Why bitcoin swings so hard on US CPI release day

The minute to watch: 08:30 in New York, 12:30 UTC until November

The US Consumer Price Index comes from the Bureau of Labor Statistics, and every row on the bureau’s CPI release schedule carries the same time: 08:30 AM Eastern. September’s figures are due on 14 October 2026, while New York is still on daylight time, which puts the release at 12:30 UTC. Daylight time ends on 1 November, so October’s figures, due 10 November, land at 13:30 UTC. If you follow markets from London, Singapore or Sydney, convert from UTC rather than from memory; the clocks change on different weekends in different countries.

The hour matters. The New York Stock Exchange doesn’t open until 09:30, so for the first hour after the print the cash equity market is shut. What is open: equity index futures, Treasury futures, currencies — and crypto, which never closes. Anyone who wants to act on the number immediately has only a handful of places to do it. That is why bitcoin so often lurches in the same minute as two-year Treasury yields and the dollar index.

So a push alert at 08:31 saying “bitcoin plunges” is rarely a crypto story on its own. More often it is one macro number being repriced across several markets at once, and the bitcoin chart is just where you happened to see it.

Markets trade the gap between CPI and the forecast, not the CPI number

Before every release there is a consensus forecast, usually the median of economists’ estimates, printed next to the actual figure on any economic calendar. Prices have already been set around that forecast before 08:30.

The only new information is therefore actual minus expected. A 0.3% monthly rise means nothing in isolation: against a 0.2% forecast it runs hot, against 0.4% it runs cool. The same number can produce opposite reactions in two different months. It’s the same logic as news that was priced before you saw it, except that with CPI the expectation is written down in public and you can check it.

A CPI report also contains more than one number. The bureau’s CPI home page summarised the August 2026 data like this:

• All items, seasonally adjusted: up 0.4% on the month.
• All items, not seasonally adjusted: up 3.4% over 12 months.
• All items less food and energy (“core”): up 0.3% on the month, 2.4% over the year.

On the annual measure, headline and core sat a full percentage point apart, with food and energy making up the difference. Whether inflation looked “high” depended on which line you read. Core strips out the two most volatile categories and is the one usually watched for persistent inflation, so at 08:30 algorithms and human traders are parsing several lines at once. An order placed off the headline in the first second can be overridden by the core reading thirty seconds later — one reason the CPI-day candle so often spikes and then reverses.

Between the percentage and the price sits the next Fed meeting

The route from a CPI percentage to the bitcoin price runs through one question: what will the Federal Reserve do at its next meeting?

The Federal Open Market Committee meets eight times a year. On the 2026 meeting calendar the October meeting falls on the 27th and 28th, two weeks after the 14 October release. With a decision that close, a hot or cool print rewrites the market’s bets on its outcome directly. CME’s FedWatch tool converts fed funds futures prices into probabilities for each possible rate decision; compare those probabilities just before and just after 08:30 and you can see how far the market actually changed its mind.

Once the bets shift, Treasury yields and the dollar move, and only then do risk assets follow. Bitcoin pays no interest, so when expected risk-free yields rise, the opportunity cost of holding it rises too; when they fall, the reverse. Where each link in that chain slackens or snaps is covered in Fed rates and bitcoin. The point here is narrower: on CPI day, what bitcoin tracks is how far rate expectations moved. If the print matches the forecast and the FedWatch probabilities barely move, the macro layer has given the price no reason to go anywhere in that minute.

Why the wildest minutes cluster around the release: the order book thins out

CPI release dates are published a year ahead. Everyone knows a large unknown will land at 08:30. A resting limit order’s worst outcome is getting filled at a stale price the instant information changes, so ahead of a scheduled event like this, market makers may pull some of their quotes or widen the spread between bid and ask.

The result is perverse but predictable. At the very moment when information is newest and people are most eager to trade, the book holds the fewest orders able to absorb a market order. A sell that would normally be soaked up by the first few levels has to eat through more of them, and the price moves further. The order book depth and impact estimator lets you hold the order size fixed and shrink the resting depth to see how the slippage grows.

Once the number has been digested and quotes come back, depth recovers and the swings usually narrow. On a one-minute chart the release looks like the market lost its mind. Zoom out to the hourly or daily candle and that long wick is sometimes just a brief liquidity gap.

Leverage turns a one-minute jump into a chain reaction

By the eve of a CPI release, the futures market has usually built up positions on both sides: longs betting on a cool print, shorts betting on a hot one. Stop orders and liquidation prices pile up just above and just below the pre-release price.

When the number drops and the price jumps one way, it can run straight into the stops and liquidations stacked on that side. A liquidation is a market order the exchange fires by rule, whatever the price; it hits a book that is already thin, pushes the price further and reaches the next layer of liquidation levels. How liquidation cascades amplify moves takes that mechanism apart step by step.

With positions on both sides, you can also get a two-stage move: the price sweeps up through the shorts, turns, then sweeps down through the longs. It looks like “good data, price fell”, yet the shape may owe more to where positions sat than to the data. The funding rate and open interest tool shows whether open interest built up ahead of the release and whether it dropped sharply afterwards; a sudden drop means a block of positions has been closed out.

Some “CPI moves” have little to do with CPI

CPI makes a very convenient headline. Any big move on release day tends to get pinned on it after the fact. Three checks tell you whether that is fair:

• When did the move start? If most of it happened an hour or two before 08:30 ET, the cause lies elsewhere and the release merely landed in the middle of a trend.
• Did rates move with it? Two-year Treasury yields are sensitive to the expected policy path over the next year or two. If they barely budged after the print and FedWatch probabilities held steady while bitcoin swung hard, macro probably wasn’t the driver; the book and leverage were likely running on their own.
• Was crypto the only thing moving? Calm index futures and a flat dollar alongside a violent bitcoin move point back inside crypto: exchange notices, large on-chain transfers, liquidation data.

The opposite case deserves a note too. When the print misses the forecast badly, yields and the dollar jump, and bitcoin hardly reacts, something inside the crypto market absorbed the macro push at that moment. That tells you more about the market’s structure than any “good news up, bad news down” story.

Matching what you see around 08:30 to the layer behind it

None of this tells you what to do after the number comes out. What it does is let you take one long candle apart afterwards: which stretch was the data changing rate expectations, which was a thin book, and which was positions liquidating each other. Once you can do that, the next “CPI sparks crypto selloff” headline becomes something you can grade for yourself.

Common questions