Four weeks down, and the gold bid finally cracked

For three weeks I kept telling dealers on the Bay Area floor that the gold pullback looked orderly, the kind of profit-taking that resets a stretched market without breaking it. This week it stopped looking orderly. Spot gold spent Friday clawing back 1.2% to settle near $4,073 an ounce, yet that bounce only softened what was otherwise the fourth straight weekly decline. Earlier in the week the metal printed near $4,000, and at the lows it had handed back essentially all of its 2026 gain. A position opened in January at a healthy premium is now flat on the year, and the desks that loaded kilo bars around $4,700 in May are wearing every dollar of the gap.

The cause has not changed since Kevin Warsh took the Fed chair. The market is pricing tighter policy, not looser. After last week's hawkish pause, futures put the odds of a December hike near 80%, with a September move running around 60%. Thursday's PCE print landed in line with expectations, enough to pull the dollar and Treasury yields back and let gold rally into the weekly close, but in line is not soft, and the bid that carried bullion through the spring is gone. The one constant under the price is official buying. The People's Bank of China extended its accumulation to a nineteenth straight month with a 9.9 tonne addition in May, lifting reserves to 2,313 tonnes, while Poland added 31 tonnes in the first quarter and Uzbekistan 25. Central banks took in an estimated 244 tonnes in the quarter and the World Gold Council still sees 700 to 900 tonnes for the year. That tells you the metal is not in trouble. The leveraged speculative bid is. I work through both sides in this week's gold breakdown.

On the physical side the price drop is a mixed blessing. A lower spot eases the sticker shock that has pushed jewelry buyers toward 10-karat and hollow product all year, but it also pulls scrap back off the bench as sellers wait for a rebound. Refiners I talk to describe kilo bar premiums as steady and scrap flow as thin, the usual pattern when the market is falling but nobody believes the bottom is in.

Watches: the listings glut nobody asked for

The secondary watch market told a parallel story. The WatchCharts Overall Market Index lost 1.4% in May, and Rolex was the weakest major at down 2.2%, with the GMT-Master off 2.4%, the Daytona off 2.2%, and the Yacht-Master down 2.0%. Patek Philippe slipped only 0.1% and Audemars Piguet 0.3%, ending a months-long rally for both but giving back almost nothing. What troubles the trade more than the price prints is supply. April and May brought record-high counts of new listings, the clearest sign yet of a sell-off of inventory that speculators bought on the way up in 2022 and 2023.

The retail side moved the other way. Rolex raised list prices on June 1, with gold models up an average of 5.0% and two-tone up 2.5%, widening the gap between boutique and secondary that I detail in the watch column. The discontinued steel Pepsi, reference 126710BLRO, still trades near $22,500 after a 12% gain in the first quarter, proof that the right reference with a real waitlist behind it can ignore the broader slide. Swiss export data backs the split. May shipments grew 0.4% to 2.1 billion francs after a brutal April, leaving the trade flat rather than falling. Selectivity, not breadth, is carrying this market.

Product kept the top of the watch market engaged even as the speculative tier cleared. Audemars Piguet used June to add three 37mm Royal Oak Offshore Selfwinding Chronograph references, a lighter read on a line built on aggression, while Vacheron Constantin showed an Overseas Dual Time Cardinal Points in titanium. These are allocation-driven releases that never touch the secondary glut, because the clients buying them were never going to flip them. That is the shape of the market in one line: the brands with real waitlists keep selling at list while the speculators mark down what they overpaid for two years ago. The gap between those two tiers is the single most useful thing a dealer can watch right now, and it is widening.

Diamonds: India takes the number two seat

The diamond story this week was structural rather than a daily price tick. De Beers data confirmed India has become the world's second-largest diamond jewellery market at 12% of global demand, up from 10% in 2019, trailing only the United States at 53%. China and Japan have each slipped to 5%. India's natural diamond jewellery market sits near 49,700 crore rupees and is tracking toward 1.5 lakh crore by 2030, with Gen Z buyers and self-purchase driving the gain. A fresh India-New Zealand free trade deal layers on zero-duty access that the GJEPC expects to lift gem and jewellery shipments to that market from 16.6 million dollars toward 50 million in three years.

Underneath the demand shift, pricing is still healing. Lab-grown now accounts for roughly 45% of United States engagement-ring purchases and sells for about 73% less than a comparable natural stone, a gap wide enough that the two products no longer compete for the same buyer. The synthetic stone has reached a production-cost floor near 100 to 150 dollars a carat for cutting and polishing, while the natural stone holds a 5 to 15 times premium at identical specifications even after the RapNet Diamond Index fell 11.3% across 2025. De Beers realized 101 dollars a carat on rough in the first quarter, down 19%, a number that explains why Anglo American wants out. The full structural read sits in the jewelry report.

Industry: Richemont keeps cutting watch weight

Richemont closed its financial year to March 31 with group sales of 22.4 billion euros, up 5% in reported euros and 11% at constant rates, and operating profit of 4.5 billion. The jewellery maisons, Cartier, Van Cleef and Arpels, Buccellati and Vhernier, produced 16.5 billion of that, roughly three-quarters of revenue, at a 30.5% operating margin. The Specialist Watchmakers division did 3.1 billion, down 4% at actual rates, with sequential improvement at A. Lange and Sohne, Jaeger-LeCoultre and Vacheron Constantin in the second half.

Against that backdrop the group is offloading loss-making Baume and Mercier to Italy's Damiani Group, a deal signed in January and closing this summer, with Richemont providing operational support for at least a year after the handoff. The message to the trade is plain. Jewellery pays, watches are being rationalized to the maisons that earn their keep, and a brand that lost money gets sold to a specialist who thinks it can do better.

The read into next week

Pull the four threads together and one theme runs through all of them. Capital is rotating out of the speculative end of hard assets and into the parts of the trade with genuine end demand. Gold's bid broke because the rate story turned, not because anyone stopped wanting metal, and the central banks buying through the dip prove the point. The watch listings glut is the unwind of the 2022 bulge, while Patek and AP, the names with real waitlists, barely moved. Diamonds found their growth at India's retail counters, not in rough speculation. Richemont is pruning a watch brand to feed jewellery counters that throw off a 30% margin. For dealers the lesson is the one that never changes in a soft tape: the references with a buyer behind them hold, and the inventory bought to flip is the inventory now setting the marks. Next week brings the June PCE follow-through and the first honest test of whether gold can hold $4,000 with a hawkish Fed in the chair. That level, not any auction headline, is the number the whole trade is watching.