Gold traded at $4,053 an ounce by mid-morning Thursday, down $64 from the same hour Wednesday and roughly 2% below the prior session’s close near $4,138. had already broken $4,100 in European hours, printing $4,089.80 for a 0.99% decline before the New York open extended the losses. Futures tracked lower in parallel, quoted at $4,092.20 in pre-market, off 1.44%. The metal had reached a two-week high on Wednesday. That high lasted less than twenty-four hours.
did what silver does and fell harder. Spot traded at $58.43 by 9:58 a.m. Eastern, down from $59.83 Wednesday for a decline of roughly 1.67%. ticked up to 69.54 from 69.03, ending a brief stretch where the white metal was outrunning gold on the tape.
The trigger was not gold-specific. Iran-backed Houthi forces claimed missile and drone attacks on two Saudi oil tankers in the Red Sea, the US completed a twelfth consecutive night of strikes on Iranian targets, and crossed $100.05 a barrel, up 6.4% and above triple digits for the first time since late May. West Texas Intermediate advanced more than 5% to $91.08. Under the old playbook, that combination should have produced a violent bid for bullion.
Instead gold sold off, and the reason sits in the rates market. Higher crude feeds directly into headline inflation expectations, which pushes the Federal Reserve further from easing and closer to tightening. hit 4.695% on Wednesday, its highest level since January 2025. reached 4.334%, and is holding above 5%. Money markets now price roughly a 78% probability of a rate increase in September. For a non-yielding asset, that is the entire story.
Step back and the numbers get uncomfortable. Gold remains up $636 over twelve months, a gain of roughly 21.4%, and it has risen 2.26% over the past month. But it is down approximately 27.7% from its all-time high of $5,602.23 set on January 29, 2026. The 52-week range spans $3,268.15 to $5,595.46 — a spread of more than seventy percent that captures both a historic melt-up and a historic unwind inside a single year.
The second quarter of 2026 was gold’s worst in thirteen years. June alone took physical gold down 10.02%. The metal briefly printed below $4,000 during that stretch, and the recent lows matched levels last seen in November 2025.
The War Is Bullish for Crude and Bearish for Bullion, Which Only Sounds Wrong
The most instructive thing about gold’s 2026 is that it has fallen during an active war. That inverts every intuition built up over the past two decades, and understanding why is the key to positioning from here.
The mechanism runs through inflation rather than fear. When a geopolitical shock arrives through the energy channel, it lifts crude, which lifts headline CPI, which lifts breakevens, which lifts nominal yields. If the central bank responds by staying tight or tightening further, real yields — the inflation-adjusted return on government bonds — rise as well. Real yields are the cleanest single macro input for gold, because they represent the opportunity cost of holding an asset that pays nothing. When they climb, gold falls, regardless of how many headlines mention missiles.
The evidence from this cycle is unambiguous. During the March-to-June war period, gold underperformed the dollar against the rest of the G10 currency complex by roughly 2.6 percentage points. Bullion fell during an active conflict because the conflict’s inflation effect prevented the rate cuts that would have driven investment inflows. The war worked against gold’s most rate-sensitive buying channel while leaving the rate-insensitive official sector as the only steady buyer.
Thursday delivered a textbook example. Maritime monitors reported a vessel struck roughly 70 nautical miles southwest of Al Shuqaiq on Saudi Arabia’s Red Sea coast, with the Houthis identifying the targets as the tankers Encelia and Layla and framing the attacks as enforcement of a naval blockade. President Trump warned of strikes on Iranian infrastructure if shipping through the Strait of Hormuz comes under attack. Hormuz crossings had already collapsed to single digits daily. The Bab el-Mandeb strait, which handles between 12% and 15% of global maritime trade annually, is now the second chokepoint under threat.
That is as severe a supply-side escalation as the oil market has seen this cycle, and gold fell two percent on the day.
There is a nuance worth holding onto. A ceasefire would cut both ways. It would remove the safe-haven premium, which is bearish for gold, but it would also collapse the crude premium that is keeping the Fed hawkish, which is bullish. The net effect of peace on gold is genuinely ambiguous — which tells you how thoroughly the rates channel has come to dominate the geopolitical one.
Real Yields Have Been the Only Variable That Mattered All Year
The Federal Reserve currently holds rates at 3.50% to 3.75% under Chair Kevin Warsh, and the entire gold complex is trading off what happens to that range. Futures assign roughly an 85.6% probability that the July 28-29 meeting produces no change. September is where the risk sits: market-implied odds of a hike range from about 61% to 78% depending on the measure, and odds of any cut in 2026 have been priced out entirely.
Thursday’s labor data pushed the hawkish case further along. Initial jobless claims for the week ended July 18 printed 187,000 against a 212,000 consensus, following 208,000 the prior week. A labor market running that tight removes the growth argument that would otherwise force the committee to look through an energy-driven inflation impulse. Roughly half of FOMC officials have already penciled in a rate increase this year.
The bond market has front-run the decision. The 10-year at 4.695% is the highest since January 2025. The 2-year at 4.334% — the maturity that most closely tracks policy expectations — hit a multi-month high Thursday. The 30-year above 5% is the level that matters most for gold, because long-dated real yields set the discount rate against which a zero-coupon, non-yielding, storage-cost-bearing asset must compete. The 10-year TIPS real yield has been running near 2%, a level that creates meaningful and continuous drag.
That drag is why the standard commentary about gold as an inflation hedge has been so misleading this year. Gold hedges monetary debasement and negative real rates. It does not reliably hedge inflation when the central bank is willing to fight that inflation with higher rates. Those are different regimes, and 2026 has been firmly in the second.
The dollar reinforces the pressure. A firmer greenback makes bullion more expensive for buyers holding other currencies, suppressing physical demand at the margin in the largest consuming markets. The index has been broadly stable through this week rather than surging, which is the one small mercy in the setup — a genuine dollar breakout on top of current real yields would be considerably worse for the metal.
The calendar compresses everything into eight days. Flash manufacturing and services purchasing managers’ indexes land Friday. The FOMC decision arrives July 28-29. Between them sits the entire near-term distribution for gold.
The Round Trip From $5,602 Is a Bear Market, Not a Correction
Precision matters here because the framing determines the strategy. Gold’s all-time high of $5,602.23 was set on January 29, 2026. Silver’s all-time high of $121.67 was set the same day. Both metals topped on a single session, which is itself a tell — that was a liquidity-driven blow-off, not two independent fundamental peaks.
From that high, gold has fallen roughly 27.7% to $4,053. Silver has fallen approximately 52% to $58.43. A drawdown of that magnitude in the senior metal, sustained across two quarters, is a bear market by any conventional definition. The second quarter of 2026 was gold’s worst in thirteen years. June produced a 10.02% single-month decline. The metal briefly traded below $4,000 during that stretch, and recent lows have matched levels last seen in November 2025.
The counterpoint deserves equal weight and is frequently overlooked. Gold is still up 21.4% year over year and $636 an ounce higher than twelve months ago. It has risen 2.26% over the past month. The 52-week low of $3,268.15 sits roughly 19% below current spot. In other words, an investor who bought gold at any point before the fourth quarter of 2025 remains comfortably ahead. The damage is concentrated among those who chased the January parabola.
That distinction matters for how the current level should be read. A market that went from $3,268 to $5,602 and back to $4,053 has retraced roughly two-thirds of its advance. Retracements of that depth after vertical moves are historically normal rather than structurally damning — they are how markets digest a demand shock that outran the physical market.
The volatility on display is extraordinary, though. January’s record highs and June’s sub-$4,000 prints occurred within five months of each other. Silver went from $121.67 to below $60 across the same window. Position sizing in that environment matters more than directional conviction, and the conventional tool for managing it — building exposure incrementally rather than at a single price — has rarely been more applicable.
What has not happened is a break of the structural uptrend that began in 2024. Gold above $4,000 with central banks buying at record pace is not a broken market. It is an overbought market working off an excess.
Western ETF Redemptions Are the Mechanical Seller
The proximate driver of the price decline is identifiable and measurable: Western investors have been selling gold funds relentlessly, and those redemptions convert directly into physical sales.
March 2026 produced more than $12.7 billion in North American gold ETF outflows, the largest monthly figure in at least five years. US-listed gold ETFs posted roughly $5.3 billion in redemptions last month, following relatively balanced flows in April and May. Gold ETF investment demand fell approximately 65% in the first quarter to 62 tonnes as 10-year Treasury yields climbed. In June, Asian gold ETFs recorded their first monthly outflow since August 2025 — a signal that mattered enough for one major US investment bank to cite it explicitly when cutting its year-end target.
The contrast with 2025 is stark. Gold-backed ETFs drew a record $89 billion of inflows that year. Private investment demand nearly doubled from 2024 to almost 2,200 tonnes. A stablecoin issuer alone added more than 100 tonnes, exceeding any single central bank’s purchases. That was a demand shock without precedent in the modern era, and it produced a price to match.
What is instructive is that despite gold being roughly 80% higher than early 2025, global gold ETF holdings remain below the November 2020 peak of approximately 3,929 tonnes. The tonnage never fully returned even as the dollar value exploded. That has two readings. The bearish one is that flows can keep leaving because they never fully arrived. The bullish one is that Western institutional re-entry has not yet happened, and represents the next available demand leg if the rate environment turns.
The mechanics of ETF flows deserve emphasis because they are frequently misunderstood as sentiment indicators. They are not. When investors redeem shares, the fund sells physical metal into the market. When they create, the fund buys. Sustained redemptions are systematic, rule-based selling that hits the spot market regardless of what any individual thinks gold is worth. That is why the flow ledger has become the highest-frequency indicator worth watching in the complex.
For now the redemption trend is intact. Reversing it requires either a fall in real yields or a price level low enough to attract value buyers. Neither condition is currently satisfied, which is why rallies keep failing at successively lower highs.
Central Banks Are Still the Bid That Does Not Blink
Underneath the investment-flow carnage, the official sector has kept buying, and that is the single most important structural change in this market.
Projections for 2026 central bank purchases cluster between 750 and 1,000 tonnes, with one major US bank modelling roughly 800 tonnes. That demand is rate-insensitive by construction. Central banks are not optimising against TIPS yields; they are diversifying reserves away from dollar exposure on multi-decade time horizons. They do not sell on bad days. That behaviour fundamentally alters the relationship between corrections and structural demand — any long-term model that treats gold’s marginal buyer as a yield-sensitive Western allocator is working from an outdated framework.
The scale of the shift is captured in one statistic: gold surpassed the share of US Treasuries in global central bank reserves for the first time since 1996. That is not a trading development, it is a monetary one. Alongside the buying, a repatriation trend has accelerated, with central banks bringing physical holdings back onshore rather than leaving them in traditional custody centres — a signal about counterparty risk rather than price.
The macro backdrop supporting that behaviour is not improving. Global debt loads reached a record $353 trillion in the first half of 2026, with the government share approaching one-third of that figure, also an all-time high. An active fiscal and inflation impulse of that magnitude is precisely the environment in which sovereign reserve managers add monetary hedges. The US Treasury Secretary’s recent remark that America holds more than $1 trillion in gold and that it does not matter for the dollar drew attention for exactly this reason — the market is increasingly unconvinced by that framing.
There is a countervailing force worth naming. Demand destruction runs in both directions. As gold climbed toward $5,600, central banks needed to buy fewer tonnes to hit reserve-share targets. At $4,053, the same dollar budget buys considerably more metal, which mechanically supports tonnage. Jewellery consumption, which accounts for roughly 40% of total gold demand, works the same way and had already shown weakness at the highs.
The practical conclusion: official-sector demand sets a floor, but a floor is not a catalyst. It defines how far gold can fall, not when it starts rising.
Silver Is the Same Trade With Twice the Beta and a Supply Story
Silver traded at $58.43 on Thursday, down 1.67% from $59.83, and is now off 17.23% since the start of 2026 despite being up 50.49% over twelve months. Those two numbers together describe an asset that went vertical and then broke. The all-time high of $121.67 set on January 29 means silver has surrendered roughly 52% of its value in under six months.
The gold-to-silver ratio at 69.54 is the number to watch. It had broken below 70 during the recent rebound as silver outran gold for four consecutive sessions, and it ticked back up Thursday. A sustained move below 70 has historically marked periods when the industrial and monetary bids are working together. A move back above 75 would signal that the industrial leg is failing.
That industrial leg is what separates the two metals. Silver has the highest electrical conductivity of any metal and sits in electronics and solar manufacturing supply chains, which means roughly half its demand is cyclical rather than monetary. In an environment where energy costs are spiking and the Fed is threatening to tighten into it, cyclical demand is exactly the wrong exposure. That is why silver falls harder than gold on hawkish repricing and why it will rally harder if the rate picture softens.
The supply side has produced a genuine dislocation this month. India’s silver imports have slowed sharply as a new licensing regime disrupted shipments, pushing local premiums to multi-month highs. That is a physical-market signal decoupled from the paper price, and premium blowouts in the second-largest consuming market typically resolve either through import normalisation or through spot catching up. Mexico remains the dominant producer with output around 6,300 tonnes, roughly a fifth of global supply, with Peru near 3,100 tonnes and holding an estimated 22% of known global reserves.
Institutional forecasts for silver still sit dramatically above spot. The 2026 average consensus clusters around $79 to $81 an ounce across major bank research and industry surveys, implying the base case still expects a substantial second-half recovery from $58. The bearish tail sits near $44, reflecting the scenario where a hawkish Fed and a firm dollar keep the metal suppressed indefinitely.
That dispersion — a range from $44 to $81 for the same calendar year — is a fair measure of how little confidence exists in the rate path.
The Miners Took the Punishment Twice Over
Gold equities have been the worst place in the complex to be, which is the predictable consequence of operational leverage running in reverse.
traded at $74.17 as of July 21 against a 52-week range of $51.37 to $117.18 — roughly 37% below its high with a market capitalisation of $24.27 billion and a trailing price-to-earnings ratio of 9.52. The technical structure is broken: the 50-day moving average crossed below the 200-day on June 26, and momentum indicators turned negative on July 10. Weekly flow data has shown redemptions, with one recent week producing roughly $103 million of outflows.
June’s damage across individual names was severe. As physical gold fell 10.02% on the month, the sector bellwether shed 14.9%, fell 13.7%, dropped 15.3%, declined 16.5%, and plunged 21.7%. Those are two-to-one downside capture ratios against the metal, which is exactly what operational leverage produces when margins compress.
The mechanism is straightforward. Mining costs are largely fixed. When gold rises above all-in sustaining cost, every incremental dollar of price drops almost entirely to margin, which is why miners outrun bullion in uptrends — the sector delivered roughly 74% over one trailing twelve-month window against half that for physical. When gold falls, the same arithmetic inverts and margins compress faster than the metal declines.
There is a new complication this quarter that has not yet shown up in earnings. Energy is a substantial input into mining cost structures, and Brent has gone from roughly $70 in early July to above $100. Management commentary on all-in sustaining costs into the coming reporting season becomes the critical variable for forecasting third-quarter margins. A gold price down 27% from the high combined with a diesel and power cost base rising 35% in a month is a genuinely difficult combination.
Worth noting for anyone building exposure: over the ten years ending July 10, 2026, the senior miners fund returned 183.15% while the physical bullion trust returned 196.51% and a cheaper competing miners fund returned 198.64%. The operational leverage story cost holders performance across a full decade rather than delivering it. Fee drag and rebalancing turnover explain part of that gap; hedging and capital misallocation explain the rest.
