Gold Rips to $4,652 With Managed Money at the 90th Percentile and Swap Dealers at the 96th — Both Sides of the COT Are Now at 52-Week Extremes Simultaneously
Copy link
Gold advanced 7.32% over the COT reporting week ending August 25 to $4,652 per troy ounce, and the positioning data that accompanied that move describes a market that has arrived somewhere it very rarely goes. The InProved Terminal’s COT dashboard flags six separate 52-week extreme readings across the four-metal complex this week, and two of the most consequential are on opposite sides of the same gold contract: Managed Money at the 90th percentile and described as crowded long, and Swap Dealers at the 96th percentile and described as near maximum short. Speculators as a whole added 21,145 contracts (+9.5% week-on-week) to reach 243,334 net long — an addition of 2,114,500 ounces, or 65.8 tonnes, of paper gold exposure in five trading sessions — while Commercials mirrored them almost exactly, extending their net short by 21,167 contracts to -279,585, likewise 65.8 tonnes. The symmetry is not coincidental; it is the mechanical identity of a futures market, in which every new speculative long requires a counterparty willing to take the other side. What makes this week’s print analytically distinctive is that both of those counterparties are now positioned at the outer edge of their own one-year distributions at the same moment, and they did so in the days immediately preceding Warsh’s speech and the hawkish repricing that followed it. A market in which the bulls and the bears are simultaneously at 52-week extremes is not a market with a clear directional edge; it is a market with an unusually high probability of a violent resolution in one direction or the other, and the risk management implications of that dominate everything else in this week’s data.
The InProved Terminal’s COT decomposition for the August 25 reporting date allows the two sides of gold’s futures market to be read against their own histories rather than in isolation, and that is where the significance lies. Commercial (Producer/Merchant plus Swap) net short stands at -279,585 contracts, equivalent to -27,958,500 ounces or -869.6 tonnes, an increase in short exposure of 8.2% week-on-week; the 52-week percentile on that reading is 88% and the five-year percentile is 86%. Swap Dealers alone are at -245,027 contracts (-24,502,700 ounces, -762.1 tonnes), having added 16,370 contracts of short (+7.2%), and their 52-week percentile is 96% with a five-year reading of 95% — a position the InProved Terminal explicitly flags as near maximum short. On the long side, Speculators sit at +243,334 contracts (+24,333,400 ounces, +756.9 tonnes) at the 86th percentile on 52 weeks and the 88th on five years, while Managed Money specifically holds +144,747 contracts (+14,474,700 ounces, +450.2 tonnes) at the 90th and 79th percentiles respectively. Notably, Managed Money’s own weekly addition was modest at +3,099 contracts (+2.2%, or 9.6 tonnes); the bulk of the 21,145-contract speculative increase came from the Other Reportables cohort rather than from the systematic macro funds, which is a meaningful distinction about who exactly is doing the buying at these levels.
The historical context for a simultaneous double extreme is what converts this from a positioning curiosity into a tradeable risk parameter. Across the past decade of COMEX gold data, weeks in which Managed Money exceeded the 88th percentile while Swap Dealers exceeded the 94th percentile short are rare — they cluster around the terminal phases of the 2011, 2020, and 2024 advances. The resolution in each case followed a consistent pattern: price continued higher for a further one to four weeks as the last marginal buyers were absorbed, and then reversed sharply when the flow of new longs stopped, because the Swap Dealer community at maximum short has no capacity to add further hedges and must instead begin buying back to reduce risk once price stalls. That buy-back is what produces the sharp two-to-three-day squeezes that punctuate the tops of gold cycles, and it is also what makes those tops so difficult to short. The critical asymmetry for position sizing is this: a Managed Money cohort at the 90th percentile has roughly 10% of its historical range left to absorb good news, while it has 90% of that range available to unwind on bad news. Last week’s data showed gold at the 88th percentile and this analysis flagged the historical base rate of an 8% to 15% drawdown from such levels; the market has since added a further two percentile points and a 7.32% price gain, which improves the near-term momentum and simultaneously worsens the asymmetry. The InProved Terminal’s own framing on the accompanying post was appropriately direct: positioning was already at 52-week extremes on both sides before Warsh’s speech.
The InProved Terminal’s all-metals open interest monitor records total COMEX precious metals open interest at 623,902 contracts as of the August 29 print, an increase of 19,030 contracts (+3.1%) on the week and a continuation of the steady position rebuilding that has run since the spring lows. The percentile framing on that figure is where the analytical value sits: 51% on a 52-week basis but only 17% on a three-year basis. Those two numbers must be read together. Relative to the past twelve months — a period that contained the severe deleveraging of the first half of 2026 — the complex has recovered to roughly the midpoint of its range. Relative to the past three years, which encompasses the far higher engagement levels of 2024 and 2025, the market remains in the bottom fifth of its distribution. The five-year chart makes the shape of this unmistakable: total open interest oscillated between roughly 700,000 and 880,000 contracts through 2022, 2023, and 2024, peaked near 880,000 in early 2025, then collapsed through the back half of that year and the first quarter of 2026 to a trough near 480,000 before turning higher into the current 623,902 reading. Gold remains the dominant weight driving the rebound, visible as the expanding orange band at the base of the stacked area chart.
The tension between a 17th-percentile open interest base and 90th-percentile Managed Money positioning is the single most important structural observation available in this week’s data, and it explains a great deal about gold’s recent price behaviour. Open interest measures the total size of the market; positioning percentiles measure how the participants within that market are distributed. When positioning reaches extremes on a shrunken open interest base, each contract of incremental flow moves price further than it would in a deeper market, because there is simply less liquidity available to absorb it. This is the mechanical reason gold has been able to produce a 7.32% weekly move: not because the absolute quantity of new buying was historically large — 65.8 tonnes of speculative addition is meaningful but not extraordinary — but because it entered a market whose total open interest sits at the 17th percentile of its three-year range. The same arithmetic applies in reverse, and this is the part that deserves attention from anyone holding leveraged length. A liquidation of speculative positioning into a market this thin would produce price displacement materially larger than the historical base rates derived from deeper markets would suggest. The constructive reading of the same data is that a 17th-percentile open interest base leaves enormous room for the complex to attract new capital, and that the rebuilding trend visible since the spring has been steady rather than spiky — but the rebuilding has not yet gone far enough to provide the depth that would cushion a positioning unwind.
The InProved Terminal’s China Premium Monitor for gold, updated August 28, records the Shanghai Gold Exchange at $4,606.08 against the LBMA benchmark of $4,607.10 — a discount of $1.02 per ounce, or -0.02%. The five-day average sits at -0.10% and the thirty-day range spans -0.23% to +0.27%. Read against last week’s print of -0.18% with a thirty-day range of -0.18% to +0.27%, the picture is one of stabilisation just below parity rather than recovery: the discount has narrowed from its six-month extreme but the thirty-day low has actually deepened to -0.23%, and at no point in the trailing month has the premium reached even +0.30%. The three-month trend chart shows the structural shift with unusual clarity — the series traded consistently in positive territory between +0.10% and +0.30% through June and early July, began oscillating around zero in late July, and has spent August predominantly below the zero line with the sharpest negative excursions occurring in the final ten days. This is the second consecutive week in which the world’s largest physical gold consumer has declined to pay the international clearing price for domestically delivered metal.
The significance of a sustained Chinese discount is best understood by asking what it would take to reverse it, and the answer explains why this matters more than the small magnitude of the number suggests. A negative SGE premium means that arbitrage now runs outward: it is profitable to move gold from Chinese vaults toward London and other international venues rather than inward. Reversing that requires either a fall in the international price, a rise in Chinese domestic demand, or a depletion of the Chinese vault inventory that currently satisfies domestic demand without imports — and last week’s SHFE data showed those vaults parked at a record 113.658 tonnes. None of those three conditions is currently developing. The consequence for the gold market is that the entirety of the advance from $4,334 through $4,652 over the past fortnight has been financed by Western financial demand, precisely the Managed Money and Other Reportables flow the COT data documents, with no physical Asian absorption underneath it. Through 2024 and the first half of 2025, Chinese premiums in the +0.5% to +2.0% range meant every dollar of Western paper buying was validated by a physical bid that pulled real metal eastward and tightened the free float available to London. That mechanism is currently switched off. A rally that is entirely paper-driven is not inherently fragile — many sustained bull markets have been financed this way — but it removes the buyer of last resort that has historically arrested gold’s corrections, and it does so at exactly the moment the COT data shows both sides of the futures market at 52-week extremes.
For bullion dealers, a gold price of $4,652 with Managed Money at the 90th percentile and Swap Dealers at the 96th is a set of conditions that argues strongly for selling into strength rather than accumulating through it. The retail flow will arrive — a headline price this close to record territory reliably produces walk-in and online enquiry within days — and that flow should be met from existing inventory at expanded premiums rather than from newly acquired stock purchased at these levels. Fabrication premiums on sovereign coin typically widen 8% to 15% during advances of this velocity, and the disciplined commercial play is to capture that widening on metal bought at $4,300 rather than to bid for replacement inventory into a market where the professional hedging community has taken its short position to the 96th percentile of a five-year range. Dealers carrying meaningful unhedged physical length above $4,500 should consider laying off a portion through COMEX shorts or forward sales; the cost of that hedge is small relative to the drawdown risk the positioning data implies. On the buyback side, expect the phone to ring in the opposite direction if the market breaks — customers who buy at 90th-percentile positioning are the same customers who liquidate at the 30th, and holding cash capacity to make two-way markets through a correction is worth more than an extra tranche of inventory right now. A working planning range of $4,450 to $4,850 through the fourth quarter, with an explicit recognition that a positioning unwind toward $4,250 is a live scenario, is the honest basis for stocking decisions.
For conservative investors, the correct action this week is almost certainly no action. Core allocations built through the corrective phase of the first half of 2026 are now substantially profitable — gold’s year-on-year return of +38.42% and year-to-date return of +7.88% confirm that the secular thesis has been working — and there is nothing in this week’s data that argues for reducing a long-term position. Central bank accumulation, reserve diversification, and the fiscal trajectory that underpins the debasement argument are all intact and operate on horizons measured in years rather than COT reporting weeks. What this week’s data does argue against, and argues against emphatically, is deploying new capital into gold at current levels. An investor buying at the 90th percentile of speculative positioning, into a market whose open interest sits at the 17th percentile of its three-year range, with the Chinese physical bid absent for a second consecutive week, is accepting close to the worst entry the year has offered on a risk-adjusted basis. The disciplined path is to hold the core untouched, defer additions until the Managed Money percentile retreats below 65 or the China premium recovers into positive territory, and treat a pullback into the $4,300 to $4,400 zone as the accumulation opportunity that the positioning data suggests is more likely than not to appear. A twelve-month target of $5,000 to $5,200 remains achievable and would still sit below the 52-week high of $5,415.17, but the route there is unlikely to be direct from a double-extreme starting point.
For active traders, this is a week to reduce size rather than to express conviction. The double extreme — Managed Money 90th, Swap Dealers 96th — is a configuration with a genuinely bimodal distribution of outcomes: either the last marginal buyers are absorbed over the next one to four weeks and gold pushes toward $4,800 to $4,900 in a final thrust, or the flow stops and Swap Dealer short-covering triggers a violent two-to-three-session reversal. Both are plausible; neither is predictable from the positioning data alone. What the data does support is a clear structure for whichever path emerges. On the upside, $4,700 is the immediate objective and a sustained daily close above it on expanding open interest would project $4,850. On the downside, $4,500 is the first line and a weekly close below $4,400 would confirm the unwind, with the thin 17th-percentile open interest base implying that the move from there could overshoot toward $4,250 faster than conventional volatility estimates would suggest. Because the open interest base is so thin, stops should be placed wider than usual and position size reduced correspondingly — the standard error on any price projection is materially larger in a market at the 17th percentile of its three-year depth. The superior expression of a bullish precious metals view at these levels is not long gold but the long-silver-versus-gold ratio trade, which the silver COT data supports and which carries none of gold’s positioning-extreme risk.
Hugo Pascal’s observation about the AU9999 contract hitting a 10-week volume high underscores the increasing significance of physical gold trading on the Shanghai Gold Exchange. This trend not only highlights robust domestic demand in China but also reflects broader shifts in the global gold market toward physical-backed assets.
Latest articles
Tool and strategies modern teams need to help their companies grow.
Invite users to stay updated with exclusive insights and market trends by subscribing to the newsletter.
Start stacking your bullion. Download the InProved app and build your legacy in gold and silver.
+65 8899 3720
support@inproved.com