We didn't get a new script on September 9. We got the same one — same instrument, same desk, same stated intent — and Bitcoin walked off set halfway through the scene.
I was in a corner of a co-working space in BGC when the August print crossed. Six in the morning, Manila time, and a friend in our trading Discord was already awake, typing in all caps. The US Treasury had stepped into the long end of the curve with a buyback, ten-year yields had peeled back from their highs, and Bitcoin was ripping from $65,000 toward $80,000 as if someone had pulled a curtain off the stage. Nobody in that channel ran a discounted cash flow model. Nobody asked for a white paper. We leaned in, because the crowd was leaning in, and in this market the crowd is the signal you can actually see.
Then September 9 arrived. Same Treasury, same tool, same goal — ease the pressure building on the long end. And Bitcoin shrugged. It didn't launch. It slid. By the time the tape settled it was pinned under $78,000, refusing to reclaim the range it had spent all of August building.
Eighteen years in this industry have taught me that the interesting part is never the price. It's the gap between two prices that should have matched. Two near-identical policy moves, two opposite reactions — that gap is not noise. It's a confession about how Bitcoin is being priced right now.
Context — The Buyback, the Bargain, and the Broken Ticker
Let me back up, because half the people trading this tape have never opened a Treasury plumbing diagram, and that's fine — until it isn't.
A buyback is exactly what it sounds like. The US Treasury goes into the secondary market and repurchases its own outstanding bonds. Historically it's a housekeeping tool: you use it to improve liquidity in off-the-run issues, to smooth the curve, to manage the mix of what's outstanding. It is not quantitative easing. It is not the Fed. It is the government buying back its own debt with cash it has to raise somewhere else. The distinction matters, and I'll come back to it with a knife.
In the reporting that set off this whole conversation, the August 19 operation was read as a surprise policy pivot — an admission that the long end was under stress and that the government was willing to intervene directly. That reading is why the market moved the way it did. When a participant you assumed was passive suddenly acts, you reprice everything downstream of them.
Here are the numbers that mattered. The September operation came in smaller than the market's implied expectation — roughly $6 billion against Wall Street whispers of up to $10 billion. The ten-year yield sat near 4.85%. The twenty- and thirty-year yields sat around 5.30%, which the piece calls the highest in three years. Bitcoin had traveled from $65,000 in August to $80,000 on the surprise, then fell through $78,000 the second time around.
Same policy tool. Opposite reception. And before we go deeper, one flag: the original account carries no year, several unsourced data points, and a timeline that doesn't fully reconcile — a hawkish Fed voice, an elevated long end, and a bond market described as openly fighting the Treasury, all jammed into one frame. So treat the numbers as directional and treat the mechanism as the real subject. The mechanism survives even if a date is wrong.
Because the mechanism is this: the market is not pricing a policy. It is pricing the difference between a policy and its own expectation of that policy. Everything below is a fight about that sentence.
There's a second layer most crypto readers skip, and it's where the adults live. A government doing buybacks while a central bank stays hawkish is a picture of fiscal dominance in progress — the borrowing arm of the state trying to hold rates down while the monetary arm refuses to help. Market commentators like the Kobeissi Letter have framed the current standoff as the bond market "fighting the Treasury," and that framing is worth taking seriously. When the market starts testing whether the Treasury can keep intervening, every intervention gets read through a suspicion lens. That suspicion is the soil September's disappointment grew in. And notice which levers nobody touched — no adjustments to the supplementary leverage ratio, no visible management of the Treasury General Account, no meaningful change in the reverse repo facility. Those are the tools that actually move dollar liquidity into risk assets. A buyback is not one of them. Keep that in your pocket.
Core — Bitcoin Is a Long-Duration Asset, and Nobody Adjusted the Discount Rate
Here's the thing nobody said out loud in August, and nobody said out loud in September either.
Bitcoin has no cash flow. No coupon, no dividend, no protocol revenue to distribute, no buyback of its own. Its price is not the sum of future payments discounted back to today — because there are no future payments. What you own is a claim on a network, a monetary premium, and the collective belief that someone else will pay more later. In valuation language, that makes it an extremely long-duration asset. It behaves like a thirty-year zero-coupon bond with a story stapled to it.
Why does that matter? Because long-duration assets live and die by the discount rate. When the risk-free rate rises, the present value of every distant, uncertain thing falls. That is not an opinion about crypto. It is arithmetic applied to any asset whose payoff sits far in the future. And when the ten-year climbs to 4.85% and the twenty- and thirty-year stretch toward 5.30%, the opportunity cost of holding a zero-yield asset climbs right alongside it.
Now look at what a buyback actually does to that arithmetic. It lowers yields at the long end — mechanically, by reducing supply, and psychologically, by signaling that someone cares. Lower yields mean a lower discount rate. Lower discount rate means a higher present value for long-duration things. On paper, that should be rocket fuel for Bitcoin.
Except the two operations landed in different rate regimes. In August, the buyback arrived as the long end was rolling over — the discount rate was falling and the surprise was landing at the same time. Two tailwinds, one ticker.
By September, the buyback was swimming upstream. Yields were pushing higher, the long end was testing that three-year high, and a hawkish Fed voice was sitting on the other side of the table. The buyback was easing the long end while the broader rate complex was tightening it. The discount-rate channel was pulling in the opposite direction from the liquidity channel — and the discount rate won.
That's the piece of the puzzle the original analysis never picked up. It attributed September's disappointment entirely to "the lack of surprise" — a demand-side story about expectations. But there's a supply-side story running underneath, and it's about valuation. You can deliver the same pleasant surprise twice and still get a different result if the discount rate has moved against you in the meantime.
And there's a confound hiding in plain sight. How much of that August run from $65,000 to $80,000 was the buyback, and how much was everything else moving at the same time — month-end positioning, seasonal flows, a general risk-on drift? Without stripping those out, the buyback's contribution is a correlation wearing a costume. I've watched this exact pattern in my own flow maps. Back in the DeFi summer of 2020, I ran a portfolio of fifteen ETH across SushiSwap and Uniswap, chasing the highest APYs in a frenzy that felt like a video game. The lesson I took from that sprint wasn't about yield. It was that the same yield number produced completely different behavior depending on what the risk-free alternative was doing that week. When the cost of capital is falling, every payout looks generous. When it's rising, even good payouts get ignored. Bitcoin here is that lesson wearing a suit.
So when someone tells you "the buyback failed because there was no surprise," nod politely and then ask them what the thirty-year was doing. Because that's the number that decides whether a liquidity headline becomes a bid or a shrug.
Core — The Buyback Was Never a Liquidity Injection. It Was a Signal.
Now let's take the other half of the trade apart.
Everyone in crypto treats a Treasury buyback like it's a firehose of dollars pointed at risk assets. It isn't. Six billion dollars is a rounding error against a Treasury market that trades hundreds of billions on a normal day. If you think $6 billion of buybacks is going to reroute the world's capital into Bitcoin, you've confused a plumbing repair with a monsoon.
A buyback does its work inside the dealer community. It improves the balance sheets of primary dealers and market makers. It takes some of the ugliest, least liquid bonds off the street. It makes the machinery of the bond market run a little smoother. That's it. The dollars don't leap from the Treasury's account into a Coinbase order book. There is no pipe. The only pipe is the one in people's heads — the signal. And a signal is a strange thing to build a leveraged position on, because a signal has no collateral.
Signals decay.
August was a surprise. Surprises get repriced fast, because nobody has a position for them. A policy move you didn't see coming forces everyone to re-underwrite their model of the government's reaction function. That's why Bitcoin went from $65,000 to $80,000 — not because $6 billion of cash found its way into the market, but because a hundred billion dollars of positioning had to be reconsidered at once.
September was not a surprise. It was a confirmation. The market had already written "the Treasury intervenes when the long end stresses" into its pricing. Worse, the September operation came in below the whisper number — around $6 billion against expectations closer to $10 billion. That's a negative surprise hiding inside a positive event. The market asked for proof the Treasury would go big and got proof it would go modest.
So the signal went from loud to silent to faintly negative over the course of two operations. That is a textbook case of what I'd call signaling decay — the same headline, delivered the second time, carries a fraction of the information and can even invert. First intervention reshapes the map. Second intervention confirms the map. Third intervention, if it's too small, becomes a warning that the map is all they've got. That third-act risk is the one nobody wants to price, because pricing it means admitting the policy reaction function has a ceiling.
We didn't lose a policy in September. We lost the novelty of the policy. And in a market that runs on novelty, that's the same thing.
There's an irony buried here, and it makes me smile at my desk. Crypto spent a decade dreaming about being "uncorrelated." What it got instead was a seat at the big kids' table — where the only thing that moves your price is what the adults decide to do with the long end of the curve. Bitcoin's price is no longer a crypto question. It's a rates question with a crypto ticker. The buyback didn't make that true. It just exposed it.
Core — The Blind Spot: Where the ETF Flows Went
Here's my biggest complaint with the original analysis, and the reason I can't fully trust its attribution: it never once mentions exchange-traded fund flows.
That is indefensible in the current market structure. Since the spot Bitcoin ETFs launched, net daily flows into those vehicles are the single most direct read on marginal institutional demand. They are the measurable heartbeat of the exact cohort — allocators, RIAs, family offices — that the buyback signal is supposedly moving. If you want to know whether a Treasury operation lifted Bitcoin, you don't guess. You look at whether the ETFs took in money that week.
I spent time in Singapore in 2024, at the financial forums where the institutions finally walked in wearing name tags. The thing that struck me wasn't how much money was moving. It was how mechanical the flows had become. A pension consultant doesn't wake up and feel bullish on Bitcoin because the Treasury bought back some bonds. They rebalance on a calendar, on a mandate, on a risk budget. The ETF wrapper turned Bitcoin's institutional bid into something closer to a spreadsheet than a mood. And spreadsheets don't move because of a headline — they move because of a rate. I watched the same dynamic play out in the NFT market, where collectors I met at launch parties in Makati treated digital assets as social capital rather than financial instruments — held for access and identity, not for price. That's fine for a weekend. It's useless for attribution. When an asset trades on identity instead of fundamentals, you can't decompose the price with a headline, and you can't decompose Bitcoin's August move either without the flow tape.
So when the original piece credits August's rally to the buyback and September's fade to the buyback's inadequacy, it's missing the middle term of the equation. How much of that move was buyback, and how much was ETF inflows, positioning, or simple seasonality? Without flow data, you can't strip the components apart. You're left with a correlation and a story, and a story is not an attribution.
This is the oracle-latency problem wearing different clothes. One of the things that has always bothered me about DeFi is that its oracles introduce a lag between what's true and what the contract believes — the price updates, but the underlying reality settles on its own schedule, and the gap between them is exactly where the risk hides. Single nodes dressed up as decentralization don't fix that; they just move the gap somewhere less visible. Macro works the same way. The buyback updates the headline instantly. What it does to the actual bid for Bitcoin arrives with a lag, through ETFs, through dealers, through the slow rebalancing of allocators who genuinely do not care about the news cycle. You can't read the lag off a daily candle. You read it off flows. The original analysis skipped the only instrument that could have measured it.
Core — Gold Moved With Us in August. That Was the Tell.
Now the part I keep circling back to, because it's the most interesting thing in the whole episode and it gets exactly one line.
In August, gold and Bitcoin rallied together.
Sit with that. A fifteen-year-old speculative asset and a five-thousand-year-old store of value both went up on the same Treasury operation. That is the "hard asset / liquidity beneficiary" trade — when money expects to be debased or expects easier conditions, it runs toward anything that isn't a promise from a government. In that frame, Bitcoin and gold are cousins. Both are non-sovereign. Both are scarce. Both get repriced when the opportunity cost of holding them moves.
Then September happened, and here's where the original reporting leaves a hole I have to peer into. If gold held its bid while Bitcoin faded — if the two cousins split — then the August co-movement wasn't a property of Bitcoin. It was borrowed. It was Bitcoin riding a hard-asset wave it didn't own, and when the wave broke, Bitcoin got left on the sand while gold kept floating.
That's an identity crisis, and it's the most important thing happening under the surface. Bitcoin is being priced by two different crowds at once. One crowd buys it as digital gold — a hedge, a haven, an insurance policy against fiscal dominance. The other crowd buys it as the highest-beta risk asset on the planet — a leveraged bet on liquidity. In August, both crowds could be right, because easier conditions lift havens and risk assets together. In September, the conditions split the crowds apart. Rising yields punish the risk-asset buyers and reward the haven buyers, so the two pull in opposite directions on the same ticker. Bitcoin gets torn in half. That's not weakness. That's unresolved.
And the unresolved thing has a cost I keep coming back to — the fee market. Bitcoin's security budget rests on the price. Miners earn block subsidies plus fees, and after the 2024 halving the fee share has been thin, oscillating in the low single digits on quiet weeks. A higher monetary premium means a higher price, which means a healthier hash rate, which means a more secure chain. Which means every macro flow into Bitcoin isn't just a price event. It's a security event wearing a delay. The reason the Ordinals inscription wave mattered — regardless of what you think of JPEGs on the base layer — is that it injected fee revenue into a budget that needs it. When the monetary premium sags, the fee question sharpens. When the premium soars, the question goes quiet. The Treasury buyback and the miner's electricity bill are the same story told at two different time scales, and almost nobody connects them. If the liquidity narrative that has funded this cycle quietly dies, the chain's security math gets ugly long before the price chart does.
Contrarian — The Buyback Didn't Fail. We Misread What It Was Pricing.
Here's where I part ways with the consensus reading.
The popular take is that the September buyback failed. I think that's the wrong frame, and it's the frame that will cost people money. The buyback did exactly what a buyback is designed to do: it put a floor under the long end of the curve and stabilized the machinery of the Treasury market. On its own terms, in its own market, it worked. What failed was our expectation that every Treasury action is a crypto liquidity event. That expectation was never true. It was a story we told ourselves during a bull market because the story was flattering.
The deeper contrarian point is about what Bitcoin has become. Bitcoin is no longer trading as a liquidity sponge. It's trading as a discount-rate instrument. That's a regime change, and it's being missed because the old lens — easy money in, number go up — is so much more fun to use.
Why is this the honest read? Because "no surprise" is a lazy explanation. It explains September perfectly and it would have called August completely wrong — in August there was also no prior template, and the move was enormous. The variable that actually separates the two episodes isn't the degree of surprise. It's the direction of the long end. In August, the buyback and the rate complex pointed the same way. In September, they diverged. Bitcoin traded the second derivative — the rate of change of expectations — and the second derivative flipped sign.
If that's right, the implications invert the doomer take. A discount-rate instrument isn't broken; it's mature. It's reacting to something real instead of to vibes. And a duration asset that's been beaten down by rising yields is exactly the asset that moves convexly when yields finally roll over. The pain of September is the setup for a response that won't be linear when the long end turns. But — and here's the part the bulls will skip — until the long end turns, every buyback is a price-of-risk story, not a bid story. You can't buy a discount-rate asset and expect it to ignore the discount rate.
There's a governance angle in here too, and it's the one with the longest shadow. The August intervention built a new policy reaction function — the market learned, in one move, that the Treasury will step in when the long end stresses. That's a credibility deposit. September's undersized operation was a withdrawal, because it revealed the deposit has limits. Policy credibility, like any other balance, runs on marginal utility: the first move reshapes the map, the second confirms it, and the third — if it underwhelms — starts to signal that the map is all there is. The buyback desk isn't the only lever the government has, but it may be the loudest, and loud levers get quiet fast.
We didn't get liquidity in September. We got a signal, a smaller one than expected, into a rising-rate tape, from a market that had already priced the first signal to exhaustion. Bitcoin did the only rational thing a long-duration asset can do: it marked down.
Takeaway
The next real signal won't come from the buyback calendar. It'll come from the issuance calendar. If the Treasury shifts its borrowing toward bills and cuts long-bond supply, that's not a signal — that's a regime change, and it will matter more than any $6 billion repurchase ever could. Watch that. Watch the five percent line on the long end like it's a heartbeat. And watch whether Bitcoin starts trading gold's book or the Nasdaq's book when the two disagree.
The question I can't shake, sitting here with the charts open and the Singapore forums already fading into next quarter's memories: when the long end finally rolls over and the discount rate falls, will Bitcoin answer like the liquidity asset we all fell in love with — or like the duration asset it quietly became? Because those are two different trades, riding on the same ticker, and only one of them is going to be right.