Goldman Says Rate Hikes Won't Stop the Bull Market. The Ledger Has a Different Timeline.

Three numbers occupied the same page this week, and at least two of them cannot both be true.

The S&P 500 sits within 2% of its all-time high. Its forward price-to-earnings multiple has been compressed from 22x to 19x. And the market has already priced more than three rate increases into next year.

A 14% haircut to the multiple. An index that refuses to break. A tightening cycle that began on September 14 and that the consensus is treating as weather rather than climate. Hold those three facts in the same frame and something odd appears: the price of equities is being defended by earnings while the price of money is being raised by policy, and the two are running a race that the commentary has already decided is over.

Goldman Sachs' strategy desk, under Ben Snider, read the arrangement as benign — rate hikes pressure valuations, but corporate earnings matter more, and earnings are strong. That is a defensible sentence. It is also far narrower than the headline built on top of it. "Rate hikes will not halt the bull market" is not the same as "rate hikes are irrelevant," and the note itself concedes the distinction: the first three months of a hiking cycle have historically cost the S&P about 2%, and the twelve months after a first hike have delivered about 9%.

I don't trade the S&P. I read ledgers. So I took the thesis to the chain and asked whether it survives contact with real capital. What I found is not a refutation of Goldman. It is something more useful and more uncomfortable: a crypto market that has already priced the rate path, and has quietly repriced something else entirely.

Context

To evaluate the claim, you have to separate it from its packaging.

The Goldman case rests on four load-bearing assumptions, only one of which is genuinely new. First, valuation compression has already occurred — the forward multiple fell from 22x to 19x, so a portion of the rate shock is in the price. Second, earnings dominate equity returns across any twelve-month window; the discount rate bites at the margin, not at the center. Third, the historical record of seven prior hiking cycles has a consistent shape — short-term drawdown, twelve-month recovery — and that shape is being used as a forecast. Fourth, the market has pre-priced three or more hikes next year, so the marginal surprise is smaller than the average commentator assumes.

Notice what is absent. No inflation path. No terminal rate. No balance-sheet policy. No dollar view. No credit-spread analysis. This is a valuation-and-earnings note wearing a macro coat. That is legitimate — and it is also a boundary. Goldman is arguing that the numerator can outrun the denominator. It is not telling you where the denominator stops.

The distinction between "priced in" and "priced through" deserves its own sentence, because the entire bull case rests on it. Markets price expectations, not events. If three hikes sit in the curve, then three hikes will not move the curve; a fourth hike, a hawkish dot plot, or a terminal-rate revision will. That is what "priced in" means. "Priced through" is different: it means the market has already allocated capital as though the hikes happened, and every marginal position has been sized to survive them. The Goldman thesis implicitly assumes the first. It needs the second to be true for the conclusion to hold. The second is an empirical question about positioning — exactly the sort of question a ledger can answer and a valuation multiple cannot.

There is also a definitional slippage worth flagging. "Bull market" is not a policy statement; it is a label applied after the fact to a period of rising prices. When a strategist says rate hikes will not halt the bull market, they are making a claim about the persistence of a trend, not about the level of valuations. A market can fall 8%, recover, and still be called a bull market. That latitude makes the headline nearly unfalsifiable — and unfalsifiable claims are the ones that get repeated, because they can never lose. I prefer claims that can lose. So I will state mine precisely and let the chain grade it.

That question matters because crypto is the purest expression of the denominator. Every dollar of leverage in this market is priced off a discount rate, and that discount rate now anchors to a Fed that has begun to tighten. Equities have earnings to hide behind; crypto has liquidity — and liquidity is a function of rates in a way that earnings are not. When you hear that rate hikes don't matter for crypto, you are hearing someone describe a market they have never priced.

Historical averages are seductive because they compress variance into a single number. Seven hiking cycles produced an average drawdown of 2% in the first three months and an average gain of 9% over the following year. An average is a statement about the middle of a distribution, and the middle is where you end up if nothing breaks. The whole job of an analyst is to ask what breaks. Goldman's average assumes no recession inside the twelve-month window; it does not say so, but that assumption is the entire load-bearing wall. Remove it and the note's own logic inverts, because a 19x multiple on falling earnings is more expensive — not less — than a 22x multiple on rising ones.

My methodology for testing the thesis is deliberately unsentimental. I pulled three public data layers: stablecoin net issuance, as a proxy for dollars entering or leaving the system; the borrow-rate curves on the two largest on-chain credit markets, which function as crypto's risk-free rate; and wallet-cohort flows around the previous three FOMC windows. I cross-checked each against cohorts I have tracked since 2017 — the wallets that, in the ICO era, first taught me that the ledger never forgets a promise it failed to keep.

Then I ran the same stress test I ran in 2022, when I mapped ten lending protocols and surfaced $2 billion in positions their marketing pages insisted were overcollateralized. The chain did not care what the marketing pages said. It never does.

The dollar that stays in the casino

Start with the most honest signal on-chain: stablecoins.

Stablecoin supply is the closest thing crypto has to a capitalization-weighted measure of risk appetite, because a stablecoin is a dollar that has chosen to remain inside the casino rather than leave it. When Treasury yields sit near zero, holding a dollar-pegged token is a costless bet on optionality — you forgo nothing to keep your seat. When the risk-free rate climbs, that same dollar acquires a genuine alternative, and the cost of parking it on-chain rises with every basis point the Fed adds. The stablecoin is not a neutral instrument. It is a thermometer for the price of patience.

Not all stablecoins send the same signal, and collapsing them into a single line is a common error. Fiat-backed tokens expand when custody banks receive deposits; their supply measures institutional intent as much as retail appetite. Crypto-backed tokens expand only when someone is willing to post volatile collateral against a loan, which makes their supply a direct read on leverage appetite. When fiat-backed supply is flat while crypto-backed supply grows, the market is not receiving new money — it is manufacturing synthetic dollars out of existing positions. That is the difference between a funded rally and a geared one, and it is visible only if you keep the categories separate.

This is the transmission channel the sell-side keeps underestimating. Rate hikes do not hit crypto primarily through the discounting of future cash flows, because most crypto assets have no cash flows to discount. They hit through the opportunity cost of the dollar. Every basis point of additional yield on a Treasury bill is a basis point a stablecoin holder is voluntarily forgoing. When that spread is twenty basis points, nobody moves. When it is four hundred, the composition of on-chain capital changes whether or not the price chart notices.

What the data shows is not an exodus. It is a rotation — and rotation is harder to see than flight. Aggregate stablecoin supply has not collapsed; it has stopped growing at the pace the price action would imply. That divergence is the tell. If the bull market were being driven by new dollars entering the system, stablecoin supply would expand in lockstep with price. It isn't. The dollars are already inside; what changed is leverage, not deposits.

I have seen this exact fingerprint once before, and it ended badly. In 2017 I manually tracked 15,000 wallet addresses tied to the ten largest ICO projects, and what sat underneath the euphoria was not a flood of new capital but a small cluster of wallets recycling the same ETH through coordinated movements. Twelve distinct bot clusters, each moving in near-lockstep. The price went up. The deposits did not. Where early ICO ghosts still haunt the ledger, they haunt it with this exact signature: rising prices on flat or falling net inflows.

The 2026 version is subtler, because the market is more sophisticated and the instruments are better engineered. There are treasury-management strategies, wrapped yield products, off-chain credit facilities that settle on-chain. Each adds a layer between the price and the dollar. But the arithmetic has not changed. A rally funded by leverage is a rally funded by borrowed confidence, and borrowed confidence has a maturity date. The stablecoin data is where you read that date.

Crypto's risk-free rate is now a Fed instrument

The two largest on-chain credit markets quote a borrow rate that functions as crypto's risk-free rate. In 2020, that rate was set almost entirely by internal supply and demand — when DeFi Summer inflated yields, capital flooded in and the rate equilibrated. That world is gone. Today the on-chain borrow rate is bounded below by the cost of the dollar itself, and that floor is set in Washington.

Consider the implication. If you can borrow a dollar on-chain at 3% and earn 5% on a short-dated Treasury, you hold a riskless spread — and the rational move is to withdraw liquidity from on-chain markets until the on-chain rate rises to meet the Treasury rate. No opinion required. No thesis required. Rate hikes don't need to scare crypto investors to drain crypto liquidity. They only need to make leaving the table profitable.

I built the first version of this model in 2020, when I scripted an analysis of 500 million tokens swapped on Ethereum mainnet and found that roughly 30% of liquidity was supplied by arbitrage bots rather than long-term holders. I published it as "The Bot Economy," and the conclusion that got quoted was the least interesting part — that the bots were not a bug but the market's immune system. The important finding was that their behavior was mechanical. They did not hold views. They responded to spreads, and they responded faster than any human could.

That machinery now reacts to rate spreads. When the Fed tightens, the gap between the on-chain risk-free rate and the sovereign risk-free rate widens in favor of the sovereign, and the bots migrate. You can watch it happen in the utilization curves weeks before it reaches the price. Utilization rises; borrow rates spike; leveraged positions get repriced; and the first domino is always a wallet that never read a Goldman note.

The mechanics of that draw are worth tracing once, because they explain why tightening bites through the plumbing rather than the price. A leveraged position is a loan against collateral. The loan carries a rate; the collateral carries a value. When the loan rate rises, the carry cost rises, and the holder must either post more collateral or unwind. When the risk-free alternative rises faster than the position's expected return, the unwind becomes rational even if the position is healthy. No liquidation engine is required. No oracle failure is required. The position simply closes because it stopped paying to stay open. In 2020 that dynamic was dormant. In 2026 it is live, and it will express itself as a slow bleed in open interest before it expresses itself anywhere else.

This is why I distrust every "this time is different" framing a bull market generates. The mechanism is not different. The participants are more professional, the collateral is better constructed, and the leverage is more transparent — but transparent leverage is still leverage. Transparency does not reduce risk; it only tells you where the risk will surface. In a zero-rate world, that surface is invisible. In a tightening cycle, it is lit.

There is a further wrinkle almost nobody models: the collateral itself is rate-sensitive. A large share of on-chain credit is collateralized by assets whose own yields are set by protocol emissions, not by cash flows. When the sovereign risk-free rate rises, those emissions must rise to stay competitive — which means the collateral's value depends on the protocol's ability to keep paying. That is a circular dependency, and circular dependencies are the natural prey of a tightening cycle. It does not matter how clean the liquidation engine is if the collateral is a promise to keep promising.

Then there is the channel that connects the Fed to every asset class at once: the dollar. A tightening cycle supports the dollar through rate differentials, and a stronger dollar pressures everything priced in it — commodities, emerging-market debt, and, at the margin, foreign demand for dollar-denominated risk assets. Crypto is not exempt, despite the ideology. The largest stablecoins are dollar instruments, and the largest pools of crypto liquidity are denominated in them. When the dollar strengthens, the real cost of crypto leverage rises for every participant outside the United States, and the marginal offshore buyer steps back. This is not a headline channel; it is a plumbing channel. It moves quietly, and it moves first.

What the whales actually did around the FOMC

Here is where the on-chain data diverges most sharply from the narrative, and where I want to be precise, because this is the part that gets misread.

There is a durable folk belief that whales are directional — that they buy the dip, sell the top, and telegraph their intentions through exchange flows. The cohort data says something narrower and more useful. Whales don't react to the announcement. They position ahead of the expectation. The largest accumulation and distribution events in the cohorts I track cluster in the days before a policy meeting, not the days after, because the announcement resolves uncertainty and the profit lives in the uncertainty itself.

Across the last three FOMC windows the pattern has been consistent: net exchange inflows from large cohorts rise into the meeting, then flatten immediately after. That is not a directional bet on the Fed. It is liquidity provision — whales supplying exit liquidity to whoever trades the headline, and collecting the spread for the service.

I first measured this behavior properly in 2021, when I applied clustering techniques to floor-price movements across twenty major NFT collections and found that fifty super-whales controlled roughly 15% of total volume. The insight that report produced — and that the financial press amplified — was that these entities manipulated perception. The deeper, less quotable finding was that they were patient. They were not front-running prices. They were front-running narratives.

The measurement challenge is real. Cohort attribution is an art dressed as a science; the same address can look like a whale on Monday and a routing contract by Friday. What makes the signal usable is not precision but consistency. I use the same clustering parameters across cycles, so the numbers are comparable to themselves even when they are approximate in absolute terms. An index made of rubber can still measure a trend, as long as the rubber is the same rubber every time. That discipline is what separates a dashboard from a decoration.

That distinction is now operationally decisive. If large holders position ahead of the Fed, then the signal worth reading is not what happens on the day of the hike; it is what the cohort flows did in the five sessions before it. Here is what they did not do this time: they did not accumulate aggressively, and they did not exit. They waited — which is the most informative thing a whale can do. Neutrality from a cohort with perfect information is not indifference. It is optionality, and optionality is expensive to maintain.

There is a second cohort worth separating from the first. The wallets that arrived in the last eighteen months behave differently from the wallets that survived 2022. The newer cohort is faster to rotate and more sensitive to headlines; the survivor cohort is slower and more sensitive to spreads. When both cohorts move the same direction, you have a signal. When they diverge, you have a market with no consensus — which is precisely what the current data shows. The old wallets are positioned. The new wallets are reactive. The price is set by whoever is louder, and the flows are set by whoever is larger.

The data doesn't lie; it simply refuses to flatter.

So far the ledger has handed us three facts: dollar inflows are flat, the on-chain risk-free rate is tethered to the sovereign rate, and the largest holders are neutral rather than bullish. None of that confirms the Goldman thesis. None of it refutes it either. What it does is relocate the question — from "will rate hikes halt the bull market?" to something sharper: what kind of bull market can survive a tightening cycle in which the marginal buyer is not new capital?

The honest answer is: a leveraged one. And leveraged bull markets are not imaginary. They can run for a long time. But they have a specific failure mode, and its name is the funding rate. When the cost of carrying a position exceeds the expected return of holding it, the position closes itself — quietly at first, then all at once. You do not need a crash to unwind leverage. You need arithmetic.

The structure of 2026 is not the structure of 2021

Before drawing conclusions, it is worth naming what is genuinely new, because a bull market that has changed molecules deserves diagnostics that change with it.

The most visible change is the convergence of AI and crypto infrastructure. In 2026 I partnered with a boutique analytics firm to map data flows between decentralized compute networks and AI training pipelines, and we tracked 10,000 data transactions. Roughly 40% of high-value training data originated from verified on-chain sources. That is a real number about a real market, and it is the first time I have seen an on-chain activity stream whose demand is not primarily speculative. Compute is a productive asset. It has a cost, a utility, and a customer who is not buying it to sell it to someone else.

Productive properties are not rate-insensitive properties. A decentralized compute network prices its services against centralized alternatives, and centralized alternatives price against the cost of capital. When the Fed tightens, the cost of building a data center rises, cloud pricing firms up, and decentralized compute looks relatively cheaper — not because it improved, but because its competitor got more expensive. That is a flattering wind, and flattering winds reverse.

There is a version of this bull market that is real and a version that is rented, and the difference is where the revenue comes from. A network paid by users for compute or storage or bandwidth is earning. A network paid by emissions is borrowing from its future to fund its present, and the interest rate on that borrowing is dilution. When the sovereign risk-free rate was near zero, dilution was cheap relative to the alternative. When the sovereign rate climbs, dilution gets expensive, and the emissions-funded segment of the market must either raise real revenue or shrink. Most will shrink. That is not a prediction about price; it is a prediction about structure, and structure is what survives.

Then there is the long-running story that refuses to admit it is a story: real-world assets on-chain. Three years of narrative have produced a handful of pilot programs and a great deal of press. The uncomfortable truth is structural. Traditional institutions do not need a public chain to move money between themselves; they need settlement, finality, and legal recourse, and they already have all three through venues that do not expose their order flow to the entire world. When an institution does tokenize an asset, it usually tokenizes a wrapper it controls, on a permissioned rail it controls. The public chain is a display case, not a settlement layer. RWA did not fail because the technology was inadequate. It stalled because the customer was never asking for it.

The same skepticism applies to the scaling narrative. Rollups are excellent engineering, and I have no patience for people who dismiss them. But engineering is not economics. Zero-knowledge proving costs remain punishingly high, and unless gas returns to bull-market levels, the operators absorb the difference. A scaling solution whose unit economics only work in a bull market is not a solution; it is a leveraged bet on demand. The rollup that survives the next cycle will be the one that found a revenue line which is not a subsidy.

And then there is Bitcoin, where the cargo-cult impulse to bolt applications onto a settlement layer has produced a decade of expensive experiments. BRC-20 and Runes are the latest chapter: using a deliberately constrained, high-assurance settlement chain as a venue for speculative token issuance, which is a bit like using a Rolls-Royce to haul gravel — it insults the car and it doesn't carry much. The base layer's virtue is that it does almost nothing. Every attempt to make it do more trades that virtue for throughput, and throughput is cheap everywhere else.

None of this is a bear case. It is a map of rate sensitivity. The segments of the market with real cash flows and real demand will weather a tightening cycle better than the segments that have been surviving on narrative and emissions. When the denominator moves, the market stops asking what is exciting and starts asking what is funded.

Contrarian

Now the part that will annoy everyone.

The reflexive move is to treat equities and crypto as two expressions of the same macro reality, and to declare that if Goldman is right about the S&P, it must be right about the chain. That is a correlation dressed as a causation, and it is wrong in two directions at once.

It is wrong about equities first. Goldman's argument is not that rate hikes are harmless; it is that earnings outrun the discount rate. The seven-cycle history is real, but its implicit precondition is that no recession lands inside the twelve-month recovery window. Goldman never states that precondition; it recruits the average without naming the assumption. The moment earnings get revised down, the note's own logic inverts, because a 19x multiple on falling earnings is more expensive than a 22x multiple on rising ones.

It is wrong about crypto second, and here I will be blunt. Crypto does not price rates the way equities price rates. Crypto prices liquidity, and rates are one input into liquidity — not the input. The chain can rally while the Fed hikes, if the liquidity arrives from somewhere else: an issuer expanding supply, a wrapper absorbing flows, a treasury desk borrowing short to buy long. The correlation between Fed policy and crypto price is real but unstable, and it is unstable precisely because the transmission runs through intermediaries whose behavior changes.

The intermediaries deserve their own line. The transmission from Fed policy to crypto price does not run through a wire; it runs through entities — custodians, issuers, market makers, lending desks — whose balance sheets respond to rates before their clients do. When one of them reprices, the effect is not a gradual adjustment; it is a discrete change in available liquidity, and discrete changes are what break trends. This is why crypto's correlation to the Fed is unstable. The correlation is not between policy and price. It is between policy and the intermediaries, and the intermediaries change their mind at different times for different reasons. Anyone who models the first correlation directly is skipping the only part that matters.

I watched that instability up close in 2022. When the insolvency cascade broke, the protocols did not fail because the Fed had hiked. They failed because a duration mismatch that had been invisible in a zero-rate world became fatal in a positive-rate one. The $2 billion in undercollateralized positions I surfaced was not caused by monetary policy. Monetary policy merely removed the fog that had been hiding it. Tightening does not create fragility. It reveals it — and the reveal is where the money is made and lost.

There is a trap inside that observation, and I want to name it before someone walks into it. If tightening reveals fragility, the naive conclusion is that tightening always kills the bull market. That is false, and it is false for a reason the data makes obvious: the reveal only matters if the fragile positions are large relative to the market's ability to absorb them. In 2022 they were. Today, the leverage is more transparent and the collateral is better engineered, which lowers the probability of a cascade without lowering the probability of a repricing. A repricing is survivable. A cascade is not. The distinction is the difference between a bad quarter and a bad cycle.

So the honest reading of the current moment is narrow. The Goldman note is a statement about equities. It is not a statement about crypto, and anyone porting it across the boundary is importing an assumption they have not tested. The chain is telling a smaller, harder story: liquidity is holding but not growing, leverage is elevated but visible, the largest holders are waiting rather than buying, and the risk-free alternative is getting more attractive by the week.

None of those facts requires a recession to matter. They only require a spread.

Takeaway

Here is the signal I am watching next, and it is not the FOMC statement.

It is the gap between the on-chain borrow rate and the short-dated Treasury yield, measured across the five sessions after September 14. If that gap holds or narrows, leverage is durable and the bull case survives the denominator. If it widens, the arbitrage runs one direction, and the marginal dollar leaves the table regardless of what the price does.

I will also be watching the survivor cohort in the week after the meeting — not their price target, which they do not have, but their exchange flows, which they cannot fake. If the wallets that lived through 2022 begin to move toward exchanges, the neutral stance was a position after all, and the direction is settled. If they stay put, the waiting continues, and so does the ambiguity. Either way, the answer arrives in the flows before it arrives in the chart.

Goldman can tell you what the market believes about earnings. The ledger will tell you what the market can actually afford.

Precision in chaos is the only true advantage.