Gold Is at $4,088. Central Banks Are Buying Faster Than Mines Can Dig It.
Gold rebounded to $4,088 an ounce as Poland added 82 tonnes in the first half and China extended its buying streak to 20 straight months. Central bank purchases are running at a third of annual mine supply, and the GLD ETF sits dead-neutral — the structural bid is sovereign, not speculative. The floor at $4,000 is holding because the buyers are not price-sensitive.
$4,088 an ounce. That is where gold rebounded to on July 14, clawing back above the $4,000 level after a sharp sell-off that had threatened to break the floor. The rally is not being driven by retail fear or ETF inflows or a weak dollar — though the dollar index did fall 0.55 percent after June CPI trailed estimates, which helped. The rally is being driven by central banks, and they are buying at a pace that the mining industry physically cannot match.
Poland added 82 tonnes of gold in the first half of 2026. China increased its reserves for the 20th consecutive month. These are not speculative positions. They are reserve diversification — sovereign decisions to hold less dollars and more metal, made by finance ministries that think in decades, not quarters. The structural bid from central banks is the single most important variable in the gold market right now, and it is the reason $4,000 is holding as a floor instead of acting as a ceiling (Newscase, July 14; Barchart, July 14).
The supply math is straightforward and ugly for anyone hoping for a pullback. Global mine production runs at roughly 3,600 tonnes per year. Central bank purchases in 2025 totaled approximately 1,040 tonnes — about 29 percent of annual mine supply. If Poland's 82-tonne H1 pace is representative of the broader trend, 2026 central bank buying could exceed 1,200 tonnes. That is a third of all newly mined gold going into sovereign vaults, never to be sold. The marginal buyer in the gold market is no longer a hedge fund or an ETF — it is a central bank with a printed mandate to reduce dollar exposure. Mine supply cannot expand fast enough to absorb that demand. New mines take 10 to 15 years from discovery to production. Poland can buy 82 tonnes with a wire transfer.
The GLD ETF — the most liquid gold proxy for US investors — tells a quieter story. RSI at 47.1 is dead neutral. Support sits at $361.30, the Bollinger lower band, then $370.50 at the 20-day EMA. The pivot is $378.00, the 20-day SMA. Resistance is $394.60, the Bollinger upper band. ATR is $20.45, meaning daily moves of 5 percent are routine. The ETF is range-bound — sitting mid-range, no momentum edge, chop rather than trend. That is the technical confirmation that the central bank bid is structural, not speculative. If hedge funds were driving this, GLD would be at the top of its bands. It isn't. The central banks are buying physical, not paper, and the paper is sitting still.
The dollar is the other variable, and it is cooperating. The DXY at 120.50 is down 0.21 percent on the latest reading, and the June CPI print at 3.46 percent — below consensus — gave the dollar an additional nudge lower. A weaker dollar makes gold cheaper for non-US central banks, which reinforces the buying cycle. If the July CPI print reverses the disinflation — which it likely will, given the nearly 15 percent oil spike this week — the dollar could strengthen and put short-term pressure on the gold price. But the central bank bid is not cyclical. Poland is not checking the CPI calendar before it wires money to buy bullion. China is not waiting for a Fed pivot. The sovereign demand is structural, and it operates on a timeline that renders monthly data noise irrelevant.
The Fed's rate path is the tactical overlay. Fed funds at 3.62 percent, 10-year Treasury at 4.56 percent, 2-year at 4.21 percent — real rates are positive, which is historically a headwind for gold. But gold is at $4,088 with real rates positive, which tells you the central bank bid is powerful enough to override the rate channel. That is new. For most of the past decade, gold and real yields moved in opposite directions. Now they don't. The central banks have decoupled gold from the rate equation, and until sovereign buying slows — which requires a geopolitical shift, not a data point — the floor holds.
The question is not whether gold goes higher. It is whether the mining sector can produce enough metal to satisfy buyers who are not price-sensitive. The answer, for the next several years, is no. Gold at $4,088 with central banks adding 82 tonnes in six months is not a speculative peak. It is a structural floor.