Tearing apart the COMEX silver registered stock drain

Silver.

Financial Twitter is calling it a gold proxy. CNBC's guest this morning mentioned "safe haven flows" and "de-dollarization." Both are watching the wrong instrument. The price of silver is not the story. The registered stock number at COMEX is the story, and it has been telling the same thing for six months while the commentary completely missed it.

Silver is not a monetary relic. It is an industrial consumable that also happens to trade on a financial exchange. Every solar panel requires roughly twenty grams of it. Every EV's battery management system runs through silver-sintered circuits. The semiconductor fabs in Taiwan and Arizona that everybody is subsidizing? Silver paste is the conductive layer that makes the chips work. Nobody on the monetary metals desk is modeling fab output. They're still writing about central bank reserve diversification.

Registered stocks are the silver that has been assayed, vaulted, and certified for delivery against a futures contract. They are the physical backing of the paper market. When that number drops, the contractual promise becomes harder to keep. I traded gold and silver spreads for long enough to know that registered stock levels are the only number in this market that actually matters. Everything else is noise with a Bloomberg terminal attached.

The number has been draining for the better part of eighteen months. The financial press noticed when gold hit its record. Nobody noticed the vault reports.

Numbers only. Here is how the COMEX silver vault system works and where the stress fracture is.

The total open interest on COMEX silver — the sum of all outstanding futures contracts — runs around 550 million troy ounces on most active trading days. The registered stock backing those contracts has collapsed to roughly 100 million oz. That ratio was sitting above 0.35 in early 2022. It has been cut approximately in half. The market has not repriced this. It has barely noticed it.

There is also an "eligible" category — silver sitting in approved COMEX vaults that has not been certified for delivery. Eligible metal can become registered, but it requires the owner to submit it for re-assay and pay the certification fees. When holders consistently choose not to convert eligible to registered, it signals something: they are not sellers, they are accumulators. The eligible pile is large. It is not moving.

Here is the supply chain as it actually exists right now:

This is not a demand spike. It is a structural deficit that has been running for three consecutive years. Mine supply grew almost not at all over the last decade — silver is primarily a byproduct of copper, lead, and zinc mining, so the silver output is constrained by base metal economics, not silver prices. You cannot simply drill more silver mines when the price moves. The ore is attached to something else.

Meanwhile solar panel installations continue to outpace every projection made in 2020. The Silver Institute's own industrial demand estimate for this year has already been revised upward twice. The models were built before anyone understood the scale of the energy transition buildout. They are catching up slowly.

The weak link is not the mine supply. It is not even the solar panel demand. It is the assumption that the eligible vault inventory acts as a permanent buffer — that whenever registered stocks get tight, the eligible holders will convert and sell, and the system self-corrects.

That assumption has been quietly breaking down. A meaningful portion of the eligible inventory in COMEX vaults is held by entities that are not in the business of selling. Sovereign wealth funds. Commodity ETFs that have shifted from paper to physical. A handful of very large family offices that treat silver as monetary insurance and have demonstrated zero inclination to part with it at current prices. Their metal sits in the eligible pile. It will not move to registered. It is scenery.

Here is what the CFTC report is telling you: the speculative community has not piled in yet. Managed money net longs are elevated but not at the levels that precede a washout. The retail crowd is buying silver ETFs at a rate not seen since the 2021 Reddit squeeze. But unlike that episode, the physical demand underneath is genuine this time. The meme and the fundamental are pointing the same direction simultaneously. That does not happen often.

The specific technical vulnerability is the July delivery cycle. Open interest for July silver futures is running above 120 million ounces. Registered stocks are at roughly 100 million. If even a fraction of July longs stand for delivery rather than rolling their contracts — which more industrial buyers are doing because they need the physical metal, not a price bet — the registered inventory cannot cover it without emergency conversion from eligible, which requires cooperation from holders who have been demonstrating they have no interest in cooperating.

This is the setup that precedes a delivery squeeze. Not guaranteed. But the machinery for one is in place, the ignition is sitting in July open interest, and I have been watching this pattern since the Hunt Brothers made it famous. The details change. The plumbing does not.

The sequence is mechanical. It requires no conspiracy, no coordinated action, no one bad actor. It just requires enough July longs deciding they want the metal.

Each wave accelerates the next. The industrial buyers are the third wave — and they are not optional buyers. A solar panel manufacturer in Anhui cannot decide to use less silver. A semiconductor fab running silver paste sintering cannot swap in aluminum and requalify the process in six weeks. They buy what they need at whatever price is required. That is the physical market's version of a hard stop.

Where does the capital go in this scenario? Not the silver ETFs that hold allocated metal — those will move with spot but offer no leverage. The edge, if there is one, sits in the primary silver miners: first-quartile cost producers whose revenue is 70%-plus silver, unhedged, with operating mines inside jurisdictions that are not currently on fire. There are not many of them. The sector is thin. That thinness is itself a feature — large capital flows into a small pool move prices significantly.

The counterfactual: July open interest rolls normally, eligible holders convert enough to cover, the market murmurs and moves on. This is the base case. Most delivery cycles resolve without drama. But the registered-to-open-interest ratio has not been this tight without producing at least a localized spike in the last decade. I have watched it resolve quietly and I have watched it not. The difference was usually a single large industrial buyer deciding not to roll.

The physical market has been running a deficit for three years. The vaults are draining. The eligible buffer is less cooperative than the model assumes. The July cycle is large. That is the machine. Watch the vault reports, not the monetary metal commentary. The vault reports have been telling this story since November.