Within three weeks, the same Treasury Secretary intervened first in the world’s most important exchange rate, then in its most important bond market, on a single philosophy: the price is wrong, and we will correct it.

In Part 2, I lift the lid on the combustible leverage machine those interventions are really guarding, before explaining why gold, silver and Bitcoin were the assets that ultimately prevailed.

The machine the fire brigade protects

Start with the question nobody in Washington likes to answer: who actually buys the bonds? The United States issues roughly USD 2 trillion of net new debt a year, and the old buyers are leaving the table. China has been reducing its Treasury holdings for a decade — a slow-motion geopolitical divorce — and Japan, the loyal old friend, finally has yield at home and buys only opportunistically. Officially, the great marginal buyer of recent years has been the “United Kingdom.” In reality, the UK line is hedge funds and levered fast-money managers booking their positions through London. The marginal buyer of the world’s most important bond market is no longer a patient pension fund or a central bank. It is a trader with a repo line.

What do these funds actually do? Andreas Steno Larsen recently supplied the missing piece, and it is worth dwelling on, because it explains every one of the Secretary’s gymnastics at a single stroke. Indeed, the funds run, in Steno’s marvellous image, a pawn-shop loop. A fund takes one dollar of real capital and buys a Treasury. It pawns that bond at the repo desk for roughly 98 cents of cash, buys the next bond, pawns that one too — and repeats the exercise, in some cases, twenty to fifty times. One saver’s dollar ends up financing an entire stack of government bonds: a wonderfully capital-efficient way of funding a two-trillion-dollar deficit without two trillion of fresh savings.

The motive is a small, boring anomaly called the swap spread: ever since post-crisis regulation made it expensive for dealers to warehouse bonds, long-dated Treasuries have often yielded more than the equivalent interest rate swap. Buy the cheap bond, fund it in repo, pay fixed in the swap, and what remains is a few tenths of a percent of almost-riskless-looking carry. Nobody gets rich on a few tenths of a percent — so the funds lever it ten to fifty times.

Chart 6: The repo recycling loop — one dollar of capital, a levered stack of Treasuries. Schematic: Incrementum AG; concept: Andreas Steno Larsen

Want institutional-grade analysis for all things gold, silver, mining and Bitcoin?

The Sound Money Report is your global source, brought to you by the fund managers and analysts behind the world-renowned In Gold We Trust report.

As a thank you for being part of the SMR community, we’re offering you 30% off your first year for a limited time only.

Upgrade to full access and support our mission to bring timely insights to sound money investors worldwide.

Offer ends September 15.

GET 30% OFF

Read more

Hold that image. It is the lens for everything that follows. Because in those very same weeks, another fire brigade was scrambled — over the long end of the US yield curve. And unlike the pilots above the Weissensee, this one cannot use water. Only liquidity.

For years, I have described the endgame of fiscal dominance as a reaction function: when the cost of servicing government debt becomes politically intolerable, the state will not accept the market’s price for its own liabilities — it will change it. Readers of the In Gold We Trust report have seen this argument in a dozen variations. Until this month, it was a thesis. Now it has a date, a press release, and a Treasury Secretary reading it aloud on television.

The fire at the long end

The setup was textbook. In mid-August, a major bond selloff pushed the 30-year Treasury yield to its highest level since 2007, peaking at 5.34%, with the 10-year near 4.7%. The market was not repricing growth; it was repricing trust. Total public debt outstanding topped the USD 40 trillion mark that same Wednesday — and as Tavi Costa has pointed out, the resulting arithmetic is untenable for a debtor of America’s size: with breakevens near 2.3%, the 10-year real yield sits around plus 2.4% — while a sovereign intent on inflating away its debt burden requires precisely the opposite: negative real rates, sustained for years.

Hold onto that number — plus 2.4%. We are going to come back to it.

Read more

Spot gold traded near USD 4,456 late in the New York session, down 3.14%. Silver went with it, down 4.24% to USD 66.21. Both metals closed the week under their 200-day averages. The cause was Fed chair Kevin Warsh, speaking at the Kansas City Fed’s annual symposium in Jackson Hole. The interesting part isn’t that it happened. It’s who was holding when it did.

Sixty-five months above target, and a chair who says so

If you hold metal, Friday was the day the interest-rate argument stopped being theoretical. Kevin Warsh has been Fed chair since May and has sat through exactly one policy meeting, at which rates were held between 3.5 and 3.75%. Until Friday he had declined to say what would make him move.

At Jackson Hole he moved anyway, without calling it guidance. On the summer’s inflation prints he said they do not tell him underlying trends have meaningfully improved. He called the 2% target firm and fixed. He said financial conditions are not currently restrictive. He said the Fed has work to do.

They do not tell me that underlying trends have meaningfully improved.

Kevin Warsh, Jackson Hole, 28 August 2026

The data behind him is the part that makes it stick. Core PCE printed 3.3 percent in July, the sixty-fifth consecutive month above the Fed’s target. Headline PCE came in at 3.7%, above estimate. Second-quarter growth annualised at 1.5% and the economy shed 23,000 jobs last month.

Markets moved the odds on a September hike from roughly a third to near half within the session, on CME FedWatch. Short-end yields lifted, the dollar firmed, and everything priced off cheap money went the other way.

Qatar has cancelled LNG cargoes into November

Part of the reason that inflation print was hot has nothing to do with the Fed, and no rate decision can fix it. Qatar extended force majeure on liquefied natural gas (LNG) supplies this week, cancelling deliveries to Asian buyers into October and European buyers into November, with traffic through the Strait of Hormuz at a near standstill. The United States and Israel have been at war with Iran for six months.

Dutch TTF gas has run up roughly 135% this year and nearly 60% since early July. Brent in euros sits half again above where it started the year. German chemical output is being read as a recovery in some quarters, when a good deal of it is European industry buying ahead of a supply problem.

Oil fell on Friday. On reports of a possible shipping arrangement through the strait and evidence that more crude is moving than the headlines suggest. So the constraint is loosening, not tightening. That matters, because a gas price built on a blockage unwinds fast when the blockage clears.

Ronald Stoeferle flagged the other half of this on Tuesday. Bank of America’s August fund manager survey puts just 1% of managers expecting oil above USD 100 by the end of 2026, with the survey’s own forecast at USD 76. The constraint is real, it’s visible, and almost nobody is positioned for it going wrong.

Storage is the part to watch from here. European fill has been running behind its long-run average all summer, which is what turns a shipping problem into a heating-season problem.

What none of this does is respond to interest rates. A cargo that cannot sail isn’t a demand problem.

Read more

Yet neither the puncture’s quiet hiss nor the frantic honking from the concerned motorists around him is enough to break through his deluded daydream of eternal dollar dominance.

Indeed, that truck is a dead ringer for the dollar today: still the vehicle powering US hegemony, but handled so recklessly that fellow drivers are increasingly being forced to consider alternatives.

Back in 2022, the old Russian Lada was the first to pull away from the dollar, when Moscow’s gold buying frenzy was supercharged by the West’s sudden freezing of $300 billion in reserves.

Fast forward to Q1 2026 and even the trusty Toyota is weighing up a lane change, after the largest foreign holder of US debt, Japan, cut its US bond holdings by the most in almost four years.

More recently, the sleek BYD has continued its acceleration away from the dollar, with China’s central bank buying gold at the fastest pace since 2023 while quietly building the plumbing for a rival currency.

Not least, the usually steadfast Mercedes has also been tapping the brakes, after Deutsche Bank became the first non-Chinese bank authorised to clear renminbi trades in Europe earlier this month.

Even the once friendly pickup in Uncle Sam’s mirror, Canada, is signalling its exit, amid Carney’s growing indifference to Trump and his Art of the Deal-style trade negotiations.

Clearly, then, nations are increasingly re-routing away from the dollar. In today’s article, we’ll explain exactly what it all means for gold and silver — hard assets that no government can print into oblivion or freeze at will.

First, though, we must delve into the heart of the problem at home, where the Trump administration remains asleep at the wheel, against the backdrop of a snowballing national debt pile.

When the fire brigade brings a water pistol

Starting with the number that grabbed the headlines, total US government debt crossed $40 trillion for the first time in history earlier this month, marking a doubling in the 9 years since 2017, while charting a trajectory toward a whopping $64 trillion debt pile by 2036, as per Bank of America.

When pressed on it, Treasury Secretary Scott Bessent put on a brave face, insisting there was “nothing magic about the $40 trillion number” and that the US can simply “grow its way out”.

For a market already nervous about the sheer scale of government borrowing, that reassurance rang hollow.

Why?

Because the bond market is voting with its feet. Long-term Treasury yields have climbed to levels not seen in nearly two decades, and it was in that precise context that the Treasury hopelessly fired a blank.

On Wednesday, it announced it would at least double the size of its long-end buyback operations, from a $2 billion ceiling per operation to at least $4 billion, beginning September 9.

The market took one look at the size of it, yawned, and sold.

By Thursday, thirty-year yields had clawed back almost the entire move, grinding back toward the two-decade high they had just set, as investors judged the intervention far too small to matter against a $32 trillion market that must absorb roughly $1 trillion of new debt every few months.

This brings us back to our bearish outlook for the dollar: when government debt crosses an unthinkable threshold, when the Treasury itself cannot talk yields down, and when a hastily doubled buyback program still cannot move the needle, trust in the underlying currency inevitably wanes.

To make matters worse, the US has been at war with Iran since late February, and Operation Epic Fury has been costly, to say the least.

The Pentagon officially pegs the cost at around $37.5 billion, though independent estimates run past $100 billion, with some early phases burning through $1 billion a day.

On top of that, the administration has requested a record $1.5 trillion defence budget for 2027 — a jump of more than 40% in a single year and the sharpest increase since WW2.

Throw in the $1 trillion-plus in annual interest payments, and it begs the question: Who in their right mind is going to keep buying US debt? And how long will they continue to do so before seeking refuge in gold?

The answer starts in Tokyo, and it is by no means reassuring.

Read more

On Tuesday evening the United States crossed forty trillion dollars of debt. On Wednesday morning the Treasury announced it would at least double its purchases of its own long bonds. If you hold metal, that sequence is the week. Everything below follows from it.

The borrower turned up as the buyer

The largest debtor in the world set a record, and the next morning the same institution said it would buy more of its own paper. For anyone holding metal, that is a working description of who is left on the bid.

Total public debt outstanding stood at USD 40,047,426,000,000 at the close of business on Tuesday 18 August, reported in the Daily Treasury Statement. The pace is the part worth holding on to. Thirty-eight trillion in October 2025. Thirty-nine in March 2026. Forty in August. Five months, then five months. A decade ago the figure was USD 19.4 trillion.

At 08:32 New York time on Wednesday, the Treasury said the maximum size of its liquidity support buyback operations for longer-dated nominal coupons would rise from USD 2 billion per operation to at least USD 4 billion. The sectors named are the 10-year to 20-year and the 20-year to 30-year. It takes effect on 9 September and runs to 4 November.

Set the two numbers side by side and the arithmetic makes its own comment.

Four billion an operation, against forty trillion outstanding. One is a ten-thousandth of the other.

Formally this is liquidity support, not policy. The deficit does not move, and the money is issued rather than printed. Which leaves the mechanism. Retire long bonds, fund the retirement with fresh borrowing, clear room to borrow again.

The debt is refinanced by more debt so that more debt can be sold. That is how the thing is built.

Hold it alongside one other fact from the same week: a USD 16 billion auction of new 20-year bonds. The issuer said it would buy more long paper days before selling more of it.

On Thursday, Scott Bessent went on CNBC to say the buybacks could run above the announced ceiling per issue, and to explain why.

The yields don’t reflect the underlying fundamentals.

Scott Bessent, US Treasury Secretary, CNBC, 20 August

He is not forecasting there. He said in the same answer that part of this is signalling. The claim is that August is thin, corporate issuance has crowded the tape, and the market is reading the wrong things. Whether that is right matters less than whether anyone acts on it.

The safe half of the portfolio moved

If you own a traditional portfolio, this was the week the half that is not supposed to do anything did something.

The 30-year Treasury yield closed at 5.31 percent on Monday 17 August, on the Treasury’s own par yield curve. That is the highest close of 2026, above the previous peak of 5.27 on 31 July, and a nineteen-year high. By Wednesday it was 5.19. The 20-year fell from 5.30 to 5.17 across the same three sessions, and the 10-year from 4.72 to 4.65.

Twelve basis points off the 30-year in two sessions is a large move in the instrument that is supposed to be the boring one.

Then it went back. By Thursday the 30-year had erased the decline, the 10-year had returned to 4.69 percent, roughly where it sat before the announcement, and both were higher again after Bessent spoke. The signal was sent, backed with money, repeated on television, and declined inside a session.

Two things. The Federal Reserve released its July minutes the same day, so anyone pinning the whole move on the buyback is claiming more than two data points support. And intraday prints circulating this week ran above 5.34 against a close of 5.31. We use closes, and we name them.

Three assets that share only scarcity

Nothing here was running before this week. Which is the reason to look at it closely.

Bitcoin closed at 64,625.40 on Tuesday, 69,487.60 on Wednesday, 73,086.97 on Thursday and 78,126.63 on Friday, on Coinbase daily closes. Up 20.9 percent in three sessions.

Silver rose nearly 6 percent on Wednesday alone and reached 69.63 on Friday on spot, up 2.29 percent on the day and 16.60 percent over the month, a third consecutive weekly gain. Gold, also on spot, went above USD 4,500 on Wednesday for the first time since early June and above USD 4,600 by Friday, its highest since mid-May. Metal prices here are spot rather than futures, which were running some sixty dollars apart on gold during the week.

Eight days earlier, on the World Gold Council’s Friday tape, gold was up 0.52 percent year to date, silver was down 9.74 percent and bitcoin was down 28.30 percent. Three assets that had done nothing or worse all year moved together inside forty-eight hours, on two fiscal events. Bitcoin remains down roughly 11.9 percent on the year even after the run.

What they have in common is scarcity. Somebody can print more dollars this afternoon. Nobody can print more of these.

Bank of America's August fund manager survey put a net 16 percent of managers in the camp that gold is undervalued, the most since March 2023 and up from 6 percent the month before. Nobody was crowded into this.

Read more

Regular readers of my articles on the R&S sector already know that this sector is characterized by high profit margins, which have received an additional boost from the recent sharp rise in gold and silver prices, leading to very strong results in Q2 2026—though these fall short of the figures from Q1 2026. This is related to the decline in gold and silver prices compared to Q1 2026, when record prices for gold and silver also led to record results for R&S companies.

Since this article focuses more on the industry’s medium- and long-term developments, I will place less emphasis on the Q2/2026 quarterly figures than in previous analyses.

I will present and analyze the highlights of each company and venture a few predictions for the future.

Let’s get started!

Cash Flow Trends

While I was already impressed by the excellent cash flow performance in the second half of 2025 in my last article in April, the first half of 2026 shows a further significant improvement.

Read more

Ronnie’s essay on Tuesday, The Incentive Standard, was about what the people setting policy are paid to do. This is the duller half of the same question. What did the people who actually hold the metal do with it this year, and how well do the numbers we all quote describe that?

Not very well, as it turns out. A half-year of American selling comes down to two months. A year of outflows gets answered in a single week. The weight the big commodity indices give gold is worked out from the one venue where gold does about half its business.

Two months did all of America's gold selling

If you have read anywhere this year that American investors have been dumping gold, that is true. It is also almost entirely about two dates on a calendar, and knowing which two changes what the number means.

March took 85 tonnes out of US gold-backed ETFs, worth USD 13bn. June took another 40, worth USD 5bn. April and May, between them, took four. The half-year came to 61 tonnes, the fourth largest H1 decline on record and the second largest in dollar terms.

Strip out March and June and the same funds would have taken in 65 tonnes.

The World Gold Council’s own reading is that this was two discrete liquidation episodes rather than a sustained withdrawal, with different causes in each. March was de-risking into the escalation with Iran, profit-taking on 2025 gains, and rebalancing out of allocations that had drifted above target. June was a firmer dollar, higher yields, and a falling price feeding redemptions that fed the price.

Both months ended a quarter, and that is not a coincidence. Funds report their holdings at quarter-end, and a position that looks wrong on a statement tends to get closed before the statement is printed. What the half-year figure measures, then, is mostly the reporting calendar. It is not eight hundred thousand Americans deciding gold was finished.

The gold demand number that was almost nothing

The retail holder never left. That is worth saying first, because the headline says the opposite.

Total US gold demand came to 7.9 tonnes in the second quarter, against a ten-year quarterly average of 90. Down 94 percent on the year. Read the line items, though, and the collapse disappears. Jewellery ran 22.2 tonnes, up 65 percent on the quarter. Technology held at 16.3, up 6 percent on the year. Bar and coin came in at 13.8 tonnes, down from the first quarter but up 28 percent year on year. Every category a person buys was flat or rising.

The ETF line was minus 44.4 tonnes, and it swallowed all of them.

The global picture separates the two holders cleanly. Bar and coin demand worldwide reached 784 tonnes in the first half, one of the strongest in the series, with spending up 22 percent to USD 86bn. North American dealers report two-way flow, and continued gold buying alongside steady silver liquidation.

The retail holder is still here. The fund line moved, and the fund line is what gets reported as demand.

Gold and bitcoin, bought back the same week

If you hold both gold and bitcoin, you have spent this year watching both go the wrong way. So has almost everyone else. Then, in one week, both got bought back.

US bitcoin ETFs recorded their highest weekly inflow since April, on Bloomberg data. That was the week gold rose 7.67 percent on the World Gold Council’s weekly monitor, its strongest week since March 2020, and gold ETFs took inflows in every region, 19 tonnes in total. Bitcoin closed the week at 64,935.4, up 3.24 percent, and down 25.91 percent year to date. Two assets with nothing in common except that Western investors spent this year selling both, and bought both back inside the same five sessions.

The honest caveat. One week is one week. We’re not going to call it a pattern on a single print, and there is an obvious common cause. The July payroll report landed that week showing the US had shed 23,000 jobs against a survey looking for 80,000 gains. Rate-hike odds collapsed, the dollar and yields fell, and both assets are priced off the front end.

But the shape is worth watching. Neither move was a story about gold, and neither was a story about bitcoin.

Read more

The author, shortly after rediscovering the concept of asymmetric risk.

Convalescence has its compensations. I have been reading more than I have in years, re-watching Papillon — a man locked up for years who never stops thinking about escape, and no, I cannot imagine why that film resonated with me just now — and, above all, I have been forced to slow down. Forced deceleration, it turns out, is a condition highly conducive to thinking about incentives. One-handed and full of painkillers, you stop watching what people say. You start watching what they are paid to do.

Which brings us directly to our topic.

I. The Emperor and the Railway

Emperor Franz I of Austria was, by the standards of his age, no fool. When engineers petitioned him in the 1830s to build steam railways across his empire, he refused — not because he misunderstood the technology, but because he understood it perfectly. Railways would move people, ideas, and unrest faster than his police could follow. Nicholas I of Russia reached the same conclusion by a different route: his empire’s entire social order rested on serfs staying on the land, so he throttled the banking system, banned industrial exhibitions, and restricted construction. Neither monarch was ignorant. Both read their payoff matrices with complete accuracy.

They owned nothing that industrialization would have made more valuable — so they suppressed it.

Contrast England. The Luddites who smashed textile machinery between 1811 and 1816 were not fools either; they were highly skilled workers who understood precisely what was happening to them. They lost anyway — because the English aristocracy owned the mineral rights beneath their estates, and coal was becoming the fuel of the industrial age. The gentry stood to profit, so the machines were protected with the full force of the state: twelve thousand soldiers garrisoned in the affected counties, machine-breaking elevated to a capital crime, seventeen men on the gallows. The technology did not win because it was good. It won because the powerful were paid to let it win.

The price of choosing the Austrian path is well surveyed. Economist Diego Comin has shown that the speed with which a country adopts new technologies explains at least a quarter of today’s wealth differences between nations. The Ottomans, after all, did eventually get the printing press. Just a few centuries late.

Charlie Munger compressed this entire literature into eleven words: “Show me the incentive and I will show you the outcome.” The aphorism is usually deployed to explain corporate misbehavior. We would argue it is the single most useful analytical instrument in monetary economics today. Central banks publish dot plots, forecasts, and carefully lawyered statements. But if you want to know where policy is ultimately headed, do not read the speeches. Read the payoff matrix.

Below, we apply Munger’s test twice: first to the Federal Reserve under its new chairman, then to Japan — the country where every incentive described in this essay has already reached its logical conclusion.

II. The Great Pretender’s Payoff Matrix

Kevin Warsh arrived at the Fed talking like a young Paul Volcker. “We’ve missed for five years, and we’re going to fix that,” he declared of the inflation target at his very first press conference, batting away any suggestion of raising the 2% goal with the dry observation that the “two” stands to the left of the decimal point. Markets briefly believed him: in the nine days to July 24, the CME FedWatch probability of a rate hike jumped from roughly 11% to 38%. At the July meeting, three FOMC members dissented — only the sixth time in three decades that so many broke ranks at once. Dario Perkins of TS Lombard rendered his verdict after the press conference with admirable economy: talking, so far, appears to be the chairman’s greatest talent.

We share the skepticism — and not because we doubt the rhetoric. We have simply read the incentives. Three of them, to be precise.

Subscribe now

Read more

That’s what reserve management looks like when the thing you defend with and the thing you defend against are the same asset. Japan is being lent dollars against its Treasuries so it can avoid selling them. Korea sold dollars outright, then opened a gold programme it has left dormant since 1967. Indian households did the domestic version of the same move, borrowing against their gold at a record pace while selling less of it than at any point in nearly three years.

The metal didn’t need a story this week. It got one anyway.

The Fed lent Japan a way out

Japan’s currency broke. The fix arrived as collateral.

In late July the United States bought yen for the first time since 2011, coordinated with Tokyo, in the first joint action of its kind in fifteen years. The funding is the detail worth holding onto. The Treasury sold euros to buy yen, keeping its own currency and its own bonds out of the transaction entirely.

Then on Sunday 2 August, Bessent went further in public. He praised the Federal Reserve’s Foreign and International Monetary Authorities Repo Facility, which lets a foreign central bank hand its Treasury holdings to the Fed and draw dollars against them, and said Washington “would encourage it to be upsized in the coming months”. Bloomberg reported it the following day. Finance Minister Satsuki Katayama confirmed Japan had bought yen on Friday and would use the facility going forward.

The mechanism is the story. Selling Treasuries to raise dollars floods the market with supply and lifts yields. Lending them to the Fed instead raises the same dollars and leaves the bond market untouched. The facility caps out at USD 60bn per counterparty per day, and lifting that ceiling would need approval from the Fed’s foreign currency subcommittee.

Read it with the labels off. The largest foreign holder of US government debt needed dollars. The US Treasury is publicly lobbying the central bank to make borrowing them easier for Tokyo. Japan already has a Fed swap line and chose not to use it.

Hedgeye circulated the chart that explains why rates couldn’t do the job, from Apollo’s chief economist. USD/JPY stopped tracking the US minus Japan ten-year yield differential after April 2025. Since then the differential has fallen while the currency kept climbing. Once a currency stops answering to rate spreads, what’s left is intervention and plumbing, and both were used inside a fortnight.

Positioning shows how crowded the other side had become. Leveraged funds and asset managers together held a bearish yen position approaching the largest since the financial crisis, on CFTC data compiled by Bloomberg.

The price fell and the structure got worse

Oil had a quiet week on the tape and a loud one underneath it.

Prices dropped roughly 8 percent after Bessent told CNBC on Tuesday that a deal to open the Strait of Hormuz with freedom of movement could come as soon as Wednesday. On Thursday the Iranian state news agency Fars published a draft plan going the other way, and Brent closed 3.8 percent higher at USD 82.49 with WTI settling at USD 77.29. That draft would bar US and Israeli vessels, require compensation from states Iran says have harmed it, and levy penalties worth 20 percent of a ship’s cargo. By Friday, Brent sat at USD 81.92 and WTI at USD 76.96, both tracking a weekly loss near 9 percent.

What’s actually being negotiated matters more than the tape. Iran wants transit fees worth 5 to 7 percent of the value of shipments. Oman has been working toward roughly 3 percent. Washington wants zero. Four people familiar with the industry told Reuters that US sanctions and insurance restrictions make any fee framework unworkable in practice, and Bjarne Schieldrop of SEB Research argued the deal as drafted hands Tehran leverage no US administration could politically absorb.

Strip the diplomacy away and the best outcome currently on the table is a permanent toll on a fifth of the world’s seaborne oil.

The market has absorbed this calmly because inventories paid for the calm. Rory Johnston of Commodity Context put it plainly to Al Jazeera in July: the crisis was blunted by an ample stock cushion, and “much of that cushion has now been depleted”. Transit data tells the same story. Roughly 130 vessels crossed the strait daily before the war began in late February. Fifty-seven crossed across an entire weekend in mid-July.

Luke Gromen has been charting the same problem from the other end, plotting the ten-year Treasury yield against USD/JPY multiplied by oil. A creditor’s currency and its energy bill move together, and Japan is where both land first.

What would change this read. The one-year US inflation swap now sits at 1.8000, down from 3.5700 on 29 April and well under the Fed’s target. If that’s genuine disinflation rather than a growth scare, the fiscal-dominance case weakens and the metal has less work to do. Pulling the other way, Steve Hanke’s Divisia M4 grew at an 8.5 percent annualised rate in the second quarter, against 6.6 percent a year earlier and 5.1 percent the year before that. Both are properly sourced. They point in opposite directions. We’re watching which one turns out to be early.

Read more

Tim Urban gave us this distinction to describe how a few people reason from raw materials while everybody else assembles borrowed conclusions. It transfers to money almost intact. You are paying a cook or a chef. If you run your own capital, you are one or the other, and the past six months in precious metals have been unusually good at telling them apart.

The Omelette Objection

Discard the obvious version of this idea first, because believing it will cost you money. The chef is not the man with the complicated portfolio and the cook is not the man with the simple one. A chef can make an omelette: three eggs, butter, salt, and a pan whose temperature he judged by holding a hand above it. Nothing on the plate is elaborate, but everything on it is deliberate.

The investor who buys bullion every month and never trades may be the most sophisticated man in the room. Ask him why it works. If he answers with real interest rates, central bank demand, and his own demonstrated inability to sit still through a drawdown, he is a chef who chose a simple dish for good reasons. If he answers only that gold always goes up, he is a cook, and his position is exactly as fragile as that of the man chasing whatever moved last week.

A recipe is the part of cooking that can be written down. It cannot tell you that this flour is absorbing more water than last week’s batch, or that your oven runs hot on the left. Hayek built a career on that gap between what can be written down and what actually decides the outcome. Portfolios are no different. Somebody can hand you a set of weightings in an afternoon. The reasons behind them take years to acquire, and the reasons are what you fall back on when the weightings stop making sense.

The distinction shows up in habits long before it shows up in returns.

Silver’s Two True Numbers

Nothing has separated the two kinds of investor this year more cleanly than silver.

Gold has had the calmer year. It crossed $4,000 for the first time in October 2025, printed more than fifty record highs across that year, and trades near $4,050 today, up roughly twenty-three percent over twelve months. That is what the monetary metal looks like while central banks keep buying and nothing structural breaks.

Two statements about the metal are both true as I write. Silver trades near $58 an ounce, roughly fifty percent above where it stood twelve months ago. Silver also sits about fifty percent below the $121.67 it reached on 29 January. A cook will select whichever number matches his mood and call the result “analysis”. He was euphoric in January, and he is despondent now, and on both occasions he was reading a price rather than a market. In contrast, a chef asks what actually changed, because that is where the real story begins…

Read more

Consider what it absorbed to get there. A “hawkish” hold. Three dissenters who all wanted to go higher. The thirty-year at its highest since July 2007. A dollar that firmed and then broke on suspected Japanese intervention. Markets pricing a September hike at better than even odds. Behind all of it, a quarter in which the LBMA afternoon price averaged USD 4,506.29, 8 percent below the Q1 record and 37 percent above a year earlier.

That is most of a bear case arriving inside five days, and the ounce finished the month higher.

Half of what follows is not about this week at all. Three governments reached for bullion in six weeks. We covered the central banks on Tuesday. This is what the metal did when they told it where to go.

The froth left the equity market. The ounce is still the ounce.

They closed the door and the metal stopped

India didn’t stop wanting silver. India stopped being allowed to buy it.

Two moves in May. On 13 May the duty on bullion went from 6 percent to 15, gold and silver together, bars and dore alike, to conserve foreign exchange. Then, with more force, a licensing requirement landed. Most banks still don’t have a permit.

Metals Focus puts June shipments at 29 tonnes against 747 in January. Less than one percent of annual Indian demand, arriving in a month.

When a market that large can’t import, the metal already inside it is worth more than the metal outside. The 30-day premium is the highest since at least 2019, as far back as Metals Focus has tracked it. That isn’t sentiment. It’s a bar in Mumbai priced against a bar in London that can’t get there.

This isn’t demand dying. India front-loaded through late 2025 and is now drawing that stock down behind a closed door.

If licences are issued in volume and imports normalise inside a quarter, it’s administrative rather than structural, and the premium goes with it.

The other door opened

The second government moved the other way, and the metal moved with it.

On 30 April the central bank and customs jointly widened China’s gold import permits, effective 1 June: more ports, longer validity, fewer limits on repeat use. An easing, not a clampdown. It still started a race, because banks had reason to exhaust existing quota before the changeover. June imports hit roughly 173 tonnes, the highest since March 2024 and a third straight monthly rise. The first half came to around 820 tonnes, nearly double last year. Hong Kong took in 150.48 tonnes in June, up nearly 29 percent on the month and the most since late 2014, as the city stood up its own bullion clearing system.

Be honest about the mechanism. Part of that 173 is quota arithmetic, not conviction. And demand underneath isn’t uniformly strong: Shanghai Gold Exchange withdrawals rose 36 percent on the month to 87 tonnes, but the first half ran 27 percent below its ten-year average.

Imports at a two-year high while wholesale sits a quarter below average means the metal isn’t going to shop windows. It’s going somewhere that doesn’t sell it back.

Two governments, opposite instruments, and the metal did what each told it.

A country that bought its own production

Ghana went further. Ghana monetised it.

The revised 2026 budget puts 5 billion cedis, USD 429 million, behind gold purchases that add to reserves, and moves the programme from the central bank onto the government’s books.

They’re doing more of it because the first round worked. The Gold Board exported 104 tonnes of artisanal and small-scale gold in 2025 and generated over USD 10 billion in foreign exchange. Reserves went from USD 8.98 billion at end-2024 to 13.8 billion a year later, the cedi rose 41 percent against the dollar, and the deputy finance minister told parliament in June that those exports were the single largest driver of it. He called it stabilisation policy, not revenue.

Read that with the labels off. A government defended its currency with gold and the currency rose 41 percent.

It isn’t a triumph. Shifting purchases onto the fiscal balance sheet moves the risk there, and the programme is already lossmaking: trading losses widened the deficit on gold transactions to 9 billion cedis last year from 5.7 billion the year before, which is what the IMF objected to. Reserves fell USD 1.2 billion between March and June. Solving one problem with one commodity buys another.

The Fed said no. The long end said it louder.

The Fed held Wednesday at 3.50 to 3.75 percent. What matters is that it barely did.

Nine to three. Hammack, Kashkari and Logan all wanted a quarter point more, the first time since September 2016 that three dissented in the same direction. Inflation has run above target for over five years. Warsh called it a good family fight and said nothing about September.

The quotation marks around “hawkish” are doing work. A hawk holds because it can afford to. Warsh is cosplaying as one, and the costume comes off the day the arithmetic below makes it unwearable.

Then the curve disagreed with itself. The 2-year fell 4 basis points to 4.236 percent. The 30-year rose 10.5, closing at 5.201 after touching 5.244 during the session, its highest since July 2007. The Dow closed 1,153 points down.

The short end took the hold at face value. The long end wanted more to lend to America for thirty years. That’s not a market worried about the funds rate, and the funds rate isn’t what fixes it. Over USD 8 trillion of privately held Treasury debt rolls inside twelve months at whatever rate investors demand.

That is the bill that decides this, not the podium. A Treasury refinancing that much at 5 percent instead of 3 pays for the difference out of everything else, and there is a level at which no chair holds. Whether the cut arrives in October, or next year, or after something breaks, it arrives. What the metal did this week is what it does in front of that: it held while the hawk talked.

Gold went up.

The World Gold Council reads it against us, and they’re worth quoting for it: a breather in a downtrend running since January, with a hold above the 2025 high in real yields likely to weigh on gold. On their series, Bloomberg’s generic inflation-indexed ten-year, that high is 2.44 percent. The Fed’s H.15 series, which we charted last week, runs lower and puts its 2025 peak below that. Same idea, two rulers, and the levels don’t carry across.

The cleanest disproof: from March to July 2008 the 2-year rose a full hundred basis points and the Fed didn’t hike.

Read more

More broadly, gold soared from $1800 in January 2022 to highs above $5600 by January 2026. During that same period, central bank demand averaged 1,018 tonnes annually — more than double the annual average of 473 tonnes from 2010 to 2021.

In light of this relationship, it would be far from hyperbolic to describe central banks as the single most powerful determinant of the gold price today.

With that in mind, today’s article examines the principal drivers of central bank demand since 2010, before highlighting the nations most likely to influence the gold price in 2026 and beyond.

Where the Gold Went: 2010 to 2025

Upon viewing the top gold buyers from 2010 to 2023, what immediately emerges is the sheer number of Eastern nations that comprise the top 10.

Headed up by Russia with 1,684 tonnes, the top accumulators read like a roll call of the East: China in second with 1,181 tonnes, then Turkey, India and Kazakhstan filling out the upper ranks. Only Poland, sitting fourth with 256 tonnes, interrupts the pattern as the lone Western representative in the top 8.

Together, Russia and China accounted for 2,865 tonnes, more than the other eight nations on the list combined. In part, this sizable gold haul was inflated by the events of February 2022, when the West sanctioned Russia for initiating the Ukraine war, with a sudden freezing of $300bn in foreign reserves.

The Eastern concentration of gold demand can also be explained by the West’s first mover advantage. The US leads the world with 8,133 tonnes of gold reserves, followed by Germany at 3,350 tonnes, Italy at 2,452 tonnes and France at 2,437 tonnes. That’s compared with ~2,310 tonnes apiece for China and Russia, and 880 tonnes for India.

Given the West’s unassailable lead, it’s understandable why the East has felt compelled to play catch-up, while G7 nations act as though their gold holdings are already sufficient. This trend of BRICS capitalising on G7 complacency is one that continued into 2024 through India and China, as well as Brazil and China in 2025.

And yet, contrary to this narrative, the most persistent gold buyer of 2024 and 2025 was neither BRICS nor Eastern. Poland, of all countries, has topped the global tables in the last two years, purchasing 89.5 tonnes of gold in 2024, followed by an even larger 102 tonnes in 2025.

Flanked by the Czech Republic, which ranked fourth globally in 2024 with 20.5 tonnes and seventh in 2025 with 20 tonnes, Poland should therefore be described as the figurehead of a golden resurgence amongst central banks, not only in Europe, but globally, too.

Pole Position: Poland Leads H1 2026

Looking at the World Gold Council’s 2026 data, Poland has shown no sign of slowing down, with a further 64 tonnes added through the first five months of the year.

Once again, that positioned the National Bank of Poland (NBP) ahead of all central banks globally, extending a buying streak that began in April 2023 and has since culminated with 610 tonnes of gold holdings.

The 2023 pivot can be traced to a single moment: Russia’s invasion of Ukraine. Since then, Poland has pledged to lift its reserves to 700 tonnes, a level that would place it among the top ten gold-holding nations on earth. Justifying his rationale for this aggressive target, NBP President Adam Glapiński stated:

“We will keep buying gold until we reach 700 tons, because that’s where we are on the political map… A country that borders a war zone is subject to special scrutiny. Investors and international institutions assessing our credibility take our reserves into account.”

Read more

Both halves of that are true, and they are the whole issue. The speculative froth got flushed out of the paper market. The bid underneath it, the one that does not trade on a 24-hour chart, never went anywhere. The charts that follow trace how far the washout went, what should have kept gold down and did not, and who was buying the whole time.

They sold until there was almost nothing left to sell

Start with how thoroughly the paper market gave up. COMEX open interest collapsed to its lowest level since 2009, down sharply from its January peak. The nine-week RSI printed 28.85, oversold enough that the WGC flagged it red on its own table. Net long positioning came down to 547.2 tonnes.

None of that is a verdict on gold. It is a measure of how many people stopped waiting for it, and the answer is most of them. A separate read of the CFTC data puts speculative net positioning back near its long-run mean of roughly 199,000 contracts, down from a stretched high above 300,000 earlier in the year. A market that has already sold is a market with little left to sell.

The headwind that was supposed to sink it

Here is the part worth sitting with. Gold did not fall because the monetary story changed. It fell because the real yield on cash pushed to a new cycle high. The 10-year TIPS real yield climbed above its January 2025 peak of 2.34 percent to 2.39 on 22 July, a new high for this cycle. The 23 July Treasury auction of a new 10-year TIPS made the point in hard money, clearing at a real yield of 2.438 percent, the richest in years. A positive real yield is the one thing that reliably competes with an asset paying no coupon, and it is as high as it has been this cycle.

It gets sharper. The market spent the week pricing in a real chance of a rate hike at the 28 to 29 July meeting, 34.2 percent by Friday, against 65.8 percent for no change. Rising real yields and a live hike are the textbook setup for gold to keep falling. Instead it ran back toward USD 4,150 on the week. The headwind that is supposed to sink the metal blew straight at it, and the metal climbed anyway.

What would change this read: if real yields keep rising because growth is genuinely accelerating rather than because deficits are forcing the long end, the bull case weakens, and that is the honest bear scenario to watch rather than the one nobody holds.

Read more

The number that governs this is mNAV: enterprise value divided by the net asset value of the bitcoin held. Above 1.0, Strategy can issue shares, buy coins, and hand every existing holder more bitcoin per share than it owned the day before. Below 1.0, the machine runs backwards. Issuing equity at a discount to the underlying asset takes value from the man already holding and gives it to the man arriving. The In Gold We Trust report 2026 states the mechanism plainly: a premium permits accretive capital raises, and a discount can shut that window entirely.

The window is shut. That single fact explains June better than any headline about Saylor’s conviction. Strategy peaked near $136 on the fifteenth and traded down to $82 on the twenty-sixth, a few cents above its 52-week low. Bitcoin fell about nine per cent across those eleven days. Strategy fell forty. The coin did not become four times worse over a fortnight. The premium evaporated, and the equity fell through the hole where it used to be.

What the IGWT Report Calls Performance Gold

If you read the In Gold We Trust report 2026, you already own the framework needed to price this correctly. It was built for a different metal.

The report’s model portfolio splits its gold exposure in two. Safe-haven gold, at fifteen per cent, is bullion: the thing itself, no counterparty, no management, no capital structure. Performance gold, at ten per cent, is silver and the mining equities. There, you take the identical monetary thesis and bolt an operating company on top of it. Now you carry balance-sheet risk, management risk, dilution risk, and jurisdiction risk that the bullion in your vault never imposed on you. What you receive in exchange is leverage to the metal.

Strategy is Performance Bitcoin. Coins in your own custody are the safe-haven allocation, and they answer to nobody. MSTR is the operating company wrapped around the same thesis, and it answers to bondholders, preferred holders, a board, and the mood of an equity market that reprices sentiment every ninety seconds.

The report draws this parallel itself, and not as a compliment. In its chapter on gold producers, it observes that Saylor turned a software company into a digital-asset treasury vehicle pursuing a thesis parallel to a miner’s and that what separates the two is capital structure rather than strategy. Then it makes the cut that matters. Gold producers hold an advantage Saylor never had, because they produce the asset they would accumulate. A miner funds his treasury by declining to sell what he has already pulled out of the ground. Saylor funds his by borrowing, and that asymmetry sets both the risk and the opportunity in everything that follows.

Two Dials, Not One

Spot bitcoin has one dial. Strategy has two…

Read more

The chart below illustrates where sound money stood at the 10 July close. Gold down 6.15% on the year, silver down 16.46%, Bitcoin down 27.21%.

It has since got worse. During the week beginning Monday 13 July, gold traded below USD 4,000 and silver near USD 55. Read those bars as a floor, not a ceiling.

Now the part worth sitting with. June CPI fell 0.4% on the month and cooled to 3.5% against the 3.8% expected. Core PPI came in at 4.7% against a 5.2% forecast. Both prints cooled in the same week the metal broke.

Before anyone files that under inflation solved: the Bureau of Labor Statistics puts the June decline down to goods, and inside that a 7.3% drop in processed energy and an 18.0% fall in diesel. By the same 10 July figures, Brent had added 24.36% this year. Disinflation built on the oil price is disinflation on loan.

So gold is not being sold because inflation is coming. It is being sold because rates and the dollar went the other way, which is what the World Gold Council said about June and what the yield charts have said since.

That is the correction. It is real, and we are not going to pretend otherwise.

The hard money complex, year to date

Three assets, one direction, and the ranking is no surprise once you say it out loud: the further out the duration risk, the harder the fall. Gold down 6.15%, silver down 16.46%, Bitcoin down 27.21%.

Nobody here is being repudiated. Sound Money is being repriced against a rate of interest that moved, in proportion to how much of their value sits in the future. That is not a verdict on sound money. It is arithmetic on a discount rate.

The first losing half in five years

Gold fell 8% in USD over the first half and 10% in RMB, the stronger currency amplifying the local pain. The World Gold Council marks it the first semi-annual decline since 2021.

Five years is long enough that plenty of people holding this metal have never seen it post a losing half. Now they have. The books closed at the end of June and they closed red.

Worth keeping the frame, though. A losing half after five winning ones is what a correction looks like from the inside: uncomfortable, and not evidence of anything having broken. After all, no asset goes up in a straight line.

The rates did it

The World Gold Council’s own attribution for June is worth quoting in substance: messages from the new Fed Chair read as hawkish, real yields and the dollar went up, and rising opportunity cost plus cooling momentum did the damage. No inflation in that sentence anywhere.

Their chart puts the 10yr US real yield at the top of a three-year sideways range, resistance starting at 2.32%, with the 2025 high of 2.44% the level that would signal a genuine turn higher. And yet, it has broken neither level. The 2yr sits at 4.21%, pushing at resistance of its own.

Not a repudiation of the metal, a competing yield. The World Gold Council’s point this week is the one worth carrying: with policy less predictable, the volatility of yields matters as much as their level. Two consecutive downside inflation surprises under a Fed that is snookered by an unsustainable US debt burden is not the end of the argument for holding sound money insurance. In fact, it is the crux of the argument for holding gold, silver, and Bitcoin in the first place, as the rest of today’s Weekly Nugget lays out.

Read more

Isaac Newton spent an enormous share of his working life trying to turn base metal into gold — not as a hobby, and not briefly. He left an alchemical archive so vast that when John Maynard Keynes bought the papers at Sotheby’s in 1936 and read them, he concluded that the founder of modern physics had not been the first man of the age of reason at all. Keynes called him “the last of the magicians”.

Newton never made an ounce of gold. What he did instead was take a job at the Royal Mint — Warden in 1696, Master from 1699 — and in 1717 fix the guinea’s value in silver at a level that quietly, almost inadvertently, put Great Britain onto a gold standard for the next two centuries. The greatest mind of his age failed for decades to create gold, then spent his final decades defining money in terms of it.

Three hundred years later, the mining industry has finally solved the transmutation problem. It simply runs the reaction backwards. Rock becomes doré. Doré becomes bullion. And bullion becomes — reliably, at industrial scale, within days of leaving the processing plant — US dollars. It is the only transmutation in monetary history that works every single time. The industry performs it on very nearly 100% of what it digs out of the ground, converting the one asset that has preserved purchasing power for five thousand years into a currency that has lost roughly 98% of its value against that same asset since 1971.

Newton was trying to make gold out of lead. Our industry has perfected the art of making paper out of gold. Only one of those two processes has ever been shown to work at scale — and it is worth asking whether it is the one we want.

II. The Test

This was the argument of Chapter 22 of this year’s In Gold We Trust report — “The Corporate Gold Standard” — developed with Chris Ritchie, former president of SilverCrest Metals, and Alexander Scanlon. I return to it now for a reason that has nothing to do with authorial vanity and everything to do with price.

Subscribe now

Read more

Consider what it held above USD 4,100 against. American strikes on Iran across two days. Iranian retaliation on US bases. Oil spiking through Hormuz. Fed minutes from a committee that has stopped discussing cuts and started discussing inflation.

Everything the textbook says should break gold arrived at once. For the week beginning Monday 6 July, gold finished down about 1.5%, holding just above USD 4,100.

Silver did worse, falling about 4% on the week to steady near USD 60. Bitcoin was quiet; Mark Valek talked it through with Niko Jilch if you want the longer view.

The half-year books closed at the end of June, and the World Gold Council has just published them. Read them next to this week and something clicks into place.

The question is not what gold did. It is who was awake when it did it.

The tape has a timezone

The World Gold Council split gold’s first half by trading session, using twenty-minute tick data. The year-to-date return during Asian hours: plus 12.97%. During European hours: minus 1.32%. During US hours: minus 15.08%.

The same metal, the same six months, three different outcomes, decided by who happened to be at the desk.

The read. Gold’s pullbacks happened while America traded. Its Asia-driven rebounds happened while most Americans were asleep. Not a curiosity: a transfer, executed at the price America set.

Key Takeaway. This is not a gold bear market. It is a change of ownership and price discovery.

Chart: World Gold Council, Bloomberg, ICE Benchmark Administration. Data as of 26 June 2026, spot gold, 20-minute tick data. Sessions in New York time.

America sold. Asia bought.

The flows say the same thing the clock does.

Over the first half, American gold ETF investors shed 60.7 tonnes. Mainland Chinese funds took in 28.7. Indian funds took in 24.1.

The read. June was ugly everywhere: USD 8.9bn out, holdings down 74 tonnes. Easy to stop there. But H1 flows stayed positive at USD 8bn, holdings rose 18 tonnes, Europe took in USD 3.2bn, and Asia posted its strongest first half on record. North America was the only region that lost, its weakest half since 2013.

Contrarian flag. “Investors are abandoning gold” turns out to mean American investors are abandoning gold, not the entire world.

Chart: World Gold Council, Bloomberg, Company Filings, Incrementum AG

India sold the paper and kept the metal

Retail is one thing. Reserve managers another.

India’s holdings of US government debt fell from USD 232.5bn in April 2025 to USD 181.0bn in April 2026. Its gold sat at 881 tonnes, worth roughly USD 102.5bn.

The read. Note what the chart does not show. India did not buy a mountain of gold this year. It added two tonnes. The mountain was built earlier: 658 tonnes six years ago, 881 now. What moved was the other side of the balance sheet. Japan, the UK, Belgium, France and Canada all added Treasuries over the same stretch.

Watch for. Whether the next TIC print extends it. Treasury collects this from US custodians and warns it cannot attribute country ownership precisely. Read the direction, not the decimal.

Chart: US Treasury TIC data, Reserve Bank of India, Incrementum AG

The month gold fell hardest

China’s central bank added 14.93 tonnes of gold in June. Largest single-month purchase since 2023. Twentieth consecutive month of buying. Holdings now roughly 2,346 tonnes.

And yet, June was the month gold briefly broke below USD 4,000.

Chart: State Administration of Foreign Exchange, Incrementum AG

Read more

And what do you call that same company when it is simultaneously one of the biggest holders of US Treasuries on the planet?

You call it Tether. And you should probably start paying attention to it…

The world’s biggest stablecoin issuer bought 26 tonnes of bullion in Q4 2025 – a pace that, according to Jefferies, ranked third globally, behind only the central banks of Poland and Brazil. At its peak, Tether CEO Paolo Ardoino said the company was buying up to 2 tonnes of gold per week. That’s more than USD 1 billion of gold. Accumulated every single month. By a company most investors had never heard of a year ago…

After a further 6 metric tons of gold reserves were added to back USDT in Q1 2026, Tether’s golden hoard now stands at a whopping 154 tonnes. Stored in Swiss vaults that the company owns outright, the stash is worth over USD 20bn at today’s USD 4100/oz gold price. Tether’s Head of Special Projects, Juan Sartori, calls it “probably the biggest private reserve of physical gold in the world.” In fact, if Tether were a sovereign nation, the current stash would place it around the 27^th largest gold holder globally, comfortably ahead of the supposedly “developed” nations of Australia, Canada, and Sweden.

So, Tether should be championed as a positive force for gold investors, right?

Perhaps.

But before crowning it so, we ought to consider the other side of the ledger…

That is, while Tether is clearly a rampant accumulator of gold, its entire business is simultaneously centred around the US dollar. As of April, roughly USD 185bn of USDT was circulating worldwide. And behind almost every one of those digital dollars sits a real short-dated Treasury bill.

That not only makes Tether a crucial, price-insensitive buyer of US government paper, but also the 17th largest holder of US debt on the planet, ahead of US-allied nations such as Germany and South Korea.

Washington passed the GENIUS Act precisely to harness that demand from whales like Tether. And the carry trade has generated more than USD 10bn in profit in the past year alone, over USD 30bn since rates began rising in 2022.

So, which is it? Is Tether the dollar’s Trojan horse, wheeled deep inside the sound money citadel? Or the vanguard of the gold standard’s return? Can one company really be both?

Our sister publication, the In Gold We Trust report, devotes an illuminating chapter to that very question. “A Golden ‘Stabilisation’ Op in Plain Sight” was superbly written by Izabella Kaminska, the journalistic force behind The Peg, a new stablecoin-focused publication, and The Blind Spot. Previously senior finance editor at Politico and editor of the FT’s award-winning Alphaville blog, few writers are better equipped to decode what Tether is actually building. Her chapter includes an exclusive interview with Sartori, and it is the basis for everything that follows.

In today’s deep-dive, we extract the key takeaways from the chapter for SMR readers, as we seek to answer an increasingly critical question for sound money investors:

Are critics right to dismiss Tether as some kind of US government psyop built to prolong dollar dominance?

Or could it actually be an ingenious fiat off-ramp that is quietly paving the way for a gold-backed monetary revolution?

Read more

Then Thursday happened.

Warsh stepped to the mic in Sintra, Portugal and eased off the hikes.

June jobs came in soft. The dollar slid from 14-month highs. Gold snapped back above USD 4,130. Bitcoin gained almost 5% for the week. And silver staged a recovery from its washed-out base with a 6.6% weekly gain.

Last week, we took inventory of a quarter-off correction and found the trade intact. This week, the tape started to agree. Let’s mark it to market.

To hike or not to hike?

Sound Money Report, Co-Founder, Ronnie Stoeferle, mapped three Warsh outcomes last week.

His base case: no hikes, maybe cuts, as one-year inflation expectations collapse from 3.20% to below 2.20%, the lowest since Trump took office.

On Thursday, Warsh co-signed it in Sintra, pointing to eased expectations and a trimmed-mean PCE down for 36 straight months. The collapse we flagged a week ago just got a co-signer at the Fed.

Key Takeaway: The dip was priced for a hike cycle that the Fed chair just talked down. Take the scare out of the tape and gold reprices higher. It already started, up 2.27% on Thursday.

Watch for: Warsh’s decision (and language) at the upcoming FOMC meeting on July 29th. A hold confirms the base case and drains the last of the hike bid. A hike would inflict more damage on Uncle Sam’s haemorrhaging balance sheet.

Smart money is (still) sitting on the sidelines.

The above pie chart is the 2024 family office portfolio structure: gold and precious metals at 2%, level with art and antiques.

UBS’s 2026 report is the update. After a record 60% of these offices reworked their allocations, gold advanced exactly one point, to 3%.

For context, that marginal increase leaves family offices a long way off the optimal gold allocation for risk-adjusted returns.

Our sister publication, the In Gold We Trust report, calculated that the optimal allocation as 14-20%, while the World Gold Council’s framework suggests up to 10%.

And yet, that’s not even considering the laughably low <1% allocation to commodities across the board by family offices.

Ultimately, that means some of the biggest buyers on earth are barely allocated to our relatively unloved and overlooked sector.

Key Takeaway: Sound Money investors are still incredibly early, while the supposedly “smart money” falls asleep at the wheel of outdated 60/40 portfolios.

The retrace was the entry.

Tavi flags the silver-to-S&P; 500 ratio pulling back to its long-term uptrend after overextending, the same shakeout the early-2000s bull ran before its best years.

Translation: 6 months after its January all-time high of $121, Silver’s recent correction to ~$60 leaves it considerably more undervalued relative to the U.S stock market.

Key Takeaway: To us contrarians, silver’s relative cheapness versus the S&P500 is a buying opportunity rather than a bearish signal to shy away.

Watch for: Rotation out of capital out of big tech and into a“commodities supercycle” that lifts silver back towards triple digits.

Read more

The price performance was equally dramatic. Gold rose 66.5% in USD after already gaining a strong 26.4% in 2024. The rally extended into late January 2026 before moderating. Such sharp price increases inevitably reshape demand patterns. The following analysis examines how the structure of gold demand has evolved over the past decade – with particular focus on the transformative shifts seen in 2024 and 2025.

Long-Term Trends vs. Recent Reality

Since 2010, the linear trend for total gold demand (excluding OTC) has been remarkably flat. The major exception was 2020, when the first year of the COVID-19 pandemic triggered a sharp drop in demand of more than 15% – even as the gold price rose nearly 30%. Central bank purchases fell around 60% and jewellery fabrication by roughly 40% that year.

This was noteworthy because, over the longer term, there is typically a mildly positive correlation between gold demand and price. That dynamic was powerfully confirmed in 2025: total gold demand rose 8.1% while the average gold price climbed 43.8%.

Let’s have a closer look at the various subsectors.

Read more