Eighty-Six Thousand Dollars of Memory

Sunday, 02:40. A spot Bitcoin ETF allocator, awake, with the screen still on.

Fine. Say it plainly. The best macro setup Bitcoin has had since September arrived on Friday at 8:30 a.m. Eastern, and the coin made it to $87,085 and turned around like a cat at a bath.

Look at what the market was handed. Payrolls of +29,000 against 84,000 expected. Unemployment at 4.2%. August revised down to 133,000. October hike odds that stood near 70% at the start of the week were somewhere around 16% to 22% by the close, depending on whose FedWatch screen you trust. This is the print every crypto account on the internet has been praying for since the Fed hiked on the 16th. Shorts got the message: CoinGlass has $244 million of the $333 million liquidated during the rally on the short side. Fuel, match, ceiling. And the ceiling held, within a rounding error of where the ceiling has always been.

By Saturday morning the price was back at $84,600 and $433.6 million of leverage had been erased, 74% of it longs who showed up to the breakout late. Second rejection near $87,000 in about two weeks. I keep staring at that number and it keeps staring back, because I know exactly what it is.

It is us.

Bitfinex's analysts put roughly 1.39 million BTC with a cost basis between $84,000 and $86,500. Run it: at $85,000 a coin that is about $118 billion sitting at the exact altitude where Bitcoin keeps stalling. Another data shop pegs the aggregate US spot ETF breakeven near $86,000. These are my people. September's $2.65 billion of net inflows came disproportionately in a single burst, $999 million on the 21st and $714.7 million on the 22nd, when the price sat in the same neighborhood where we all sit now. We bought a breakout and the breakout became our basis.

So every rally to $86,000 meets a wall of holders who finally see green and want to leave. Fidelity's FBTC lost $60.7 million on Thursday and Grayscale's GBTC lost $31.4 million while BlackRock's IBIT took in $195.6 million. Net, $102.7 million came in. Net is a polite word for who is selling to whom.

Do I hate the structure? I wrote the structure. In 2023 we told every committee that the ETF wrapper would professionalize the asset, and it did. Professionals do something retail never learned: they check their cost basis. Nobody who bought at $60,000 in a panic ever drew a wall at $60,000, because nobody who panics remembers. The wall at $86,000 exists because the buyers at $86,000 are people who write memos.

Now the macro, because I have to defend a book on Monday. The 10-year hit 5.342% on Thursday, highest since 2002, and Friday's dovish jobs print could only drag it to 5.16% before it climbed back near 5.30% for a fifth straight weekly gain. Hold odds for the October 28 meeting sit around 78% to 84%. That should be the dream. Instead the long end is telling me the Fed's next move is no longer the point. Diesel is. The G7 spent a hundred million barrels on Friday and Brent shrugged to $102.25. I own an asset that is supposed to be the hedge against all of this, and it traded this week like a Nasdaq name with worse hours.

Which brings up the thing the committee will ask. Gold. Gold was $4,239.50 on Friday morning and rising on the same payrolls. Central banks are buying it by the tonne. I have not found the central bank buying Bitcoin by the tonne. The debasement story has two beneficiaries and only one of them has a sovereign bid. We keep calling the other one digital gold. It would be more accurate to call it digital cost basis.

Okay. Cold read of the tape, because I am tired and the spreadsheet is honest.

Bull case: ETF flows came back this week after a $148.7 million outflow on September 30. The price held an $83,000 to $87,000 range through a Friday-night leverage flush. Last October's $19 billion liquidation crisis is a distant memory because leverage is genuinely lower now, and a market that can absorb $433 million of forced selling without a cascade is a sturdier one. If the Fed holds on the 28th and the long end calms down, the people with a basis at $86,000 stop being sellers and become the floor. Cost basis is a wall from underneath and a shelf from above.

Bear case: the same shelf is a ceiling until it is cleared, the 10-year has not agreed to anyone's soft landing, and the long end does not care what the labor market is doing. Each failed breakout transfers coins from impatient hands to patient ones and teaches both sides which direction the next test breaks. A break below $83,898 and the market starts talking about $80,000.

Neither case needs the other to be wrong. That is the dull and expensive truth of mature markets.

04:10 now. I will do the sensible thing. I will not sell below my basis, because my basis is $86,000 and so is everybody else's, and the market is a room full of people politely refusing to be first through the door.

Somebody will be. The only open question is whether it is us.