Oil prices turn lower as Bessent says U.S. may have Iran deal "today or tomorrow"
Gold ended July’s final session pinned near $4,076 an ounce, down roughly 0.5% on the day after two sessions of gains, with COMEX futures trading $4,110.10 and slipping 1.21% in early hours. Spot XAU/USD traveled between $4,028.77 and $4,120.16 across the prior 24 hours against a previous close of $4,066.38, and the metal rejected $4,100 for the third session running.
The monthly number is what closes the book on a brutal stretch. Gold finished July up somewhere between 1.36% and 2.2% depending on the measurement, its first monthly advance after four consecutive months of declines. That is the entire bull case in one statistic — a 2% month, arriving after gold fell 16% in the second quarter, its worst quarterly performance since 2013.
The damage from the peak remains severe. Gold printed its all-time high at $5,595 to $5,602 on January 29, 2026, and trades roughly 27% below it. The 52-week range runs $3,273.77 to $5,595.46. The metal rallied hard after the U.S.-Iran war began on February 28, then surrendered nearly 23% from that post-war peak as energy prices drove inflation higher and forced the Fed hawkish rather than dovish. Year over year gold still shows a 21.51% gain, and 2026 opened near $4,932, which puts the year-to-date decline around 17%.
Friday’s specific drag came from currency mechanics rather than gold fundamentals. The Bank of Japan held at 1% on an 8-1 vote, the yen resumed weakening after Thursday’s intervention-driven spike, and the dollar firmed into month-end. Gold sells off on dollar strength with metronomic reliability, and it did again.
The rest of the precious complex went with it. COMEX silver fell 1.71% to $58.01 after September futures opened at $59.26 and rolled to $57.94 by 8:37 a.m. ET. The gold trust dropped 2.04% and the silver trust fell 3.31%, both underperforming the underlying metals, which is what happens when Western fund flows lead the move.
The setup into August is the tightest gold has faced this cycle. Central banks just posted a record second quarter of buying. Western investors just pulled 45 tonnes out of physically backed funds. The 10-year Treasury sits at 4.731% and the 30-year at 5.263%. Bank targets have been slashed across the board. Every structural argument for gold remains intact, and every cyclical argument runs against it.
Tokyo Spent $53 Billion and Gold Got Two Days Out of It
The two-session rally that preceded Friday’s fade traces directly to Japan. Tokyo conducted yen-buying, dollar-selling intervention during Thursday’s New York session, its first foray in three months, and the scale was unprecedented. Bank of Japan account data compared against money broker forecasts put the operation at roughly ¥8.45 trillion, or $52.8 billion — likely the largest single-day intervention Tokyo has ever executed.
The price action was violent. Around 10:30 p.m. Thursday Tokyo time, the yen jumped from about ¥162.80 to the ¥157 handle inside an hour. USD/JPY fell from above 163 to as low as 157.95, a 3.35% move and the largest single-day yen gain since August 2024. The currency had been trading near a four-decade low. The dollar’s 2.4% drop was its worst session since January 2023, and gold caught the entire mechanical bid off that move.
It did not hold. By Friday the dollar had recovered to roughly 160.18 yen, the BoJ held its policy rate at 1% following June’s increase from 0.75%, and hawkish board member Hajime Takata dissented in favor of a hike. Governor Kazuo Ueda said CPI should accelerate clearly above 2% in the second half of the fiscal year and that the bank expects to keep raising rates. Markets read the hold as insufficient given the currency stress, and the yen resumed sliding.
The coordination angle matters for gold. Treasury Secretary Scott Bessent publicly described the yen as seriously undervalued, and reports indicated U.S. authorities carried out a rate check on the currency alongside Japanese action. Japan’s finance minister declined to confirm coordination while noting Tokyo stands ready to act. This would be Japan’s second major intervention of 2026, following ¥11.73 trillion deployed across April and May.
For gold the read-through runs one direction. Intervention produces a mechanical, temporary dollar decline that lifts dollar-denominated metals for hours or days. It does not change the rate differential driving the yen lower, and the differential is what determines the dollar’s trend. The dollar firmed Friday and was still headed for a July loss overall, but the intraday direction is what gold trades.
Gold’s two-day advance was borrowed from a central bank operation rather than earned from investor demand. That is the weakest possible foundation for a move, and it broke on the first session after the intervention faded.
A 4.73% Ten-Year Is Costing Gold $20 an Ounce per Basis Point
The single mechanical relationship that has governed gold in 2026 tightened again Friday. The 10-year Treasury yield jumped almost 7 basis points to 4.731% with an intraday print at 4.737%, the highest since January 2025. The 30-year surged 5.6 basis points to 5.263% after touching 5.244% on Wednesday, a 19-year high. The 2-year rose 6.6 basis points to 4.295%.
The sensitivity has been quantified. Gold has dropped approximately $20 per ounce for every one basis point rise in 10-year real yields since late February. Applied to Friday’s move alone, that mechanism accounts for a $140 headwind on the day. Applied across the second quarter, it explains most of a 16% decline that arrived while central banks were buying a record 289 tonnes.
The curve shape compounds it. The front end priced a Fed on hold while the long end priced inflation the Fed is not containing — a bear steepener after a hawkish-dissent hold. For a zero-yield asset, the long end is what matters, because that is where the opportunity cost of holding bullion gets set. A 5.26% thirty-year with core PCE at 3.3% leaves a positive real yield near 2%, and gold has never competed well against that.
The inflation data offered no rescue. June PCE fell 0.1% month over month, putting the annual rate at 3.7% from 4.1% in May, with core rising 0.1% and holding 3.3% year over year. Gasoline and energy goods prices sank 9.2% in June, the largest monthly drop since August 2022, and that single line is why headline cooled. Gasoline is back above $4 a gallon with the ceasefire broken and fighting resumed, so the July print reverses.
Consumer inflation expectations stayed elevated. The final July University of Michigan reading came in at 55.2, up from 49.5 in June, with one-year inflation expectations at 4.2% and five-to-ten-year at 3.3%. Both remain far above the 3.4% one-year reading in February before the conflict began.
That combination — sticky 3%-plus underlying inflation with a central bank leaning hawkish rather than easing — is the worst configuration gold can face. Inflation without monetary accommodation produces higher nominal and real yields, and gold loses on both. Oil rising roughly 21% in July, its steepest monthly gain since March, adds to the pressure by keeping breakevens elevated without triggering a policy response.
The 9-3 Vote That Removed the Rate-Cut Trade
The Federal Open Market Committee held the funds rate at 3.50% to 3.75% on Wednesday, a fifth consecutive meeting without a move, on a 9-3 vote. Cleveland’s Beth Hammack, Minneapolis’ Neel Kashkari and Dallas’ Lorie Logan dissented in favor of an immediate quarter-point increase — the most hawkish dissent since September 2016.
Gold caught a brief bid on the hold before the bond market corrected the interpretation. Markets now price roughly a 63% probability of a September rate hike, down from near 80% before the decision but still the dominant scenario. Logan argued rates should be modestly higher and that inflation is not on track back to 2%. Hammack and Kashkari both cited the risk of five-plus years of above-target inflation becoming entrenched. On Friday, dissenting officials reinforced those positions publicly, adding to the yield move.
Chair Kevin Warsh gave the market nothing to anchor on, stating there is no soft inflation target and no implicit target — only 2% — while declining to offer forward guidance. He described the stance as watchful thinking rather than watchful waiting. The policy statement ran 166 words, roughly a third the length of the prior chair’s final one.
For gold this removed the last remaining bullish macro catalyst. Goldman Sachs explicitly cut its year-end target after stripping all remaining 2026 rate cuts from its forecast and pushing easing to June and December 2027. That single assumption change was worth $500 an ounce on the bank’s model. The metal spent 2024 and 2025 rallying on an easing cycle that has now been deferred by eighteen months and replaced with a live hike debate.
The dispersion inside the committee is itself a source of volatility. Three regional presidents voting for a hike, with the majority waiting on July and August CPI before September, means gold trades every inflation print as a binary. A hot July CPI puts a September hike near certain and takes gold toward $3,900. A cool print buys the metal a rally toward $4,300 without changing the structural picture.
The bond market’s message is more useful than the committee’s. A 30-year at a 19-year high while the Fed holds says the market doubts the central bank’s inflation resolve. That doubt is historically gold-positive. It has not been in 2026, because the same doubt is being expressed through nominal yields that gold has to compete against rather than through currency debasement gold benefits from.
Central Banks Bought a Record 289 Tonnes Into a 16% Drawdown
The demand data released July 30 delivered the strongest official-sector quarter on record and gold did not rally on it. Central banks purchased 288.9 tonnes in the second quarter, a 62% increase from 177.9 tonnes in the same quarter of 2025, and more than five times the revised first-quarter total of 57 tonnes. It was the strongest second quarter ever recorded.
They bought into a 16% price decline. That is the behavioral signal worth extracting — official buyers treated the worst quarter since 2013 as an accumulation window, which is the opposite of what Western fund flows did.
Poland led at 51 tonnes, lifting its reserves to 632 tonnes by the end of June. China’s central bank added 33 tonnes, its largest quarterly purchase since late 2023. A dozen smaller buyers filled out the total. The UAE led Middle East investment demand, surging 34% quarter over quarter and 30% year over year to 5.3 tonnes.
The sell side of the ledger is where the story gets more complicated. The Bank of Russia was the quarter’s largest seller at 22 tonnes, with first-half sales totaling 43.5 tonnes against a federal budget deficit approaching 6 trillion rubles. Russia spent a decade building a gold stockpile as sanctions insulation and is now liquidating it domestically to generate rubles. Turkey and Azerbaijan also sold.
Forward-looking survey data stayed strong. The World Gold Council’s 2026 Central Bank Gold Reserves Survey found 89% of respondents expect global official gold reserves to rise over the next twelve months, a record 45% plan to increase their own holdings, and 74% anticipate lower dollar holdings over the next five years.
That last figure is the structural thesis stated plainly. Three-quarters of reserve managers intend to hold fewer dollars in five years, and gold is the primary alternative. Mine production cannot expand quickly to meet that, and recycling shows little sign of increasing.
The problem is timing rather than direction. Total 2026 central bank purchases are now expected to finish below 2025’s pace. First-half net purchases came in at 345 tonnes, the lowest since 2022 and the weakest first half in four years, because the record second quarter was offset by a first quarter that barely existed.
The Revision That Erased 187 Tonnes of Demand
The most consequential number in the second-quarter report was not the 289 tonnes. It was the correction applied to the quarter before it.
Metals Focus and the World Gold Council revised first-quarter 2026 central bank gold demand from 244 tonnes to 57 tonnes, a 77% reduction, following an erratum notice issued in July and incorporated into the Gold Demand Trends report published July 30. The 187-tonne difference was reclassified as over-the-counter and other demand.
The original 244-tonne estimate had been presented as a continuation of elevated official-sector buying above the five-year average, consistent with multi-year accumulation. Investors positioned on it. The revised 57 tonnes marks the weakest first-quarter official demand in more than a decade.
This matters because the central bank bid has been the load-bearing argument in every bullish gold thesis since 2022. Bank price targets, the de-dollarization narrative, and the debasement trade all rest on official-sector accumulation running structurally above historical norms. A 187-tonne downward revision to a single quarter does not break that thesis, but it does establish that the data underlying it carries meaningful measurement error in the direction that flatters gold.
Reclassifying the volume as OTC changes its character entirely. Central bank purchases are sticky, policy-driven, and price-insensitive. OTC demand is discretionary, positioning-driven, and exits when the trade stops working. Moving 187 tonnes from one bucket to the other converts what looked like permanent official demand into what was actually institutional speculation — and institutional speculation is exactly what left in the second quarter.
The rebound to 289 tonnes restores official demand to levels typical of the prior four years, which is the reassuring read. The less reassuring read is that the first half aggregate of 345 tonnes is the lowest since 2022, and the World Gold Council described first-half activity as continued but uneven rather than accelerating.
Total gold demand including OTC held flat year over year at 1,269 tonnes in the second quarter, bringing first-half demand to 2,522 tonnes, up 2% from a year earlier and valued at a record $380 billion. Flat volume at record value is a price effect, not a demand effect. Strip the price and the demand picture is a market treading water while its most-cited support gets restated downward.
How much is hesitation costing you?
Every day you wait is a day the market moves without you. InvestingPro puts institutional-grade data and AI-powered insights in your hands—without the institutional price tag.
Stop waiting. Start investing with confidence.