AI has arrived in a very different environment.

The largest technology companies are committing hundreds of billions of dollars to GPUs, data centres, networking, cooling and power. Governments are issuing heavily as well, while utilities and infrastructure developers need fresh capital for grids, generation and transmission. Several large pools of demand are now leaning on the same financing system at once.

The size of the AI opportunity is already well understood. What matters now is the return earned on the capital required to build it.

AI can still be one of the strongest growth themes of the decade. The hardeest part is turning that growth into free cash flow while financing costs, depreciation and power requirements are all moving higher.

AI Is Becoming an Industrial Build-Out

The public conversation still presents AI mainly as software: better models, new applications, automation and productivity. The spending profile looks increasingly industrial. Compute has to be bought, housed, cooled, connected and powered before a model can generate revenue at scale.

The scale is easier to appreciate when AI spending is placed alongside previous investment cycles. The five largest infrastructure spenders - Meta, Google, Microsoft, Oracle and Amazon - accounted for about 30% of combined S&P 500 and Nasdaq 100 capital expenditure in December 2025. Consensus estimates put that share at 48% by December 2028. By S&P Global’s estimates, eight years of AI-era capital spending could exceed the previous 26 years combined.

It does make capital efficiency increasingly important. When investment grows this quickly, even small differences in utilisation, financing costs, hardware lives and power availability compound into very different shareholder outcomes.

Read more

Does it help identify companies whose outlook is improving? Can it distinguish a genuine upgrade from reassuring management language? And does the information it extracts tell us anything about returns that the market has not already priced in?

I use financial reasoning benchmarks as an initial screen for accuracy and cost. They help shortlist a model; the investment test comes next.

My starting point is a narrow research question: can changes in a company’s outlook improve a simple stock-selection process?

I want a rule that is transparent enough to explain, repeatable enough to test, and practical enough to trade after the information becomes available. The language model has a specific role within that process: extracting comparable forward guidance from earnings releases.

Earnings announcements offer a useful setting. They bring together reported numbers, management expectations and changes in the business. Comparing these across reporting periods gives us a way to turn narrative information into a measurable signal.

Here I develop the scoring rule, work through an example and specify the test I would use before risking capital.

Read more

One of the more intriguing examples came from Steven Edwards, who created 100 GPT-4o investor personas. They included value investors, growth investors, momentum traders, contrarians, quants and ESG specialists. Each selected 50 US-listed equities. Edwards aggregated their choices into a consensus portfolio and tested it over 561 trading days beginning in January 2024.

The market-capitalisation-weighted version returned 29.5% annualised, against 15.3% for the SPDR S&P 500 ETF. Even after controlling for six familiar market factors, the paper estimated annualised alpha of roughly 9.65%, with a t-statistic of 3.27. On the surface, the artificial committee had earned its fees.

Yet the most important result was not the return. It was what happened when 99 of the investors were dismissed.

Read more

The true signal emerges only when these forces are evaluated together: markets are hitting a dense catalyst window precisely as structural absorptive capacity is being tested. Central banks are shrinking balance sheets, reserve managers are diversifying, and private AI investment is competing directly with record government borrowing. Price-insensitive buyers are gone; the investors replacing them demand a concession.

This environment demands a new macro framework to diagnose which variable is actually in command. When inflation cools and Treasury auctions clear smoothly, markets easily digest heavy issuance. But when long-term yields climb while policy rate expectations remain anchored, the term premium is doing the damage—investors are demanding real compensation for fiscal supply risk. Conversely, when yields tumble alongside widening credit spreads, supply concerns fade and economic growth becomes the dominant threat.

The traditional playbook—buying duration simply because Fed easing is near—is dangerously outdated. Investors can no longer treat yield moves as a simple vote on monetary policy; they must identify whether supply, term premium, or growth is running the show to position risk correctly.

Read more

September puts that relationship to the test. AI is creating customers for chips, computing services and electricity infrastructure. But building that capacity requires large payments before all the benefits arrive. Governments also need to borrow, and investors must decide how much compensation they require to finance them. The interest rates that emerge affect both business expansion and the willingness to hold gold.

This helps explain an apparently contradictory market: gold can suffer when the Fed signals that rates may stay high or rise even while the reasons for owning it remain intact. A trader responding to next month’s interest-rate expectations and a central bank diversifying reserves over several years are making different decisions. The first may sell on a policy surprise while the second continues to accumulate.

Equities have their own version of that tension. A company may report excellent sales yet disappoint shareholders if spending rises faster, or if its share price already assumes even better results. Likewise, a rising S&P 500 does not tell us whether profits are improving across the economy. The largest companies have more influence on the index, so their strength can conceal weakness elsewhere.

The next four months provide a sequence of tests. September’s inflation figures and Fed meeting will shape financing expectations. Autumn earnings will show which businesses are converting demand into cash. November’s US election will change expectations for taxes, borrowing and energy policy. Iran-related disruption could alter the inflation picture throughout that period.

The useful question is therefore more precise than “gold or stocks?” Which companies can keep growing if money stays expensive? When does stress in government debt strengthen gold’s appeal? And what would need to improve before silver and the wider equity market become more attractive?In this article, I examine those questions through valuations, earnings, rates, positioning and the conditions that would change the outlook.

Read more

Goods became cheaper.
Technology improved efficiency.
Global trade expanded.
Interest rates trended lower.

It felt permanent because it lasted for decades.

From the 1990s through the early 2020s, the global economy benefited from an extraordinary combination of disinflationary forces. China entered global trade at scale. Companies moved production to lower-cost regions. Supply chains became leaner. Technology improved productivity. Capital became cheaper.

Consumers received more for less.

A better smartphone at a lower price.
Cheaper clothing.
Affordable electronics.
Faster delivery.
More choice.

For investors, the same environment created a remarkably durable playbook. Falling inflation supported lower interest rates. Lower rates lifted asset valuations. Globalisation protected margins. Cheap capital encouraged leverage and expansion.

That world is beginning to look increasingly historical.

The forces that drove prices lower for 30 years are weakening, and several of them are moving in reverse.

So will we be returning to the old regime at all?

Because if the structure of the economy has changed, the investment strategies built for the previous 30 years may become less reliable in the next 30.

The End of the Disinflation Era

Veteran macroeconomist Lacy Hunt has spent decades studying debt, inflation and long-term economic cycles. One of the most important ideas in his work is that the previous few decades were unusually favourable.

The integration of China into global trade created one of the strongest disinflationary forces in modern economic history.

Companies gained access to vast pools of lower-cost labour. Manufacturing migrated toward regions where production could be done more cheaply. Shipping networks became more efficient. Inventory systems became leaner. Global supply chains were optimised around cost.

For years, the direction of travel was clear.

Produce where it is cheapest.
Source globally.
Hold less inventory.
Use scale to drive costs down.

That model worked exceptionally well while the geopolitical environment remained stable.

Now stability can no longer be assumed.

Governments and corporations are reassessing how much risk they are willing to tolerate in exchange for lower production costs. Supply chains built around maximum efficiency are being redesigned around security, redundancy and strategic control

That shift has consequences.

A factory moved closer to home usually means higher wages.

A second supplier adds resilience, but also cost.

Larger inventories improve security, yet they tie up more capital.

New domestic production requires years of investment in infrastructure, labour and energy.

The old system minimised cost.

The emerging system pays more for resilience.

That is one of the clearest reasons inflation may prove more persistent than many investors expect.

The AI Paradox: A Digital Revolution Built on Physical Infrastructure

Artificial intelligence is often presented as the counterargument.

If AI can automate work, improve productivity and reduce inefficiency, surely technology will once again push costs lower.

Over time, that may well happen.

The near-term picture is more complicated.

The AI boom is extraordinarily physical.

Every large model depends on advanced chips. Those chips depend on specialised manufacturing facilities. Data centres require enormous quantities of electricity. Electricity demand requires grid investment. Cooling systems, transformers, copper, industrial equipment and construction all sit behind what appears to the end user as a purely digital product.

The front end is software.

The economic foundation is infrastructure.

This matters because the spending is happening now, while much of the productivity benefit may arrive later.

The transition therefore creates an unusual dynamic: a technology associated with future efficiency is currently intensifying demand for scarce real-world resources.

That can support higher capital expenditure, stronger commodity demand and additional pressure on energy systems.

The AI story may eventually become deflationary. However, its build-out phase is highly capital-intensive.

The Debt Problem Is Becoming Harder to Ignore

The next structural shift sits in government balance sheets.

The United States is running very large fiscal deficits even outside a recession or major war.

That changes the macro backdrop because interest expense itself becomes a growing source of pressure.

As debt rises, the government becomes more sensitive to interest rates.

As rates remain elevated, refinancing costs increase.

As financing costs increase, fiscal flexibility narrows.

This creates a system where debt management becomes increasingly intertwined with monetary policy, market liquidity and financial stability.

The size of the deficit matters.

The cost of servicing it matters even more.

A heavily indebted economy can function for a long time, especially when the underlying currency remains globally important. Yet the margin for error becomes smaller as debt service absorbs a growing share of national resources.

Investors therefore have to watch a broader set of variables than they did in the low-rate era.

Growth still matters. Inflation still matters.

But fiscal credibility, debt issuance and the long-term purchasing power of currency matter more than they used to.

The K-Shaped Economy: Why the Same Inflation Feels Different to Different People

One of the defining features of the current environment is the widening gap between asset owners and everyone else.

Inflation does not hit every household equally.

People who own equities, property, businesses or scarce assets can benefit when nominal asset prices rise.

People who hold most of their wealth in cash and wages face a different experience.

Their bank balance may remain stable while the cost of living moves higher.

That creates a K-shaped economy. The top branch benefits from asset ownership. The lower branch absorbs the rising cost of everyday life.

This is one reason financial markets can look strong while many households still feel poorer.

The stock market can rise. Property values can rise. Nominal wages can rise.

And purchasing power can still decline. Inflation works quietly.

It does not announce itself as a tax.

It simply reduces what each unit of money can buy over time.

That is why understanding the monetary regime matters far beyond macroeconomic theory. It changes how people should think about savings, portfolio construction and wealth preservation.

The Investment Playbook for a Different Monetary Regime

Up to this point, the argument is macroeconomic.

The more interesting question begins here.

What does this environment imply for investors?

If globalisation is becoming more expensive, AI is increasing physical capital demand, fiscal deficits remain large and inflation erodes purchasing power unevenly, then the traditional definition of “safe” starts to shift.

Cash looks stable in nominal terms.

Bonds can provide income.

Equities still benefit from growth.

Yet each of these assets behaves differently once inflation, fiscal pressure and currency debasement become more persistent features of the system.

The central idea is simple:

When Cash Stops Feeling Safe

Cash feels safe because its nominal value does not move.

$100,000 remains $100,000.

The problem appears when we measure wealth in terms of what that money can actually buy.

If inflation runs at 4%, purchasing power falls by roughly 18% over five years.

At 6%, the erosion is much faster.

This is the distinction between nominal wealth and real wealth.

Nominal wealth asks: “How much money do I have?”

Real wealth asks: ”What can that money buy?”

In a low-inflation environment, the difference is easy to ignore.

In a persistent inflation regime, however, it becomes central.

This does not make cash useless. Liquidity has enormous value. Cash provides optionality, stability and the ability to act when opportunities appear.

But investors need to stop treating a stable number in a bank account as the same thing as stable purchasing power.

Those are two different concepts.

Why Scarcity Matters More When Financial Claims Expand

Modern economies can create financial claims very quickly.

Governments issue debt. Banks create credit. Central banks expand their balance sheets. Companies issue securities.

The supply of financial assets can grow rapidly.

Physical scarcity moves at a different speed.

A new copper mine can take years to develop.

Energy infrastructure takes time to build.

Gold production expands slowly.

Industrial capacity cannot appear overnight.

And all of this creates an important asymmetry. Financial claims can expand much faster than the supply of scarce real assets.

When that happens, the relative value of scarcity rises.

This is one reason real assets tend to attract attention during periods of persistent inflation, fiscal expansion and monetary uncertainty.

The logic is straightforward.

If the denominator is weakening, scarce assets can reprice higher even without dramatic changes in their underlying physical characteristics.

Monetary Insurance in a Fiat World

I have written before about gold as monetary insurance: scarce, liquid, globally recognised and free of issuer risk. That case still stands.

What is more interesting now is that the market is beginning to behave differently from the framework investors have relied on for years.

Historically, gold and real yields had a fairly intuitive relationship. When inflation-adjusted bond yields fell, the opportunity cost of holding gold declined and the metal usually benefited. When real yields rose, gold often struggled.

That relationship has weakened.

Why? Demand function for gold is broadening.

Real yields still matter. But so do fiscal credibility, reserve diversification, geopolitical risk and central-bank demand. Gold is increasingly being priced not only against the return available on government bonds, but against confidence in the system those bonds represent.

That is a more consequential shift than another cyclical rally.

If gold can remain firm in an environment that would historically have been a headwind, the market may be telling us that investors are assigning more value to assets outside traditional financial claims.

Beyond the inflation, repricing is is also about trust, scarcity and the quality of the denominator.

The usual silver argument is also familiar: it has a monetary history like gold and an industrial use case tied to electronics, solar and electrification.

Silver can underperform for long stretches when investor demand is weak. But when monetary demand and industrial demand reinforce one another, the market can tighten quickly.

This is why I see silver less as a cheaper version of gold and more as a different expression of the same macro regime.

Gold is driven more directly by confidence in money and financial assets. Silver adds an additional sensitivity to the physical economy.

Real Assets Are Becoming Part of the Monetary Conversation

The investment case for real assets is broader than gold and silver.

Energy.

Industrial metals.

Infrastructure.

Commodity producers.

Real estate in the right locations.

Productive businesses with pricing power.

All of these assets interact differently with inflation. The common thread is that they are tied to something real.

Something scarce.

Something difficult to create instantly.

The economic regime appears to be shifting from an era dominated by abundance and efficiency toward one increasingly shaped by scarcity, resilience and capital intensity.

The previous 30 years rewarded investors for owning duration. The next 30 may reward investors more for owning scarcity.

The New Rules of Wealth Preservation

The old economic system was built around efficiency.

The old world benefited from cheap labour, cheap capital, abundant global supply and falling inflation.

The new world faces geopolitical fragmentation, higher fiscal deficits, energy constraints, capital-intensive technology and more competition for physical resources.

What retains value if the unit we measure wealth in keeps losing purchasing power?

It changes how you think about bonds.

It changes how you think about real assets.

The world has moved into a new phase.

The investment framework should move with it.

— Alina Khay

_{This is a research note and a record of my own market thinking, not investment advice or a personal recommendation. Nothing here considers your objectives, financial situation, or risk tolerance. Markets can move against this view; use your own judgment and seek professional advice where appropriate.}

When silver moved vertically higher, decades of price-sensitive metal suddenly became economical to sell. Jewelry, silverware, investment holdings and other scrap flowed back toward refiners. Recycling reached 197.6 million ounces in 2025, the highest level in 13 years, and the Silver Institute expects another 7% increase this year to roughly 211 million ounces. More interestingly, the flow became large enough to create refinery bottlenecks: some refiners extended settlement times and temporarily stopped accepting certain types of scrap. I would therefore describe the past several months as a supply-reset phase. The spike pulled dormant inventory into the market; the correction gave the system time to process and absorb it.

Higher prices have already produced a substantial supply response, while industrial demand is simultaneously being softened by silver thrifting and substitution, particularly in solar. Yet the market is still projected to run a sixth consecutive annual deficit in 2026.

This is why I find the current setup more interesting than the January squeeze itself. The first move was driven by scarcity recognition. Then came speculative acceleration and overshoot. What followed was supply mobilisation - exactly the mechanism markets use to respond to scarcity. Silver may now be entering the next phase: finding the price at which new supply, recycling and existing inventories can actually balance persistent demand. If that equilibrium exists near current levels, silver can spend a long time consolidating. If the market discovers that the supply response is still insufficient, the next repricing could become considerably more important than the first.

***

There are already small signs that silver is evolving as a financial market as well. CME will extend 24/7 trading to its new 100-ounce Silver futures contract from September 11, pending regulatory review, after seeing strong participation in its smaller gold contract. The product is financially settled against benchmark COMEX silver and has already averaged around 17,800 contracts a day during the first half of 2026. It is not a bullish signal by itself. But it is another indication that precious metals are becoming more continuously accessible macro assets, with a larger retail and global investor base able to react to news outside traditional market hours.

The market has already absorbed the price shock, the speculative overshoot and a major supply response. Yet structural tightness remains.

That leaves the real question:

If higher prices have already pulled more silver into the market - and supply still cannot fully close the gap - how high does price need to go to restore balance?

— Alina Khay

_{This is a research note and a record of my own market thinking, not investment advice or a personal recommendation. Nothing here considers your objectives, financial situation, or risk tolerance. Markets can move against this view; use your own judgment and seek professional advice where appropriate.}

The Market Setup

The August tension is blunt: gold is rising in the part of the cycle where the textbook said it should struggle, and it is doing so as the U.S. debt problem becomes harder to hide.

On 18 August, Treasury data put U.S. public debt at $40.05 trillion. Two days later, Treasury’s long-bond buyback plan became the market’s live stress test: the cap for some longer-dated operations was doubled from $2 billion to at least $4 billion per operation, according to The Wall Street Journal, while officials framed the program as liquidity support. The market read was less polite. A government with a $40 trillion debt stock was trying to take pressure off long borrowing costs.

Gold noticed. On 20 August, Comex gold settled at $4,516.30 per ounce, according to Dow Jones market data. By early European trading on 21 August, gold futures hit $4,600, heading for a weekly gain of roughly 4%. The tension is that U.S. real yields remain high by post-2008 standards. The Federal Reserve’s H.15 release dated 20 August showed the 10-year inflation-indexed Treasury yield at 2.35% on 19 August and the 10-year nominal Treasury yield at 4.65%. Real rates are interest rates after inflation and they matter for gold. When inflation-protected bonds offer a high positive yield, gold needs a stronger reason to rise - and right now that reason is inflation persistence, debt pressure, and fading confidence in policy control.

The old model still matters. It is just leaving too much money flow unexplained.

The Fed held the federal funds target range at 3.50% to 3.75% on 29 July. The minutes released on 19 August kept the inflation issue alive: many officials judged that additional firming could be needed if inflation failed to decline. July CPI, released by the Bureau of Labor Statistics on 12 August, cooled to 3.4% year on year, with core CPI at 2.5%, but the energy index was still up 14.7% from a year earlier.

For gold, those details matter. They define the setting in which the metal’s role changes. Treasury’s late-August buyback signal adds a new ingredient: the market is now watching the boundary between ordinary debt management and an official attempt to lean against long-end yields. That is where debt meets the gold price. When investors believe central banks and treasuries can control inflation, term premia, refinancing risk and currency risk at the same time, gold behaves like an expensive insurance policy. When investors doubt that control, gold starts to behave like a neutral reserve asset again.

A real-rate-driven rally is fragile. A debt-confidence rally can turn high-yield pullbacks or ETF weakness into accumulation windows.

Read more

The Market Setup

The headline volatility gauge and the stocks beneath it are sending very different messages. The size of that gap is unusual enough to change how investors should interpret risk, selection and portfolio exposure.

• What we can see: The index looks calm while individual stocks are making unusually large moves.

• What we need to know: Whether that gap represents a durable opportunity or an early warning.

The VIX traded near 15.3 on 11 August 2026. On its own, that number describes an ordinary market with some uncertainty, but no visible stress.

Yet beneath the index, the picture is far less ordinary. VIXEQ—the options market’s estimate of volatility across S&P 500 companies—closed at 39.64 on 7 August. In simple terms, options traders were pricing much larger moves in the average large-cap stock than in the index holding those same stocks.

This is already visible in daily returns. Dell gained 32% in one session after earnings. Hewlett Packard Enterprise rose 28%. Marvell gained 25% after a favourable comment from Nvidia’s chief executive. Large companies have been making moves normally associated with much smaller, more speculative stocks.

So why has the S&P 500 remained calm?

Because the stocks do not agree. An AI winner rises while a software incumbent falls. A strong earnings report lifts one company while a rate-sensitive sector weakens. The moves are large, but they occur in different directions and at different times. By the time they are combined inside the index, much of the volatility has cancelled out.

That makes the current market more interesting than the headline VIX suggests. The gap creates a wider field for stock selection and relative-value trades, but the gap alone cannot tell us whether conditions are constructive or fragile. That distinction is where the strategy begins.

Read more

Gold is rallying because the market has left itself very little room to be wrong.

That is the useful signal. A normal gold rally can be a fear trade. This one is more interesting because it is happening while investors remain committed to the growth story. The risk is concentration, not simply optimism.

The Bank of England’s July Financial Stability Report says AI-related companies now account for around half of S&P 500 market capitalisation, up from around a quarter in 2022. The Philadelphia Semiconductor Index was up 97% since the start of the year at the time of the report. BofA’s July fund manager survey also found long global semiconductors to be the world’s most crowded trade.

That does not make AI a bad theme. It means the index has less diversification than it appears to have. A wobble in AI capex, credit conditions, real rates, energy prices, or Fed credibility can travel through the whole benchmark.

Gold settled at $4,340.70 on Friday, August 7, its best week since January, after falling to a 2026 low near $3,985.60 in mid-July. The timing matters more than the level. Gold is rising at the moment when the equity market has become most dependent on a narrow version of good news.

So we ask the natural and simple question. If portfolios are already long AI, where does capital go when the growth story needs a hedge?

Read more

Markets have the exact same two variables, yet almost nobody separates them. Valuation is a slope-angle measurement. It tells you, with useful precision, how far the market falls once something breaks. It tells you close to nothing about whether anything is about to break, and it never has.

This matters right now because the slope is close to the steepest ever recorded. The cyclically adjusted price-earnings ratio (CAPE), which compares today’s index level to the last decade of inflation-adjusted profits, hit 41.4 in July 2026 against a long-run average since 1881 of 17.8. Only December 1999 was higher. Twenty-four stocks now account for more than half the S&P 500’s value, a narrower base than the dot-com peak.

The same market that prices S&P 500 volatility at roughly 16, which is to say “we expect nothing much to happen,” is simultaneously pricing insurance against Nvidia defaulting at a record 82 basis points, and against Oracle at 215 after S&P Global cut it to the lowest investment-grade rung on 9 July. Two markets, looking at the same companies, in the same week, reaching opposite conclusions about how fragile they are.Let’s stop asking whether stocks are expensive; that question has been answered repeatedly while the index continued rising. The more useful question is what releases a market priced this way, and how deep the run-out zone becomes when it does.

Read more

But this hawkish repricing misidentifies the regime. In a system operating under structural fiscal dominance, central banks lack the balance-sheet clearance to hike rates without destabilizing the sovereign debt model itself.

Markets are confusing hawkish rhetoric with operational policy capacity. While futures price higher terminal rates to combat supply-side commodity pressures, annual U.S. national debt service costs have breached $1.0 trillion—consuming roughly 19 cents of every single tax dollar collected. Every basis point of nominal tightening accelerates government deficits, compounding federal interest burdens and forcing the creation of secondary liquidity to absorb continuous Treasury issuance.

The short-term yield spike suppressing hard assets the terminal mispricing of a trapped central bank.

Read more

Markets rarely break in a single, clean sweep. Instead, they shift from a regime of unconstrained discovery to one of aggressive liquidity rationing. While the broader index might look deceptively flat and volatility appears contained, the tape has been pricing in a far less comfortable reality since early July.

Three superficially isolated headlines this month actually describe a single, interconnected process:

• The $3.2 Trillion Reallocation: An unprecedented rotation where capital has fled highly crowded semiconductor stocks, pulling out of over-owned winners to seek shelter in the rest of the S&P 500 and cash-rich mega-caps.

• The $SPCX Fracture: Weeks after the largest public debut on record, SpaceX ($SPCX$) slipped below its $135 fixed IPO floor to touch an intraday low of $132.15—a stark signal that the market’s capacity to absorb massive new supply is hitting a wall.

• The CapEx Arms Race: Taiwan Semiconductor Manufacturing Co. (TSMC) lifted its capital expenditure forecast to an eye-watering $60B–$64B to keep building out AI infrastructure, raising the high-stakes question of who will ultimately fund the return on this massive cash burn.

Separately, these financial stories look as idiosyncratic events. Collectively, they reveal that the expansion regime is over, and a classic zero-sum game has begun.

When the pool of capital stops growing and hits a fixed boundary, laggards can only catch a bid if fresh assets are cannibalized from somewhere else. To navigate what comes next, we must look past daily index noise and analyze the dangerous internal cycling of a secular bull market starting to exhaust itself.

Read more

The numbers refuse to cooperate. Passive vehicles pulled in huge money through 2025 and the first half of 2026, often at trillion-dollar annualized clips. Large-cap active funds mostly lagged, with roughly 79% underperforming the S&P 500 in key readings. The pattern showed up across assets. Gold had real structural support from central banks and Asian buyers, yet ETF flows and price action created swings that played out differently for systematic versus discretionary approaches.

Markets do not get harder because opportunities dry up. They get harder because more people see the same ones, and the table they sit at has changed.

The central question is how this filtering process actually works and what it means for extracting edge in equities, rates, gold, and the broader tape.

Read more

Real yields set the swings; central banks set the floor. Confusing the two is expensive.

For two decades gold behaved like a mirror held up to real interest rates: yields up, gold down, with metronomic reliability. Around 2024 the mirror cracked. Gold returned some 65% in 2025, its best year since 1979, in the same twelve months that ten-year American real yields averaged close to 2%, their highest since 2007. The textbooks say these two things cannot happen together; rising real rates are supposed to be gold’s natural predator. Anyone still pricing the metal off the TIPS curve has been reading the wrong instrument.

The correction sharpens the question rather than settling it. Gold touched roughly $5,405 an ounce on the LBMA benchmark in January 2026, then gave back a hefty slice of the gain, trading near $4,170 by early July. Now that the momentum crowd has left the building, the awkward question is being asked out loud: what, exactly, is holding the price up?

Most trading desks still price gold off real rates, its old cyclical driver, and keep missing the level—because the buyer who now sets the price at the margin does not appear in that model at all. That gap is the argument of this piece. Sitting on the wrong side of it has been an expensive way to be theoretically correct.

There is a tidy way to take the metal apart: separate the part that sets the floor from the part that sets the amplitude. Do that, and 2025’s melt-up and 2026’s stumble stop looking like a contradiction. They run on different clocks—and most of the industry is still checking its watch against the fast one.

In fairness to the crowd

The real-rate model earned its reputation the honest way: it worked. From 2004 to 2021, gold traded like a very long-duration inflation-linked bond that had simply forgotten to pay a coupon. Regressed against the ten-year American real yield over that window, it carried an empirical duration of roughly 18 years: a 100-basis-point rise in real yields knocked about 18% off the inflation-adjusted price. The long-run correlation sat near minus 0.8. For the better part of a generation, knowing the TIPS yield was as good as knowing gold’s direction.

Then the relationship broke, and not politely. Through 2025, real yields sat at an eighteen-year high while gold rose by two-thirds. The model did not merely miss its forecast, it pointed the wrong way. What replaced it is a reserve-allocation story: a slow, structural, and dull grind—right up until the moment it shifts.

The buyer changed, the model confirmed

Prices are set at the margin, by whoever trades next, not by whoever holds most. From 2010 to 2021 the marginal buyer was Western ETF money, and ETF money is a pure real-rate machine. It holds a yieldless metal only when doing so is cheap: it buys as real yields fall and sells as they rise. That is why the TIPS relationship looked so clean for so long. The model was never describing gold. It was describing the buyer.

From 2022 the marginal buyer became the official sector, and central banks do not keep an opportunity-cost spreadsheet. They buy gold for reserve diversification and sanctions insurance, not for carry. The trigger is specific and dateable. Freezing Russia’s central-bank reserves in 2022 proved that dollar assets parked offshore can simply be switched off, and every reserve manager not aligned with Washington took note. Gold, alone among reserve assets, carries no counterparty and no issuer: nobody’s promise, nobody’s balance-sheet. Basel III then made the point official, reclassifying gold as a Tier 1 asset alongside cash and government bonds.

The scale settles any doubt. Between 2021 and 2025 central banks bought around 225 tonnes a quarter, roughly double the pace of the previous five years. Each of 2022, 2023 and 2024 cleared 1,000 tonnes; 2025 added another 863, the fourth-largest annual total on record. Monthly estimates put official buying near 60 tonnes, against a pre-2022 average closer to 17. Central banks’ share of total gold demand has climbed from about a tenth in the 2010s to roughly a fifth today. A buyer that big, that steady and that indifferent to price does not move the price around. It moves the floor underneath it.

Two clocks, one price

Gold’s price is best read as the sum of two things—a structural floor and a cyclical amplitude—running on two different clocks.

The structural floor is set by the price-inelastic buyer, the official sector, plus a slow debasement updraft off sovereign balance-sheets. It moves in years and could not care less what the TIPS yield did this week. The cyclical amplitude is set by the price-sensitive buyer: Western capital reacting to real rates, ETF flows and the dollar. It moves in weeks, and it moves loudly. The real-rate model reads the amplitude with precision. It is simply blind to the floor beneath it.

The interesting part is what happens when the clocks interact, because the arithmetic turns non-linear fast. Most of the time they simply add: the floor rises quietly while the amplitude swings the price around it.

When they align—real yields falling while the reserve bid runs hot—the effect multiplies , and you get January’s near-vertical run to $5,400. When they diverge—real yields rising against a floor that refuses to budge—you do not get a crash. You get a cushioned correction, because the inelastic buyer quietly absorbs what the elastic buyer sells. That asymmetry is the tell: gold rises hard when real yields fall and gives back comparatively little when they rise. The old relationship has gone convex.

The floor shows up in the reserve data

The best evidence for the floor sits in the reserve data. On the ECB’s own analysis, gold has overtaken the euro to become the world’s second-largest reserve asset—a first since Bretton Woods ended. By end-2024 gold made up close to 20% of global official reserves, against roughly 16% for the euro and about 47% for the dollar. Rewind to 2018 and gold was 11.6% of reserves while the dollar commanded 54.6%. The reallocation moves in a straight line with a clear direction: gold is taking the share the dollar is giving up.

That is the whole case for durability, in one fact. A new Fed chair, or one hawkish surprise on rates, can crush the amplitude overnight. It cannot reverse a decision taken independently by more than thirty emerging-market central banks, all rebalancing away from the dollar since 2022 for reasons that have not gone away. The floor moves slowly because it is being poured by institutions that have never once looked at the TIPS chart.

The cross-asset effect of Gold

This is where the framework earns its keep: the same force turns up in 4 markets at once.

The Bond Market (Rates)

Global debt exploded to a staggering $340 trillion by mid-2025, and governments own a record 30% of it. Investors are getting terrified of holding long-term government debt because countries are drowning in deficits. As a result, they are demanding higher premiums to hold bonds, while quietly bidding up gold.

The Currency Market (Foreign Exchange)

Gold and the U.S. dollar are on a seesaw—when the dollar’s grip slips, gold wins. Central banks aren’t dumping the Euro but they dump Dollar, and gold is catching the fallout. Even better for gold bugs: math models show the U.S. dollar index is still 9% overvalued. As the dollar gradually adjusts toward its fair value, it is poised to give dollar-priced gold a natural, long-term boost.

The Stock Market (Equities)

American stock-bond correlations climbed to thirty-year highs through the post-Covid inflation shock, so the bond leg of a standard 60/40 portfolio has stopped hedging the equity leg. When your insurance policy stops paying out, you shop for another, and a metal with no counterparty is the obvious quote. Some of the institutional flow into gold is simply the fallout from bonds failing to diversify. With a record $7.5 trillion parked in money-market funds, there is no shortage of dry powder for that rotation, and it is released as cash yields fall.

The Metals Market (Commodities)

In commodities, the breadth confirms the story but is easy to misread. Silver and platinum joining the rally says this is a monetary repricing, not a gold-specific quirk. But silver is the high-beta amplitude play and gold is the floor; treat them as interchangeable and a leveraged book gets hurt the moment the cyclical clock turns. The metals rhyme. They do not share a floor.

What does AI do to gold?

The obvious objection is AI. If it delivers the productivity boom its boosters promise, the equilibrium real rate rises and so does the cost of holding a yieldless metal. History has run this experiment: the 1990s boom, with its high real rates, strong dollar and heavy official selling, ground gold down to roughly $252 by mid-1999. On the naive chain—AI equals technology equals risk-on—gold is the one asset the future forgot to invite.

Sort the channels onto the two clocks, though, and the picture changes. Every bearish vector sits on the amplitude side: a higher natural rate, disinflation as AI compresses the cost of cognition, capital crowding into the earnings story. Powerful, cyclical—and already in the real-rate model.

The bullish vectors sit on the floor, and they are heavier. The AI buildout—data centres, grid, power—is one of the largest capex waves in history, and it is inflationary while it is being built. Gold mining is energy-hungry, so AI’s appetite for power lifts gold’s own cost of production. AI rivalry hardens the US–China split and speeds up the very reserve diversification this piece is built on. And if AI displaces labour at scale, the tax base shrinks while claims on the state grow. A sovereign at 100–120% debt-to-GDP cannot absorb that with orthodox policy; the path of least resistance is financial repression—negative real yields, on purpose.

That last vector inverts the 1990s comparison, which assumes a government balance-sheet that no longer exists. AI may raise the natural real rate while fiscal arithmetic caps the realised one—and gold trades on the realised rate. The question is whether long real yields track r* or get pinned below it. Whether AI succeeds is almost beside the point.

Add it up and gold is a strangle on AI outcomes: short the middle, long both wings.

The bearish case needs one scenario—the goldilocks middle, in which AI delivers steadily, fiscal pressure eases and geopolitics behaves. The bullish case owns both tails: AI disappoints (the concentration unwinds, cuts are forced through, real rates collapse) or succeeds too well (labour displacement, debasement, escalation). And unlike a strangle bought in the market, this one is paid to wait.

Sequencing matters too. The real-rate headwind is cyclical and arrives early; the debasement and trust tailwinds are structural and compound. One channel has no 1990s precedent: AI degrades the reliability of information itself. In a world of synthetic media and machine-written narratives, the premium rises on the one asset with nothing to verify. Bitcoin competes for that premium but drags its own AI-adjacent tail risk behind it—quantum computing threatens its cryptography. A lump of metal is unbothered.

The position

The upshot: price gold as floor plus amplitude, rather than a single real-rate trade. My odds, not advice: roughly 70–75% that the floor holds above $4,000 over the next twelve months, whatever real rates do. The amplitude—sideways in $4,300–$4,700 or a re-test of $5,400—is close to a coin-toss on the Fed. Conviction on the floor; none on the amplitude. On AI, I weight the debasement branch above the 1990s replay, which makes the strangle modestly constructive, with positive skew.

The downside belongs on the record; a position without one is just a forecast in disguise. If the floor gives way, official demand was more price-sensitive than the thesis assumed, the correction runs toward $3,600–$3,800, and the two-clock model needs rebuilding. That is the bill I would owe.

What confirms the thesis and what kills it, with levels:

• Confirms: the floor holds in the $4,000–$4,300 band through this drawdown, even when price action looks soft.

• Confirms: OTC and Swiss-refinery estimates of official demand stay above roughly 200 tonnes a quarter, and Chinese net imports stay elevated.

• Confirms: the next reserve-composition print shows gold taking another step of share from the dollar.

• Confirms (AI regime): realised long real yields lag rising r* estimates—repression by another name.

• Confirms (AI regime): electricity PPI runs hot, and term premium rises during equity drawdowns rather than falling.

• Invalidates: a weekly close below $4,000 that holds. The floor was thinner than the reserve data implied.

• Invalidates: two consecutive quarters of net official-sector selling on the OTC-adjusted estimate. The marginal buyer stepping back is the one thing the thesis cannot survive.

• Invalidates (AI regime): realised long real yields tracking r* upward for several quarters with no fiscal strain in term premium—the goldilocks middle asserting itself.

The cleanest tell is estimated official flow—specifically the gap between the reported number and the estimated one. In the first quarter of 2026, reported central-bank buying looked to be cooling, with one large seller dumping 60 tonnes. The OTC and refinery-derived estimate showed buying rising, to around 244 tonnes from 208. The reported line arrives late, incomplete and sometimes pointing the wrong way. Watch the estimate.

Step back and the regime is the real subject. The risk-free asset is being re-rated in slow motion; stocks and bonds fall together often enough to impair the 60/40 hedge; the dollar’s privilege erodes at the margin. In that world gold is still the hedge against the monetary base itself, a structural role the old crisis-alarm-bell framing misses entirely.

For twenty years gold answered to the real-yield chart. It still reads it, it has simply stopped taking orders from it, because the buyer who now sets the price never took orders from it either. Gold is not a bet on AI. It is the hedge against every world in which AI and the fiscal state cannot both be right.

— Alina Khay

_{This is a research note and a record of my own market thinking, not investment advice or a personal recommendation. Nothing here considers your objectives, financial situation, or risk tolerance. Markets can move against this view; use your own judgment and seek professional advice where appropriate.}

The second time was two weeks ago, and it is why this piece exists, written to hand you a reusable test for any crowded trade rather than as a victory lap or a mea culpa. The midyear rerun shows the allocation gap widened past 30 percentage points, its most stretched reading since November 2007 - the month large banks began writing down the credit losses that became the financial crisis. At the same time, the rally presented as proof that markets have recovered is largely confined to the US. Even within the US, performance has been concentrated in a small group of high-earning companies, while most non-US equity markets fell over the same six months.

This article opens the process, showing how crowding, historical precedent, and exit conditions are translated into explicit probabilities, position sizing, and predefined conditions that would invalidate the thesis.

The consensus and I are looking at the same market. We disagree because we are asking different questions. The consensus is trying to forecast earnings. I am trying to forecast how prices behave when investors have already positioned for those earnings. Those are fundamentally different problems, and at the moment they lead to very different conclusions.

The output odds in uncertainty

A forecast without error bars is just a marketing claim. This is the method’s current output, expressed in its only honest form: as probabilities, recorded so it can be tested against reality. Roughly 55 to 60 per cent on bonds outperforming equities over the coming twelve months, 25 to 30 on the melt-up extending as earnings keep outrunning the discount rate, and the remainder on a muddle where neither side wins by much. Those are judgment-weighted probabilities. What follows from odds like these is not “sell everything”. It is a tilt: equity exposure trimmed back toward neutral against a strategic benchmark, the released risk parked where it is paid to wait, and convexity owned against both tails.

The process that produces those numbers has five stages. Here is each one, run live on July 2026.

Stage one: measure the crowding

Stage one asks a single question of the custody data: is the crowd’s position growing or shrinking relative to its own history? Institutional investors did not use the rally to trim their record equity overweight. They, in fact, added to it. The holdings data shows the equity-minus-bond allocation gap rising from just over 28 percentage points in January to more than 30 by midyear, against a 25-year average near 20. The March war shock briefly cut equity allocations by 1.6 points, the sharpest monthly de-risking since August 2023. April reversed it with a 2.1 point increase, one of the largest single-month jumps in a data history reaching back to 1998. June’s buying was funded by drawing down cash, with bond holdings untouched.

Read more

Most of Wall Street reads that as a warning sign: a market this narrow, the argument goes, is one bad earnings report away from taking the whole index down with it.

That’s the wrong read, and the data from the last few weeks backs it up. Extreme concentration like this has happened before, and it has rarely been the thing that actually breaks a market. What breaks markets is a recession or a rate shock landing at the same time as the concentration peak. Neither has happened. What has happened is quieter and, I think, more important: a rotation. Money has been moving out of the most expensive AI-linked names and into everything else, without the index falling apart, and the market’s own correlation data confirms it, not a forecast, a measurement, already visible as of this week.

Here’s the part most commentary is missing. A 40.7% concentration figure tells you nothing about who owns those stocks or how fast each type of owner would sell if conditions changed. That gap, between the money that sells first and the money that sells last, is what decides whether Q3 turns into a healthy rotation or something worse. I’ll show you the mechanism, the data behind it, and the exact level that would prove me wrong.

Four things matter for Q3. Everything else, daily capex headlines, the latest hot take comparing Nvidia to Cisco, single stock target changes, is noise sitting on top of these: how mega-cap growth is being financed, why the Fed can’t lean against it even if it wanted to, a shift in real interest rates hitting every long-duration asset at once, and the passive money quietly working to keep the index looking calm regardless of what’s happening underneath. I’ll take these one at a time, then get to the position and what would prove it wrong.

Read more

Why now

The disinflation victory lap

The cause is obvious. The Fed held, scrapped its easing bias and floated a hike, sending 2-year yields to a one-year high. Oil’s collapse is doing the Fed’s job on inflation, the Iran premium has unwound, and Treasury Secretary Bessent is out narrating a tidy world of strong dollar, tamed inflation, 3% growth and AI-doubled productivity — all at once. Higher real yields and a firm dollar are gold’s natural kryptonite, so the move is coherent. It is also extremely crowded.

What’s actually going up

Here’s the one line printing fresh records while everything real bleeds: borrowed money. Margin debt just hit an all-time high $1.3 trillion, roughly 5.2% of GDP. Leveraged ETF assets sit at a record $198bn — the 99.9th percentile in history — with ~$40bn in 3x Nasdaq and ~$35bn in 3x semis. Nothing says “sound fundamentals” like a record margin balance propping up a record-crowded long.

The crack under the floor

And the buyer of last resort is tapped out. US real disposable income just turned negative year-over-year: households have less real money to spend, courtesy of the inflation that already happened — even as the forward narrative pivots to disinflation. A levered-up market resting on a shrinking consumer is not a sturdy structure.

The setup is everywhere. The trade isn't. So why this looks like a washout rather than a top, and the one signal that's flipping?

Read more

The verdict is wrong, or at least it is answering the wrong question. What the past five months killed was the monetary engine of the trade. The reason the trade exists is still running, and it does not report to the Fed.

That distinction is the entire argument, and almost nobody is making it. On 10 June the Bureau of Labor Statistics reported May CPI at 4.2% year over year, the hottest print since early 2023, with energy responsible for more than half the monthly gain. A week later Warsh’s first FOMC left rates at 3.50 to 3.75% but did something louder than a move: the Summary of Economic Projections lifted the 2026 PCE forecast to 3.6% from 2.7% and quietly retired the easing bias. Futures that began the year pricing two cuts now price a roughly two-thirds chance of a hike before December. The market spent 2026 asking when the Fed would rescue risk assets. It has its answer, and the answer is a cap.

Gold fell because real yields rose and the dollar firmed. That is the textbook response of a non-yielding asset to a hawkish repricing, and it tells you nothing about whether the case for owning it has weakened. The crowd has confused the Fed’s hand on the price with a verdict on the thesis. They are not the same thing. One is a rate story playing out over months. The other is a balance-sheet story playing out over a decade, and it is the one that put gold near $5,600 in the first place.

So the question that matters is not whether Warsh is hawkish. He’ll be hawkish or dovish enough for the headlines. The question is which engine of the trade he can actually reach, and which one is accelerating while he is not looking.

Read more

Before dismissing them: momentum is genuinely crowded in some asset classes like commodity trend-following, FX carry. It is decaying in others - volatility-targeting strategies in 2022. But in the one market where institutional momentum capital is deepest - U.S. equities - neither explanation holds. The crowding is unchanged from 2010, the decay is selective, not universal. This suggests the disease isn't momentum itself.

New research on three decades of VIX behaviour has an answer. It’s not the direction of volatility that broke momentum. It’s the tempo. And once you see how the two halves of that finding compound each other, the underperformance stops looking like a mystery and starts looking like arithmetic.

The mechanism is specific, the data is clean, and the position it implies is actionable. Here it is.

The consensus read is wrong in the right direction

Crowding doesn’t hold up.

Read more