27.7% of Gold Positioning Is Options Again (18 August 2026): The Dealer-Hedging Amplifier That Cuts Both Ways

Delta-equivalent options made up 155,557 of the 561,817 contracts of combined open interest in COMEX gold as of 18 August 2026 — 27.7%, up from 22.1% three weeks earlier, on a straight subtraction of two CFTC series. That is the arithmetic behind Goldman Sachs's warning on 21 August that gold call demand has created "a mechanical price amplifier to both the upside and downside". The amplifier is real and it is measurable. It is also, and this is the part the headlines drop, direction-neutral.

Gold spent the week to 21 August doing what the macro story said it should. Spot traded around $4,587 an ounce by mid-morning in New York on the Friday, up about 2.6% on the day, per Fortune's quote; December futures reached $4,661.70, the highest since mid-May, on Forbes's account. The week added close to 5% and was the third consecutive weekly gain. The dollar sat near a three-month low. Treasury had surprised the market by increasing its planned purchases of longer-dated debt, a fiscal signal this site examined in the 30-year round trip.

None of that is the subject here. The question is not why gold rose but why it rose so fast, and the answer is in a market-structure series most gold coverage never opens.

What the CFTC data actually shows

The Commitments of Traders report is published in two versions of the same week. One counts futures only. The other, headed "Commitments of Traders with Delta-adjusted Options and Futures Combined", converts every option position into its futures equivalent using its delta and adds it to the futures total. Subtract the first from the second and you have isolated the option book, expressed in the same units as everything else.

Do that for COMEX gold on 18 August 2026 and you get 561,817 minus 406,260, or 155,557 contracts. At 100 troy ounces a contract that is about 15.6 million ounces of delta-equivalent exposure, and 27.7% of the whole market's combined open interest.

The level matters less than the trajectory. Run the same subtraction across every weekly report of 2026 and the series has a clear shape.

Two things stand out. The first is the speed of the rebuild: 42.8% growth in the option layer in three weeks, which is a fair reading of Goldman's "risen sharply". The second is the small decline in the final week — the overlay shrank by 8,719 contracts between 11 and 18 August while futures-only open interest rose by 5,951. Some of the option exposure was being converted into, or replaced by, outright futures. That is a normal thing to see when a rally becomes a consensus, and it is a reminder that the overlay is not a one-way accumulation.

Outright speculative positioning tells a complementary story. Non-commercial traders were net long 219,495 combined contracts on 18 August, against 174,131 on 28 July. The 2026 peak was 230,463 on 13 January. So speculative length is back within 5% of its high for the year while the price is still about 18% below January's record — which is another way of saying that a lot of people are positioned for a recovery that has not yet arrived.

How a dealer hedge becomes a price move

The mechanism deserves stating slowly, because it is routinely described as though someone were choosing to push the price around. Nobody is.

An institution that sells a gold call has taken on an obligation whose value moves with the gold price. The sensitivity of that obligation to a one-dollar move in gold is its delta. Delta is not constant: a call far below its strike barely responds to the gold price, and a call deep above it responds almost one-for-one. As the price climbs toward the strike, the delta of the short call grows, and the dealer's book — which it wants flat — becomes progressively short gold. To flatten it, the dealer buys gold. Price rises again, delta grows again, dealer buys again.

Three properties follow, and all three are worth holding onto.

It is pro-cyclical. The hedging flow always trades in the direction the price is already going. That is the amplification.

It is symmetric. The identical machinery that buys into strength sells into weakness. Goldman's note said so directly — a renewed increase in Fed-hike expectations "could likewise trigger dealer hedge unwinds and produce a sharper-than-usual correction". Any account of this dynamic that only mentions the upside has removed half of it.

It is concentrated at strikes. The hedging demand is not spread evenly across the price range; it clusters where the options are, and it is most intense where delta is changing fastest, which is around the strike near expiry. This is why option-driven moves tend to feel discontinuous rather than gradual.

The skew is where the composition shows up

The CFTC series has a hard limit: it reports the delta-adjusted option book as a single number and never separates calls from puts. It can confirm the size of the overlay. It cannot confirm that the demand is in calls.

For that, the evidence comes from the desks. Murphy told CNBC on 17 August that gold skew — the relative pricing of options at different strikes — "has shifted materially away from downside puts and toward upside calls, reversing the earlier summer setup when put protection was relatively richer; that shift has already shown up in recent flow". He cited a purchase of 8,000 November 460 calls on the SPDR Gold Trust at about $5.55, against an ETF close of $405.49 that day — a strike roughly 13% above the market, bought in size. He also flagged the other side of the same coin: about 25,000 September 350 puts bought at $0.62, taking advantage of downside protection that had become cheap. Gold funds, he added, were seeing their strongest inflows since January.

Read those two trades together and you have the whole regime in miniature. Upside is being bought with conviction and downside is being bought because it is on sale. Both add to the delta a dealer has to hedge.

Why February's 44.3% is the caution, not the endorsement

Here is the test that should settle whether a large option overlay is a bullish signal. In 2026 the overlay was at its largest, 44.3% of combined open interest, on 10 February. Gold then delivered its worst quarter in more than a decade, falling 16% in the three months to 30 June and sliding from a January record near $5,600 to a summer range of roughly $4,000 to $4,200, on Forbes's figures.

The amplifier worked perfectly. It simply amplified a decline. The reason it declined is the subject of a separate piece on this site — why gold fell in 2026 — and it has nothing to do with options: an energy shock lifted US inflation, inflation moved the Federal Reserve from cuts to hikes, and the real yield gold competes with rose across every maturity. Market structure decided how fast that repricing travelled. Macro decided which way.

That distinction is the practical use of this data. The option overlay is a gain setting, not a direction. Knowing it is high tells you to expect larger moves from a given surprise, in whichever direction the surprise points, and to discount the informational content of the move's size. It does not tell you what the surprise will be.

What has to be true this week for the amplifier to fire

Goldman named the trigger specifically: a renewed increase in Fed-hike expectations. This week's calendar is unusually dense with inputs to exactly that variable, and they are stacked into three days.

The Bureau of Economic Analysis schedule puts both "Personal Income and Outlays, July 2026" and "GDP (Second Estimate) and Corporate Profits, 2nd Quarter 2026" at 8:30 a.m. Eastern on Wednesday 26 August. The Cleveland Fed's nowcast, as of 21 August, had July core PCE at 0.25% on the month and 3.29% on the year — essentially unchanged from June and more than a point above the 2% objective. Then the Kansas City Fed's symposium runs 27 to 29 August, with the chair's keynote on the Friday morning, covered here in the Jackson Hole preview.

The link from those events to gold runs through the dollar and the real yield, not through the options market. Options only decide the magnitude. Forbes reported the futures market pricing a 70.9% probability of a rate increase at the December meeting as of 21 August — so hike expectations are not absent from the price; they are deferred. A print or a speech that pulls them forward is the specific input Goldman flagged as capable of forcing hedge unwinds.

What would change the picture

Three things, none of them a price level.

The first is the overlay itself. The report released on Friday 28 August covers positions as of Tuesday 25 August and is the first look at whether the option layer grew into this week or was trimmed before it. A larger overlay going into the events means a larger hedging response to whatever they produce.

The second is implied volatility. If one-month implied volatility rises materially from the lows Murphy described, the amplifier weakens at the margin — options get more expensive, less delta gets bought for the same premium, and the hedging load stops growing as quickly. Persistently cheap volatility against a rising price is the condition that keeps building it.

The third is the composition. Skew is observable daily. If it rotates back toward puts, as it was earlier in the summer, the asymmetry Goldman described reverses and the mechanical pressure sits on the downside instead. That would not make gold more likely to fall. It would make a fall, if one came, faster.

And a standing caution on all of it. Every figure in this piece is a Tuesday snapshot released three days later, describing a market that trades continuously. The 18 August data cannot tell you what the option book looked like on Friday, when most of the week's move happened. Positioning data explains mechanisms well and timing badly, and the distinction between the two is the difference between understanding a market and predicting one. If you are new to how the rate and dollar channels reach a commodity at all, the about page explains what this site measures and why.

Educational macro context only — not investment advice.