Bitcoin skyrocketed past $86,000 on a $262 million squeeze. Can it hit $87,000 by Friday?

Will bitcoin close below $87,000 on Friday?

The plan, posted before it settles

The goal is the full dollar. This is written to be held to Friday's settle, where a correct call pays $1.00 a contract. Everything else is a way out, not the plan.

1. Hold to $1.00. The default and the intention. If bitcoin is under $87,000 at 5 PM Eastern on Friday, every contract settles at a dollar and the position returns $7,795.82 with no exit fee, because settlement is fee-free. That is a profit of $2,795.82.
2. Sell at 80¢, optional. An order rests at 80¢. If bitcoin breaks down early and the contract reprices, it fills and the position closes for $1,149.34 rather than waiting out the week. We do not need it to fill, and we will not chase it lower to make it fill.
3. Stop out above $87,000, if it happens in market hours. If bitcoin trades through the strike while US equities are open, the thesis is wrong in the window that matters and the position is sold rather than held to zero. The full arithmetic is below.

The economics, per contract

A Kalshi contract pays exactly $1.00 if the answer is yes for the side you hold, and nothing if it is not. So the arithmetic does not depend on how much anyone puts on it, and neither does anything below. Everything here is per contract and in percentages, which is the only honest way to publish a trade idea other people might size differently.

The book showed a 62¢ NO ask when we read it.¹ The 64.13¢ we quote is higher because an order of any real size walks up through more than one price level and because Kalshi's taker fee is baked into the average it shows.² We quote the worse number rather than the screen price, because the screen price is not what anyone gets.

Disclosure. The desk holds this position. Every trade idea on this site is one the desk has taken itself and grades on Kalshi's own settlement afterwards, win or lose. Sizes are not published, because the arithmetic above is the same at any size and ours is not a recommendation about yours.

It is also on Exclusive Tuna's record, where it will be graded on Kalshi's own settlement, win or lose, next to every other call this desk has made.

What the market says right now

These prices update themselves every ten seconds from Kalshi's public API while this page is open. A figure that moves flashes green or red for a moment and carries an arrow, so you can see the board is live. The paragraph below was written against the 10:08 AM PT read.

The board is priced sensibly around spot: with bitcoin at $86,107.44 the coin flip sits right at the $86,000 strike, which is where it belongs. The $87,000 line has traded 8,314 contracts with 4,350 still open, and there were 4,657 contracts bid on the YES side when we read it,¹ so there is a real market here rather than a quote nobody will honour.

Why bitcoin went up, which matters more than that it did

Nothing was announced. No company bought anything, no agency approved anything, no bill passed. Three mechanical things happened at once, and all three are the kind that stop.

1. A 45-week signal fired on the weekly close. Bitcoin ended last week at $81,159, just above its 50-week moving average of $78,786. That was its first weekly close above that line in 45 weeks.⁷ Systematic money watches that line, so the close itself is what triggered the buying, not anything that happened on Monday. This is the weekend part of the story, and it is a chart event rather than a news event.
2. Shorts were forced out, fast. As bitcoin went through $84,000, CoinGlass counted $262.30 million of short positions liquidated in a single hour.⁸ That is not people choosing to buy bitcoin. It is people being made to. The full day is below, and it is the heart of this article.
3. The macro tape helped. Brent crude fell for a fourth straight day to about $102, and the 10-year Treasury yield slipped under 5%, which is the ordinary backdrop for risk assets catching a bid.⁷

The same leverage has now been flushed twice in six days

This is the part worth sitting with. Bitcoin's crash last Monday and its rally this Monday were the same event pointed in opposite directions: a crowd that was too levered getting closed out by an exchange rather than by choice.

September 15 figures from CoinDesk's coverage of the vote.¹¹ September 21 figures from CoinDesk, which also had the most recent hour running $159.9 million, 95% of it shorts.¹⁰ Both are 24-hour windows across all venues.

Two numbers that change how we read it

The first is CoinDesk's own summary of the day, which we could not put better: the move is being driven by forced buying more than fresh conviction.¹⁰ Trading volume was up 39% to $224 billion, which is what a liquidation cascade looks like from the outside.

The second is the one that actually matters for Friday. Open interest went up 7.59% to $156 billion while all of this was being closed out, and bitcoin futures open interest reached 700,000 BTC for the first time in weeks.¹⁰ Read that twice. Positions were not being reduced. They were being replaced, at the highs, by someone new.

The crypto equities, and what they are telling us

Bitcoin trades all weekend. The stocks that track it do not, so Monday's US open is the first time the leveraged crowd can act on a weekend move. That makes the crypto equity complex worth watching on a day like this one, in both directions.

Moves and prices as reported Monday, Sep 21, 2026, with bitcoin at $85,131 and up 6% on the day.⁹ CLSK (CleanSpark), HOOD (Robinhood), BTDR (Bitdeer), HUT, WULF and CIFR are the rest of the complex; we do not have sourced closes for those and are not going to make them up.

Strategy moves with bitcoin because it holds bitcoin. The miners move through hash rate, energy cost and what they actually mined, which is why they lag on an up day.⁹ Coinbase ran hardest of all, which is what happens when volume is the product and everybody is trading at once.

Where it still matters: these names are how leveraged crypto exposure gets expressed by people who cannot or will not hold the coin, and their options chains are deep. After a run this size, call sellers and profit-takers in MSTR and COIN are a genuine source of supply, and four more US sessions stand between now and Friday's 5 PM Eastern settle, which itself lands in the middle of one.

Three reasons we are on the under

1. Bitcoin has not printed $87,000 once in 350 hours. From September 7 at 04:00 UTC through this morning, across 350 hourly Coinbase Exchange candles, the highest price bitcoin reached is $86,354.94, set at 15:00 UTC today.³ Not a close above $87,000. Not a wick. It has never been there in this move. For the YES side to pay, bitcoin has to travel $892.56, which is 1.04%, into air it has not been in, and then still be there on Friday afternoon.
2. The 14-period RSI on hourly closes is 85.5. Computed from Coinbase hourly closes through the hour ending 17:00 UTC on September 21.³ The conventional overbought threshold is 70. Worth being honest about the timeframe: on the 15-minute chart the same indicator reads about 65, which is ordinary. The stretch is in the hourly and daily picture, not the last two hours.
3. The move is six days old and 15% tall. Bitcoin bottomed at $74,887.50 at 18:00 UTC on September 15. From that low to $86,107.44 is 15.0%. The 24-hour range alone runs $80,570.77 to $86,354.94, a 7.2% span in a single day.³ Moves that shape do not usually add a clean extra leg without resting first.

Still bullish underneath, and that is the point

On September 15 the US Senate failed to advance the CLARITY Act on a cloture vote, ending the crypto industry's market-structure push for 2026.⁵ Bitcoin went to $74,887.50 that evening. Six days later the entire decline is recovered and the market is making new highs for the month anyway. That is a tape absorbing its worst regulatory headline of the year and going up, which is bullish, not bearish.

Contemporaneous reporting also cuts against the tidy story about that dump. CoinDesk's live coverage that day described thin leverage, funding drifting toward zero and futures premiums below 5%,⁴ which means traders had already stepped back before the vote rather than being flushed out by it. We mention it because the neat version, a leveraged long wipeout at the lows, is not what the reporting says happened.

So the call is narrow on purpose. It is not a top call and it is not a trend call. It is a bet that a vertical six-day move does not print a brand new high and hold it through 5 PM Eastern on Friday.

The case against our own plan

• The entry is above the market's own price. Kalshi prices NO here at about 61% to 62%. An entry of 64.13¢ with the fee in puts break-even at 64.14%. Crossing the spread with size costs roughly two and a half points, so this only works if the market is underpricing NO, not merely if NO is the better side.
• Four days is a long time in bitcoin. This does not settle for roughly 100 hours, and $87,000 is 1.04% away. A single strong session covers that.
• Overbought is a condition, not a signal. RSI can sit above 80 for days in a real trend. It says the move is stretched. It does not say when it stops.
• Untested air cuts both ways. Nobody is sitting on a losing position above $87,000 either, because nobody has held one there. A clean break has nothing overhead to slow it down.
• A stop can take us out of a winner. The rule fires on bitcoin touching $87,000 in market hours, not on Friday's settle. Bitcoin can trade through $87,000 on Tuesday, sell off on Thursday and settle under the line, and we would have taken a 28% loss on a call that was right. That is the price of not risking the whole stake.
• The exit is thin. To reach the 80¢ sell, bitcoin probably has to be back near $84,000 with the clock running out. There is no gentle version of this where we get paid for being slightly right.

The arithmetic

Kalshi's taker fee is ceil(0.07 × contracts × price × (1 - price)), rounded up to the cent, charged going in and again on any sale. Settlement carries no fee.⁶ In dollar mode the entry fee is already inside the 64.13¢ average, so that number is the whole cost of getting in.

The stop, priced before it can be needed

A stop nobody wrote down in advance is a decision made in a bad moment. This one is written down: if bitcoin trades above $87,000 on Coinbase while the US stock market is open, 6:30 AM to 1:00 PM Pacific, the position is sold. Not held, not averaged, sold.

Where would it fill? At $87,000 the contract is sitting on its own strike, so NO should be near a coin flip. The check available today: when bitcoin was $86,107 this morning, the $86,000 strike, also sitting on its own strike, was bid 48¢.¹ Selling into a bid rather than at it is the honest assumption, so 48¢ is the number we publish now, before it can be tuned after the fact.

That is the whole case for having one. Being wrong costs about a quarter of the stake instead of all of it, and the difference between those two outcomes is a rule written before the event rather than a judgement made during it. The cost is real and worth naming: a stop can take the idea out of a position that would have come back, and on a contract this close to its strike that will happen sometimes.

What this plan costs at your stake

Your money, not ours. Pick a stake and the slip works out how many contracts it buys at the 62¢ ask that was on the book when we read it, what Kalshi's entry fee takes, and what each outcome pays. We paid more than 62¢ because an order of real size walks up through the book; a smaller one should not have to.

55.9% up against 100% down, which is exactly why the break-even sits at 64% and not 50%. Those percentages hold at any size, which is the point of publishing them this way. It is not a recommendation to put money on it, and it is not advice.

What actually settles this market

The simple average of the sixty seconds of CF Benchmarks' Bitcoin Real-Time Index immediately before 5 PM Eastern on September 25, 2026. Not a Coinbase print, not an hourly candle close, not the price on any single exchange. The ticker is KXBTCD-26SEP2517-T86999.99 and trading closes at 5 PM Eastern.¹

Note the clock. 5 PM Eastern is 2 PM Pacific. Every other time on this site is Pacific; this market is named in Eastern because Kalshi names it that way. If you are on the West Coast, this resolves over lunch.

The same index runs the perpetual

Kalshi's BTC perpetual, KXBTCPERP, marks against the same CF Benchmarks index that settles the Friday contract, so a view on one is a view on the other. One contract is 0.0001 BTC.

The difference matters for what you are choosing between. The Friday contract is a fixed question with a fixed end: you know your worst case the moment you enter, and it is what you paid. A perpetual has no expiry and carries leverage, which means the position can be closed against you before Friday ever arrives even if the eventual call is right. Being right late is the same as being wrong on a leveraged position. That is the trade-off, in both directions, and it is worth understanding before you pick one.

Sign-up offers change; Kalshi shows the current terms on the sign-up page. 21+. Not available in every state.

PM Countdown is not affiliated with or endorsed by Kalshi. Nothing here is financial advice, and we are not licensed advisers. Prediction markets can lose you money fast.

New to the mechanics? Read how Kalshi prices and fees work and setting a sell target first, then watch a live window on the BTC 15-minute desk or the hourly strike ladder before putting money on anything.

Quick answers

Will bitcoin close below $87,000 on Friday, September 25, 2026?

The market says it probably will. Kalshi priced the $87,000-or-above contract at 39¢ YES and 62¢ NO at 10:08 AM PT on September 21, an implied 38% to 39% chance of settling at or above $87,000. Bitcoin was $86,107.44 on Coinbase Exchange, about $893 under the line. Our featured trade idea takes the NO side at 64.13¢.

Why did bitcoin go up on September 21, 2026?

No announcement: three mechanical things at once. Bitcoin closed the previous week at $81,159, its first weekly close above the 50-week moving average of $78,786 in 45 weeks, which is a signal systematic money acts on. As it cleared $84,000, CoinGlass counted $262.30 million of short positions liquidated in a single hour and about $648 million across the day. Brent crude fell a fourth straight day to around $102 and the 10-year Treasury yield slipped below 5%. A squeeze is short covering rather than new demand, which is the basis of our position.

How much was liquidated in the September 21, 2026 bitcoin short squeeze?

$746.6 million across all crypto in 24 hours, of which $647.9 million, or 86.8%, was short positions. Bitcoin shorts accounted for $277.5 million and ether shorts $122.8 million. CoinGlass counted $262.30 million liquidated in a single hour as bitcoin cleared $84,000, and the most recent hour CoinDesk reported ran $159.9 million at 95% shorts. Open interest still rose 7.59% to $156 billion over the same period, so the positions were being replaced rather than reduced.

How did crypto stocks like MSTR, COIN, MARA and RIOT trade on September 21, 2026?

All higher, and most of them harder than bitcoin. Coinbase (COIN) rose 11.66% to $194.25, Strategy (MSTR) 6% to $163.68, MARA 5% to $13.84, Riot (RIOT) 4% to $24.80 and the CoinShares mining ETF (WGMI) 3% to $51, with bitcoin around $85,131 at the time. Strategy tracks bitcoin because it holds it; the miners move through hash rate, energy cost and mined output, which is why they lagged.

Has bitcoin ever traded at $87,000?

Not in this run. Across 350 hourly Coinbase candles from September 7 to September 21, 2026, bitcoin did not print $87,000 in a single hour. The highest is $86,354.94 at 15:00 UTC on September 21. YES needs a brand new high for the move, held into Friday afternoon.

What exactly settles it?

The simple average of the sixty seconds of CF Benchmarks' Bitcoin Real-Time Index before 5 PM Eastern on September 25, 2026. The ticker is KXBTCD-26SEP2517-T86999.99. Trading closes at 5 PM Eastern, which is 2 PM Pacific.

Is bitcoin overbought right now?

On the hourly, by the usual reading, yes: the 14-period RSI on Coinbase hourly closes was 85.5 through 17:00 UTC on September 21, and above 70 is conventionally overbought. On the 15-minute chart it reads about 65, which is ordinary. Bitcoin had gained 15.0% from its September 15 low of $74,887.50 in six days.

What happened on September 15?

The US Senate failed to advance the CLARITY Act on a cloture vote, ending the crypto industry's market-structure push for 2026. Bitcoin fell to $74,887.50 on Coinbase Exchange at 18:00 UTC that day and has since recovered the whole decline.

What are the returns on this trade idea?

Entered at 64.13¢ with Kalshi's fee included. Held to a correct settlement each contract pays $1.00, a gain of 35.86¢ or 55.9%. If the 80¢ sell order fills, 14.74¢ or 23.0%. If the stop fires near 48¢, a loss of 17.88¢ or 27.9%. Wrong and held, the contract expires at zero and the whole entry is lost. Break-even is 64.14%. Those percentages are the same at any size.

Where is this call posted publicly?

On Exclusive Tuna's record here, and on the desk's X account @exclusivetunapm, where the entry went up with the order confirmation at 10:04 AM PT on September 21, four days before the market settles.