Gold, the hedge everyone wanted in January, is 19% below its record and has just failed a breakout. Nine months ago the consensus on these two assets pointed in opposite directions. Both consensus views have now been wrong in the same direction.
The print
At 09:00 UTC bitcoin was 83,670, up 4.29% from Sunday’s 81,162. Ether was 2,692, up 4.6%; solana 115, up 6.5%. The move began on expiry Friday in New York and has run through a weekend in which nothing obvious happened — no central bank, no approval, no fund launch.
The week’s macro news, such as it is, is that the Federal Reserve hiked on Wednesday, the 10-year is back at 5.00% this morning, and the dollar index is 100.34. None of that is a textbook bitcoin tailwind.
And now the discipline. Our own rule, printed in Sunday’s letter, is that a level is taken on a closing basis. 82,139 is a weekly-close line. Bitcoin trades every hour of every day, which makes “the close” a convention rather than an event, and the convention we chose was the weekly one. That means today’s 83,670 does not score. The score comes on Sunday 27 September. We are saying so now, while the number is pleasant, rather than in a week’s time when it might not be.
The equities went first
The coin stocks did not follow bitcoin this morning; they led it on Friday, and they led it hard. Friday’s closes from the Closelook data lake:
Name, Friday close Friday
Strategy (MSTR)153.92+16.4%
Marathon (MARA)13.24+13.8%
Coinbase (COIN)194.25+11.7%
Robinhood (HOOD)119.82+9.1%
Riot (RIOT)23.75+8.6%
Hut 8 (HUT)98.74+8.5%
CleanSpark (CLSK)14.47+8.4%
Circle (CRCL)91.78+7.9%
Bitwise 10 (BITW)53.41+6.8%
iShares Bitcoin (IBIT)46.02+6.3%
Friday’s closes, Closelook data lake. Bitcoin rose 5.9% the same day — so every name above except IBIT and BITW outran the coin.
Strategy and Coinbase both retook their two-hundred-day averages — 137.04 and 189.22 — on that session, which was the specific thing Sunday’s letter said to watch. Both are indicated higher again in pre-market trading this morning. We are not printing those pre-market numbers as facts: extended-hours prints are not closes, and the difference has embarrassed people before. Today’s closes are what count and we will carry them tomorrow.
The part that is actually interesting
Now rewind to January. Gold had just made a record. The hedge trade was the crowded, obvious, universally recommended position: the dollar was in question, the deficit was in question, and gold was the asset that had proved it. Bitcoin, meanwhile, was a yesterday-man investment — down from its October record of 126,198, associated with a cycle people had already decided was over, and written about in the past tense.
Here is how the two have actually done, measured from the same starting point:
Gold (GLD)BitcoinYear to date−3.7%−6.8%High of the year495.90 (29 Jan)96,929 (14 Jan)From that high−19.1%−16.1%Low of the year364.96 (16 Jul)58,559 (30 Jun)From that low+9.9%+39.0%
Gold via the GLD ETF and bitcoin in dollars, both from a 26 December 2025 base, Closelook data lake. Bitcoin’s figure is this morning’s live price; gold’s is Friday’s close.
Read the top line and nothing happened: both are down on the year, and the S&P 500 ETF is up 11.2% and beat them both without needing a thesis. Read the bottom two lines and the shape is completely different. Gold peaked in the fourth week of January and has spent eight months making lower highs. Bitcoin bottomed on 30 June and has spent three months making higher lows.
The failed breakout
Gold’s recent path deserves stating precisely, because it is the mirror image of what bitcoin is doing. The gold ETF bottomed at 364.96 on 16 July — an area that corresponds to roughly $4,000 an ounce in spot terms. It then rallied 17.3% to 428.07 by 25 August. That was the breakout attempt. It failed. Gold has not made a higher high in the four weeks since, closed Friday at 401.17, and remains below its two-hundred-day average at 416.22 — a line it has not reclaimed since the Fed hiked. Gold futures were $4,389.60 this morning, down 0.8%.
Silver is the same story with the volume turned up: down 15.7% on the year, and 43.2% below its 28 January high.
So the hedge complex rose from a base, ran seventeen percent, stalled at a level, and rolled over. The digital one fell to a base, ran thirty-nine percent, and is now at its level with the question still open. Same year. Same macro. Opposite chart.
What is obvious is obviously wrong
The old adage holds: what is obvious on markets is obviously wrong. It is worth being precise about why that is not merely a clever line. In January both of these positions were consensus, and consensus is a statement about who has already traded. Everyone who wanted the gold hedge at a record had bought it, which is exactly the condition under which good news stops moving a price. Everyone who had given up on bitcoin had sold it, which is exactly the condition under which the absence of bad news starts moving one. The fundamentals did not need to reverse for the prices to. Only the positioning did.
That is also the honest limit of this entry. We are describing a rotation that is at most three months old, in two assets that are both still negative on the year, on a morning when one of them is up four percent. None of that is a conclusion. It is the reason we wrote levels down in advance instead of opinions.
What would settle it
Three things, in order of weight. First, Sunday’s weekly close: above 82,139 the recovery is a trend by our own stated rule, below it this was a strong week inside a range. Second, gold reclaiming 416.22 — if the hedge complex turns back up alongside bitcoin, this is a liquidity story about both and not a rotation between them. Third, the rails against the miners: Sunday’s letter measured that spread at −4.7 points and argued that if bitcoin runs through 82,139 the miners should lead again, because leverage to the coin beats yield on the reserves in a rising tape. Friday’s tape said exactly that, with Strategy and Marathon at the top of the table. One session is not a verdict, but it is the first data point on a question we set before we knew the answer.
Elsewhere on Friday’s closes, the Closelook indices: Rubin Build-Out +2.64% on the day and −0.40% on the week; HALO Growth −0.69% and −1.46%; Euro-AI +0.89% and −0.23%; AW40 −1.54% and flat; the Agentic Ecosystem index −0.75% on the day but +6.57% on the week. The Money Temperature composite sits at 50 on Friday’s data — mixed, and by its own construction not a signal.
This is an investment diary. It records what we watch, the levels we wrote down in advance, and whether they held. It is not investment advice and it is not a recommendation to buy or sell anything.