was trying to manufacture a squeeze. This time the squeeze is manufacturing itself, one depleted vault at a time, and most of the market is looking the other way.

Story off. Numbers on.

Global silver mine production: 846.6 million ounces in 2025. The 2026 forecast from Metals Focus: 844.1 million ounces. Essentially flat — a decline of 2.5 million ounces at a time when the market needs 46 million more than it’s producing. That flatness is the whole problem.

Where did those 762 million ounces come from? From vaults. Deliverable exchange inventories dropped from 290 million ounces at the start of 2024 to below 210 million by October 2025. COMEX registered silver sits at 96.3 million ounces as of last week. LBMA London vaults held 27,729 tonnes at end of January 2026, but 225 million ounces of that migrated from London to CME vaults between December 2024 and October 2025 on tariff arbitrage. They didn't produce more metal. They moved it from one room to another and called it supply.

Now look at where the metal is mined. North America posted decade-low silver mine production in 2025 as ore grades declined in Mexico. Peru — 130 million ounces annually, roughly half of China's silver concentrate imports — issued Emergency Decree 003-2026 on May 11 to bail out Petroperú, signaling an energy infrastructure crisis that puts mining operations directly at risk.

The thrifting narrative says demand is weakening. The vault data says the opposite. Both can be reported. Only one of them can be true at the physical delivery window.

India. The world's largest physical silver consumer. Meets more than 80% of its demand through imports. In 2025, it inhaled 7,158 metric tonnes — 230 million ounces — 22% of global supply.

On May 16, 2026, the DGFT reclassified high-purity silver bars from “Free” to “Restricted.” In June, they added grain and powder. More than 90% of silver imports now require a government-issued license. Most banks haven't secured one yet. The result: imports collapsed from 747 tonnes in January to an estimated 29 tonnes in June.

Delhi built a dam to protect the rupee. Behind that dam, 230 million ounces of annual demand is pooling. So far in 2026, India has imported 1,837 tonnes — a 16% year-over-year decline. The demand didn't evaporate. It's waiting. Wedding season starts in October. Festival buying ramps in September. Industrial users are already paying 10% over benchmark and the seasonal surge hasn't even begun.

Meanwhile, the paper market is retreating from the metal. SLV has shed over $3.1 billion in ETF outflows this year. Managed money is sitting at a modest net long of about 11,282 contracts — nothing aggressive, nothing committed. The gold-silver ratio hit 69:1, near its 50-year historical average. That ratio is the market's way of saying: we don't believe in you.

I've seen this configuration before. October 2025. London silver suffered an unprecedented liquidity squeeze on October 10 despite over 800 million ounces sitting in vaults — only about 140 million of which was free float available for trading. Shorts couldn't source metal for near-term delivery at any price. The World Silver Survey described what followed as a squeeze that “became self-fulfilling” as shorts covered and liquidity vanished.

That's the gap. And nobody on the sell side has written a note about it because the licensing story is a Reuters wire buried under three Iran headlines. They have, to my knowledge, not checked what silver premiums are doing in Mumbai this week.

There are two triggers. Either one is sufficient.

First: India relaxes the licensing regime. Delhi built the import wall to defend the rupee, but the domestic premium blowout is already creating political pressure. Silver is not a luxury item in India — it's wedding jewelry, temple offerings, industrial feedstock. A $6.50 premium is a tax on every family planning a wedding this fall. If even a handful of banks secure import licenses and the floodgate cracks, 230 million ounces of deferred demand hits a global market that's already running a 46-million-ounce annual deficit.

Second: the gold-silver ratio compresses. At 69:1, it's near its long-term average. Every time it's crossed above 80 in the last thirty years, the mean reversion was violent — silver outperformed gold by 40% or more within 18 months. The ratio doesn't need to hit 80 for the trade to work. It just needs to stop climbing.

Coin and bar demand already jumped 14% to 217.7 million ounces in 2025. The Silver Institute forecasts another 18% jump in 2026. That's retail absorbing metal that the industrial supply chain also needs. Two lines of buyers converging on the same shrinking pool of deliverable silver.

Where does capital go? Not into SLV — the ETF becomes a source of metal during squeezes, not a beneficiary. The edge, if there is one, is in mid-cap primary silver producers with significant unhedged output and reserves outside the Peru and Mexico risk zones. Companies whose revenue is a direct function of spot price, not locked-in forward contracts signed at $24.

The sell side says the thrifting story killed the silver bull case. The vault data, the Indian import dam, and six years of cumulative deficits say something else entirely. In my experience, when the paper narrative and the physical plumbing disagree this completely, the plumbing wins. It just takes longer than anyone with a quarterly P&L is comfortable waiting.

The thrifting is real. The conclusion drawn from it is not.