Why Gold Prices Are Falling While Demand Hits Record Levels

Gold and silver just posted their third consecutive weekly loss, and yet the money flowing into these markets is arriving at a near-record pace. That is the contradiction worth sitting with: prices falling while institutional and sovereign buyers keep accumulating.

The timing sharpens the stakes. The Federal Reserve’s next rate decision lands on 16 September 2026, four days out, with crude oil trading above $100 per barrel and rising real yields doing the heavy lifting on the downside. The macro pressure is real, and so is the demand signal, which means one of the two is likely to give way once the Fed shows its hand.

This analysis breaks down why prices are sliding, why the largest buyers have not blinked, how the current pullback stacks up against past corrections, and what the next few days are positioned to settle. The read you should take is less about the weekly chart and more about the conditions that would reverse it.

Three weeks of losses, and the buyers are still there

Start with the coordinates. As of 11-12 September 2026, spot gold traded between $4,348 and $4,409 per ounce, with silver in the $64.48 to $65 per ounce range. The gold-to-silver ratio sat near 67 to 67.6:1, roughly in line with where it has held through the recent weakness.

The correction is not trivial. Gold is down approximately 16% from its January 2026 high of around $5,589 per ounce, according to Crux Investor. Key technical support now sits near $4,300 per ounce, with resistance stacked between $4,400 and $4,500 per ounce, per SD Bullion’s James Anderson.

Now the number pointing the other way. In August, physically backed gold ETFs pulled in a striking sum while the price kept sliding.

$18 billion / 121 tonnes August 2026 net inflows into global gold ETFs, the second-largest monthly inflow on record in value terms, per World Gold Council data. Total holdings hit a record 4,189 tonnes, with assets under management up 16% month-on-month to $615 billion.

Hold both figures in view at once. A 16% correction and near-record institutional accumulation are pointing in opposite directions on the same asset in the same week. That tells you this is not a straightforward momentum story, and it is worth resisting the urge to read the falling chart as evidence of collapsing demand. The price and the flows are telling two different stories, and the rest of this analysis is about which one dominates, and when.

The August 2026 figures arrive on top of an already-exceptional base: the broader historic ETF inflow cycle of 2025 established the institutional allocation infrastructure that allowed August’s $18 billion to clear without price support, because the buying was absorbed into a holdings base already built for sustained accumulation rather than tactical trades.

What is actually suppressing prices while demand stays elevated

If demand is this strong, why are prices falling? The answer runs through two forces that non-yielding assets cannot escape: rising U.S. real yields and a stronger dollar.

The mechanism is straightforward. Gold pays no income, so when the real return on Treasuries climbs, the opportunity cost of holding metal rises with it. A firmer dollar compounds the drag by making dollar-priced gold more expensive for international buyers. WisdomTree’s July 2026 “Gold Monthly” framed the correction exactly this way, arguing that rising real yields, a stronger dollar, and higher Fed rate expectations have outweighed safe-haven demand in the near term.

The relationship between real interest rates and gold is more precise than the headline correlation suggests: it operates through the opportunity cost channel, where each incremental move in inflation-adjusted yields reprices the carry cost of holding a non-yielding asset against the entire sovereign curve.

The institutional chorus is consistent. ING cut its Q3 and Q4 2026 gold forecasts, reasoning that Treasury yields and the dollar are overpowering investor demand. Kitco, writing on 31 August 2026, stressed the same hawkish Fed and real-yield pressure as the dominant short-term headwinds.

Here are the three forces working against price right now:

• Rising real yields: Higher inflation-adjusted Treasury returns raise the opportunity cost of holding a non-yielding asset, pulling capital away from gold.
• Dollar strength: A firmer dollar makes dollar-denominated metal costlier for overseas buyers, softening international demand.
• Hawkish rate expectations: The market’s pricing of a possible Fed hike keeps upward pressure on yields and the dollar before the decision even arrives.

Now the piece that resolves the contradiction. There are two different markets operating here. Futures-market participants are deleveraging under margin pressure, and that selling drives the spot price lower. Strategic ETF buyers, meanwhile, keep accumulating regardless. Ainvest’s February 2026 flow analysis attributed the divergence to short-term liquidity, margin hikes, and dollar strength driving futures selling even as strategic investors continued allocating through ETFs.

That split is the mechanism to keep in mind. It explains how prices fall and inflows rise simultaneously, and it tells you the spot price alone is an incomplete picture of what large-scale investors are actually doing.

The September 16 rate decision as a binary inflection point

The Fed funds target has sat at 3.50%-3.75% since 18 June 2026. What happens on 16 September will move real yields and the dollar directly, which is why metals traders are watching it as the near-term pivot.

The probabilities show genuine disagreement. CME FedWatch put a 56-58% probability on a 25-basis-point hike to 3.75-4.00% in early September, down from 65.9-66% on 31 August. Yet a Reuters economist poll conducted 4-9 September found roughly 70% of surveyed economists expecting no change at all.

That gap between futures pricing and economist consensus is itself a signal. Informed observers are split, which makes this a live decision rather than a formality. A hike would reinforce the dollar and real-yield pressure that has been suppressing prices; a hold would ease both. Complicating the Fed’s call, crude above $100 per barrel keeps inflation in the picture, giving the hawkish case a foothold it might otherwise lack.

China’s central bank and the structural demand floor

While futures traders unwind positions week to week, one buyer is operating on an entirely different clock. China’s central bank has now bought gold for 22 straight months.

22 consecutive months The People’s Bank of China extended its gold-buying streak through August 2026, adding approximately 20.22 metric tonnes (650,000 ounces) during the month, the largest single-month addition since purchases resumed in November 2024, per PBOC data released 8 September 2026.

That August addition lifted PBOC reserves to 76.73 million ounces, up from around 76.08 million ounces in July, when the streak stood at 21 months. This is not a buyer reacting to the weekly chart. It is a multi-year reserve-management decision that has continued straight through the 16% correction.

The broader central bank picture reinforces the point. Union Bancaire Privée projects central banks will add roughly 800 tonnes to global reserves across 2026, though UBP flags this as a forecast subject to policy and market risks, not a settled outcome.

Central bank gold buying in 2026 has been driven by reserve diversification logic that sits entirely outside the yield-and-dollar framework governing futures markets, with purchasing decisions anchored to currency exposure management and geopolitical hedging rather than near-term return expectations.

It helps to separate the three demand categories operating at once, because they behave very differently:

• Sovereign central banks (PBOC and peers): Multi-year reserve diversification, structurally insensitive to short-term price swings.
• ETF strategic allocation: Medium-term portfolio positioning, tactical enough to add on dips but slower-moving than futures.
• Tactical futures and physical buying: Short-horizon, leverage-sensitive, and the source of the current spot-price selling.

An unbroken 22-month streak at accelerating monthly volumes tells you at least one class of large buyer is treating this correction as irrelevant to its thesis. That is a materially firmer signal than retail dip-buying, and it is why the demand floor beneath gold looks more durable than futures positioning alone would suggest.

How this correction compares to past episodes, and what changes the trajectory

None of this is unprecedented. GoldSilver.com characterised the recent stretch as gold’s worst quarter since 2013, and the parallel is instructive.

In 2013, Fed policy expectations shifted sharply as the taper debate took hold. Gold underwent a steep, multi-month decline before stabilising once real-yield pressures eased. The driver then was the same as now: a repricing of monetary policy, not a breakdown in physical demand.

The more recent precedent is closer to home. In March 2026, global ETFs shed $12 billion (84 tonnes), partly offset by $1.9 billion (10 tonnes) of Asian inflows, according to the World Gold Council. That selling was driven by momentum factors and COMEX long unwinds rather than any demand collapse, and prices recovered once the deleveraging pressure abated. A further 16 tonnes left ETFs in May 2026 under the same dynamic.

ETF gold flows in April 2026 turned positive after the March redemption episode, a reversal that followed the pattern the current correction is now replicating: futures-driven selling creating a price dip, followed by strategic allocation buying as the macro picture clarified.

The pattern is recognisable, which reframes the question. It is not whether gold can recover from a policy-driven correction; history says it can. The real question is whether the specific conditions that ended prior corrections are forming now.

Analysts are specific about those conditions. Lombard Odier, writing on 19 May 2026, framed the current consolidation as similar to past rate-yield pauses, with recovery contingent on real yields declining and the dollar weakening. UBP echoed the framing on 29 May 2026, treating the weakness as a pause in a broader uptrend supported by central bank buying. Both firms’ bullish targets carry conditions: UBS holds a $5,500 per ounce year-end target contingent on rate cuts materialising, while Morgan Stanley’s $5,200 per ounce H2 2026 target depends explicitly on a meaningful ETF rebound.

The three variables to watch as the FOMC week unfolds

The 16 September decision produces three observable outputs that will most directly influence whether the correction extends or reverses:

• Real yield direction: A decline in inflation-adjusted Treasury yields is constructive for metals; a further climb extends the pressure.
• Dollar index movement: Dollar weakness supports international demand and metal prices; continued strength keeps the headwind in place.
• The Fed’s updated dot plot: A projection signalling fewer future hikes is constructive; a hawkish revision pointing to higher-for-longer rates is destructive for the recovery case.

Where the risk-reward sits ahead of a live FOMC decision

Put the two sides together honestly, because both are grounded in real data rather than speculation.

The structural case is intact. Central banks have bought for 22 straight months, ETF holdings sit at a record 4,189 tonnes, and the historical pattern shows policy-driven corrections tend to resolve once macro conditions shift. None of that has weakened during the pullback.

But the cyclical forces are winning in the short run, and only the Fed can change that. The bearish constraints are credible, not fringe. Standing on each side:

The structural case:

• Central bank buying continues unbroken at 22 months and accelerating monthly volumes.
• ETF holdings reached a record 4,189 tonnes in August despite falling prices.
• Historical corrections of this type have recovered once real yields eased.
• Longer-term diversification demand remains, per UBS and UBP framing.

The cyclical headwinds:

• Rising real yields raise the opportunity cost of holding gold.
• Dollar strength suppresses international demand.
• Hawkish rate expectations keep upward pressure on yields.
• Loss-making ETF holdings pose a potential supply overhang if momentum stays negative.

On that last point, Standard Chartered’s Suki Cooper has flagged a large block of ETF holdings in loss-making territory as a structural supply ceiling that could fuel further redemptions. The specific tonnage figure is unverified in secondary sources, so treat it as a risk to note rather than a number to anchor on. ING’s forecast cuts and Morgan Stanley’s ETF-rebound dependency sit alongside it as genuine bearish constraints.

“Current weakness resembles past episodes where gold paused amid rising real yields and a strong dollar, but advanced again when real yields declined and the dollar depreciated.” Lombard Odier, 19 May 2026

The FOMC outcome is the pivot. A hold reduces the headwind; a hike extends it, with $4,300 per ounce as the first technical test if pressure continues. Both outcomes are plausible, which shifts the reader’s question from “should I exit” to “what am I waiting for, and what would change my view.”

This article is for informational purposes only and should not be considered financial advice. Investors should conduct their own research and consult with financial professionals before making investment decisions. Past performance does not guarantee future results, and financial projections are subject to market conditions and various risk factors.

What the next few days settle, and what they do not

Whatever the Fed does on 16 September, it is worth being clear about the limits of what a single decision resolves.

A hold would remove the most proximate headwind. It would not, on its own, reverse the 16% correction from January or dismantle the yield and dollar dynamics that built up over months. It buys relief, not a reversal.

A hike would extend the cyclical pressure and put $4,300 per ounce support squarely in play. But it would not negate the 22-month central bank buying streak, the record 4,189 tonnes in ETFs, or the longer-term structural case that has held through the entire pullback.

That is the dual frame to carry into the week. The near-term path runs through a binary policy decision where the bearish outcome remains the more likely futures-priced scenario. The longer-term anchor is a set of demand signals that have not flinched.

UBS’s $5,500 per ounce year-end target captures the directional case, but it comes with a condition attached: it depends on the rate path turning favourable. The forecast is only as good as the macro shift it assumes.

So the honest close is not a prediction on the Fed. It is a framework. On 16 September you will know more about the near-term direction of real yields and the dollar. What you will not know, regardless of the outcome, is whether the structural buyers are wrong, because they are operating on a horizon this decision does not reach.