Goldman’s $4,900 Gold Call Meets a China Buying Gap—and a Fed Risk

Goldman Sachs’ case for higher gold prices rests on sustained central-bank purchases, including buying its analysts believe may not appear promptly in published reserve figures. The numbers warrant a closer look—but so does the forecast itself: reports published in September give different impressions of whether $4,900 per troy ounce remains Goldman’s operative end-2026 figure.7209

Is $4,900 still Goldman’s year-end forecast?

Goldman published a $4,900 end-2026 price forecast in August, citing central banks’ efforts to diversify their reserves. A September 23 report still described $4,900 as its year-end target. Other September reports, however, say Goldman lowered its end-2026 fair-value estimate from $4,900 to $4,650, while retaining a $5,400 target for the end of 2027. One report describes the updated 2026 fair-value outlook as a $4,650–$4,900 range. The available accounts do not establish that $4,900 remains Goldman’s single, current central estimate for year-end 2026.7209102

July’s gold-buying figures: estimate versus reported reserves

Goldman’s nowcast put July 2026 central-bank purchases at 44 tonnes; World Gold Council figures put reported net buying at 23 tonnes. Goldman estimated 35 tonnes of buying by China, while China’s published reserves rose by about 20 tonnes. Reuters reported that the People’s Bank of China’s holdings increased from 75.44 million to 76.08 million fine troy ounces that month.203027

China’s estimated 35 tonnes is 15 tonnes, or 75%, above its reported addition. It is an estimate of total buying, not 35 tonnes of hidden purchases on top of the reported 20. Nor can the 21-tonne difference between Goldman’s global estimate and reported net additions simply be labeled unreported buying: the figures come from different approaches, and the published series measures reported changes in reserves.20274030

What points to buying beyond official disclosures?

The strongest indication in the supplied reporting is Goldman’s estimate itself, which exceeds the reported figures for both China and central banks overall. An account of its approach describes using London trading signals to infer purchases that may not yet appear in reserve disclosures. That offers a reason to investigate a reporting gap, not independent confirmation of who bought the gold or how much remains undisclosed.20322730

The broader demand thesis does not depend on proving every July purchase was hidden. Goldman said in August that it expected central banks to buy an average of 50 tonnes a month in 2026, compared with an average of 17 tonnes a month before 2022. Those are Goldman’s estimates and forecast, rather than official additions to reserves.7

How the Fed, ETFs and options could change the outcome

Goldman’s outlook has a rate-sensitive side as well as a central-bank-demand side. A more hawkish Federal Reserve can weigh on gold-backed ETF demand and on the appeal of holding non-yielding gold; reports of Goldman’s lower $4,650 fair-value estimate reflect that near-term risk. An earlier account of Goldman’s downside scenario put gold at roughly $4,400–$4,440 by year-end if Fed hikes prompted investors to unwind policy hedges. That was a conditional scenario, not a new base-case forecast.9101113

Options trading could make either outcome more volatile. If demand for gold call options rises as prices climb, dealers hedging those contracts may need to buy more exposure, potentially adding momentum to a rally. If prices fall or positions unwind, hedge adjustments could instead intensify a decline. Strong central-bank demand may support Goldman’s longer-term view, but it does not remove the risks posed by rates, ETF flows or positioning.482