The Curious Case of Gold - Part 2: The Crash

On Wednesday, January 29, 2026, gold peaked at $5,594. Silver touched $121.64. Both were all-time records - set on the same morning, in the same frenzy, after a month in which silver had gained 68% and gold had gained 29%. Traders who had been in the market for thirty years said they had never seen anything like it.

Forty-eight hours later, gold was at $4,700. Silver was at $78. The largest single-day decline in precious metals in over four decades had just happened. Somewhere in the range of $7–15 trillion in paper market value had evaporated. Mining stocks fell 10–15% in a session. Leveraged ETFs lost half their value in an afternoon.

This is the forensic account of those two days: what happened, in what order, and why the story you probably heard - that a Fed Chair appointment caused it - is true only in the narrowest possible sense.

(This is Part 2 of a three-part series. Part 1 covered the rise. Part 3 asks what comes next.)

Table of contents

• The week that preceded everything
• The match: Kevin Warsh
• The powder keg: how it was packed
• The cascade mechanics
• Silver: why it was so much worse
• What the data shows
• What the crash reveals
• The rebound, and what it means

The week that preceded everything

To understand the crash, you have to understand the week that built toward it. January 2026 was not a normal month for precious metals. It was a month in which markets appeared to lose contact with normal gravity.

Gold opened January at roughly $4,300 - already a level that would have seemed extraordinary a year prior. By the third week, it had crossed $5,000. Then $5,200. Then $5,400. Institutional investors who had missed the 2025 rally were scrambling to get in. Retail traders, who had watched gold double over the prior year from the sidelines, were now piling in with leverage. ETF inflows were running at the highest pace since March 2020. China's retail gold ETFs alone had seen billions of dollars of inflows in the prior weeks. The Gold RSI - the Relative Strength Index, a technical measure of overbought conditions - had touched 90. Silver's RSI hit 93.86, a level not seen since 1980.

The gold-to-silver ratio, which tracks how many ounces of silver it takes to buy one ounce of gold, had compressed to around 46 - nearly the lowest in decades. That compression alone was a warning signal. Historically, whenever the ratio gets that tight, it means silver has run far ahead of its fundamentals on speculative momentum.

On Tuesday, January 27, the Federal Reserve held its first meeting of 2026. It left interest rates unchanged at 3.5–3.75%, having cut three times at the end of 2025. The statement was read as neutral-to-slightly-hawkish. Markets absorbed it, and gold barely moved.

On Wednesday morning, January 29, gold set its all-time record: $5,594 per ounce. Silver's record came the same morning: $121.64.

Then things started to break.

The match: Kevin Warsh

At some point on the afternoon of January 29 - the exact timing of the leak matters - markets began pricing in what was confirmed the next day: President Trump was nominating Kevin Warsh to succeed Jerome Powell as Federal Reserve Chair, effective when Powell's term expired later in 2026.

Warsh is a former Fed governor, a Wall Street veteran, a Stanford economist. He is known for one thing above all others: hawkishness. He has consistently argued that central banks print too much money, cut rates too aggressively, and fuel financial instability through excessive accommodation. In markets that had been pricing in two or three rate cuts over the course of 2026, the prospect of a Fed Chair who might deliver zero - or even a hike - was a genuine shock.

The reaction was immediate. The dollar strengthened. Expectations for rate cuts repriced sharply. Treasury yields rose. And gold, which depends on the implicit logic of falling real rates and a weakening dollar, fell.

But here is the critical thing to understand: Warsh did not cause the crash. He lit the fuse on a powder keg that had been packed for weeks.

If Warsh had been named during a normal market environment - one with moderate positioning, reasonable leverage, and orderly futures markets - the gold price would have dropped 3%, maybe 5%, and recovered within a few days. What happened instead was a 21% decline from peak. The difference between a 5% and a 21% move is not the spark. It is everything that was already in place when the spark landed.

The powder keg: how it was packed

There were at least five structural conditions that turned a routine piece of news into a historic crash. Each one was independently significant. Together, they were catastrophic.

1. The CME's shift to percentage-based margins (January 13)

On January 13, 2026 - two weeks before the crash - the CME Group, which operates the COMEX exchange where gold and silver futures are traded, quietly made a procedural change that would prove consequential. It switched from fixed-dollar margin requirements to percentage-based ones.

Previously, a trader holding a gold futures contract had to post a fixed dollar amount as collateral - say, $15,000 - regardless of whether gold was at $3,000 or $5,000 an ounce. Under the new system, the margin was calculated as a percentage of the contract's current value. That sounds like a technicality. It isn't. What it means in practice is that as prices rise, the collateral required rises automatically and continuously.

As gold climbed from $4,000 to $5,500 during January, the effective margin requirements tripled in dollar terms for every trader in the market. This was happening silently, without announcements, in the background - a slow tightening of the screws on every leveraged position in the market.

2. The CME's explicit margin hikes (January 31 and February 2)

On January 31 itself, after the market had already begun crashing, the CME announced it was raising percentage-based margins further. Gold margins went from 6% to 8% for standard accounts; 6.6% to 8.8% for high-risk accounts. Silver margins went from 15% to 18%. A second round of hikes followed days later, taking silver to 19.8% for high-risk accounts.

Each hike sent a fresh wave of forced selling through the market: traders who had been barely meeting their margin calls at the old rates suddenly couldn't meet them at the new ones.

3. Overleveraged Chinese retail positions

Throughout January, an enormous and extremely concentrated wave of leveraged buying had come from Chinese retail investors. Gold-backed ETFs in China had absorbed billions of dollars in a matter of weeks. Chinese futures traders on the Shanghai Futures Exchange had accumulated positions that were, by the standards of the market, extraordinary.

When the selling began, these positions unwound rapidly and without regard for price. Leveraged traders don't get to wait for a better offer - when the margin call comes, they sell whatever they can, as fast as they can. The Chinese unwinding was one of the dominant forces driving the speed and severity of the decline.

4. The Shanghai Futures Exchange had already been tightening

Before January 31, the Shanghai Futures Exchange had quietly made multiple rounds of adjustments to restrict silver futures trading - raising its own margin requirements and reducing position limits over the course of January. These moves had been generating pressure on Chinese silver traders for weeks, setting up a situation where any trigger could cause a rapid unwind of accumulated positions.

5. Month-end and options mechanics

January 31 was the last trading day of the month. In futures markets, month-end is already a period of elevated volatility, as traders roll positions, close books, and rebalance. The day also coincided with significant options expiration. As prices fell, options dealers who had sold puts (essentially insurance against declining prices) found themselves suddenly needing to sell the underlying futures to hedge their exposure - a dynamic called a gamma squeeze in reverse, where the hedging itself accelerates the move. Each wave of selling forced more selling.

The cascade mechanics

Once these conditions were in place, the actual mechanics of the crash were almost classical. Gold fell $380 - nearly 7% - in just 28 minutes on January 29, as the Warsh news began to circulate. That initial drop was large enough to trigger stop-loss orders across tens of thousands of accounts. Stop-losses are automatic sell orders that activate when prices fall below a threshold. They are designed to limit losses - but in a market with many people at the same stop levels, they cause avalanches.

Goldman Sachs later described the mechanics with unusual clarity: as prices fell, dealer hedging flipped from buying into strength to selling into weakness, investor stop-outs were triggered, and losses cascaded through the system.

By the morning of January 30, gold had fallen more than 10% from its peak. By the close of January 31, it was down 21% from the record high. Silver, which had peaked at $121.64, hit an intraday low just below $75 before recovering slightly. The ProShares Ultra Silver fund - a leveraged ETF - fell more than 62% in a single session.

The trading was not orderly. Reports emerged of halts on certain contracts, of bid-offer spreads blowing out to extraordinary widths, of institutions unable to execute large orders without moving the market by several percent. One analysis suggested that at certain points during the afternoon of January 30, the effective liquidity in silver futures had essentially evaporated.

Silver: why it was so much worse

Gold fell 21% from peak to trough over the crash period. Silver fell 40%. The gap is not a mystery - it reflects the fundamental differences between the two markets.

Silver's total market size is roughly one-tenth of gold's. A move of $1 billion into or out of silver has ten times the price impact of the same flow in gold. This is why silver is sometimes called the "Devil's Metal": it amplifies everything. In the rally, silver's smaller size meant that the same force of buying pushed prices much higher relative to gold. In the crash, the same mechanics worked in reverse with brutal efficiency.

Silver also had a specific problem: a large fraction of its late-January price had been driven by momentum traders - retail investors and short-term speculators - rather than by the industrial or long-term monetary buyers who tend to hold through volatility. When prices started falling, there was no deep base of committed holders to absorb the selling. The momentum crowd all ran for the same exit simultaneously.

The chart above shows the full arc - the parabolic January rally in both metals, the peak on January 29, and the violent snap back. Note that while gold had largely recovered to the $5,000 range by mid-February, silver was still struggling to reclaim even $85. The asymmetry reflects everything about the difference between these two markets.

What the data shows

A few numbers tell the story of January 31, 2026 more precisely than any narrative can:

The RSI for silver at its peak - 93.86 - is the number that best explains what happened. An RSI above 70 is conventionally considered overbought. Above 80 is extreme. 93.86 is, by any standard, a parabola nearing its apex. The last time silver RSI was anywhere close to that level was in 1980, the year the Hunt brothers cornered the silver market and it crashed 80%.

At the same time, the gold-to-silver ratio had compressed to roughly 46:1. Within two days of the crash, it had snapped back to 57:1 - a move of that magnitude in the ratio, in that short a time, is essentially unheard of.

What the crash reveals

Strip away the drama, and the January 31 crash reveals several things about modern commodity markets that are worth sitting with.

Paper markets are not physical markets. The gold and silver that was "lost" in January 31's crash existed only on screens - as futures contracts, ETF units, and leveraged positions. The actual gold bars in actual vaults did not change. Physical demand, from jewellers in India, from manufacturers in China, from coin dealers in the United States, was if anything higher during the crash, not lower. Premiums on physical silver rose sharply as dealers struggled to source metal while paper prices were collapsing. The two markets - physical and paper - had decoupled.

Leverage creates synchronised exits. One of the defining features of a leveraged market is that everyone's stop levels tend to cluster in predictable places. When prices are rising, leverage amplifies gains and creates a self-reinforcing loop of optimism. When prices fall through key levels, the same architecture runs in reverse: margin calls, stop-losses, and forced liquidations all hit simultaneously. The cascade that Goldman described - dealer hedging flipping from buying to selling, investor stop-outs triggering, losses cascading - is not a unique pathology. It is the normal operating logic of a futures market under stress. The January crash did not happen despite the market working correctly. It happened because it was working correctly, just under conditions of extraordinary leverage.

The fundamental thesis did not change. On January 28, the reasons for owning gold were: a weakening dollar, de-dollarisation, central bank buying, fiscal deficits, and geopolitical uncertainty. On February 1, all of those things were still true. The Warsh nomination changed the short-term rate path expectation, but it did not change any of the underlying structural factors that had driven the 2025 rally. Within two weeks of the crash, gold had recovered to $4,800. Within three weeks, it had crossed $5,000 again.

Silver's move was, in part, a meme. Not entirely - the fundamental case for silver, anchored in supply deficits and industrial demand, is real. But by late January 2026, a significant portion of silver's price was being driven by the kind of momentum-chasing and retail-FOMO dynamics that have more in common with GameStop in 2021 than with a commodity repricing. One analyst at Interactive Brokers described the silver trade as "momentum trading that exceeded even the type of outsized moves that we have seen in a wide range of speculative assets." When a precious metal starts getting compared to a meme stock, it is a sign that price has temporarily separated from value.

The rebound, and what it means

The rebound was almost as striking as the crash. By February 5, gold had recovered to $4,800. By February 10, it was back above $5,000. By February 20 - the day before this piece was written - it was trading at $5,062. Silver had recovered to the low $80s. The gold-to-silver ratio had stabilised around 63.

The speed of the recovery supports the view that the crash was a liquidity event - a forced liquidation driven by leverage and mechanics - rather than a fundamental repricing. Markets that undergo genuine fundamental re-evaluations tend to recover slowly, if at all. Markets that undergo forced liquidations tend to snap back once the selling pressure is exhausted.

Goldman Sachs, which published a note during the peak of the chaos, maintained its $6,000 target for year-end 2026. UBS raised its 2026 target to $6,200. JP Morgan kept its $5,000 forecast, calling the selloff "a positioning correction within an ongoing uptrend." Not everyone was sanguine - some analysts argued that the parabolic nature of the January rally had permanently impaired the market's credibility, and that a long consolidation phase was likely. But by mid-February, the evidence was pointing toward recovery rather than collapse.

There is one lingering detail that deserves mention. On January 31, 2026, at precisely the moment the silver price hit its intraday low of $78.29 - its absolute nadir during the crash - JP Morgan reportedly closed approximately $10 billion in silver short positions: 3.17 million ounces, all settled at that exact price. The timing has attracted significant commentary in financial circles. Whether it represents a well-timed trade, a coordinated rescue operation, or simply coincidence is, as of this writing, unclear. What is clear is that $10 billion in short covering at the exact bottom of a market creates buying pressure that can stabilise prices. The crash stopped. Then the recovery began.

So what comes next? Is the bull market in gold and silver structurally intact, or did January 2026 mark the peak? What should you actually do with any of this?

Part 3: What Now? - coming next.

Sources: CNBC Markets; Bloomberg Commodities; CME Group Clearing Advisory CH-26-041; Goldman Sachs Metals Research; JP Morgan Global Research; World Gold Council Gold Demand Trends; SD Bullion Market Analysis; Bullion Trading LLC; Finance Magnates; TradingKey.