Arthur Hayes 2026: Long BTC/ETH, Then $22.28M Dumped in Nine Days

On 15 October 2026, Arthur Hayes posted on X that he wasn't trying to guess whether AI would crash — he was betting straight on "the eventual trillion-dollar bailout" and going explicitly long BTC and ETH. Nine days later, on 24 October, he posted the line "I dumped it all" — and on-chain data shows he had cleared roughly $22.28 million in directional positions, including about 6,400 ETH alone (roughly $15.2 million).

Same person, same October: bullish first, flat nine days later. This isn't gossip — it's an excellent case study in how to use a KOL's views. This article lays out, in chronological order, his public statements from June to October 2026, the positions on-chain data caught him actually holding, and the calls he has admitted getting wrong — then offers three reusable rules for reading him.

Two opposite moves in nine days: bullish on 15 October, flat by 24 October

Start with the two original posts (both have clear sources you can check yourself):

1. 15 October: On X he said he wasn't sure whether AI would crash, but rather than bet on the direction itself, he would "trade the eventual trillion-dollar bailout," and wrote that he was long $BTC $ETH, with a chart of CoreWeave's share price attached.
2. 24 October: His exact words were "I dumped it all. The Maelstrom directional book, took profits on $ETH, $ENA, $UNI and other shitcoins. I think we gotta go down to go up. It looks a bit dicey out there in macro and geopolitical land. Going to sit on the sidelines and wait for the all clear signal" — meaning he emptied Maelstrom's directional book, on the grounds that macro and geopolitics "look a bit dicey," choosing to sit on the sidelines and wait for a clear signal.

According to the on-chain monitoring firm Lookonchain, the 24 October sell-off included roughly 6,400 ETH (about $15.2 million), 14 million ENA (about $2.76 million), 323,000 UNI (about $2.24 million), plus ETHFI, AERO, PENDLE and ZAMA, for a total of about $22.28 million.

Yet just nine days earlier, on 15 October, Lookonchain recorded him buying about 546,610 AERO (about $460,000), 120,000 PENDLE (about $232,000) and 3,000,000 ZAMA (about $197,000), for a total of roughly $889,000. In other words, all three of those buys — AERO, PENDLE and ZAMA — were cleared out less than ten days later.

His core macro thesis: AI bubble → trillion-dollar bailout → spillover liquidity

To understand why he dared to go bullish on 15 October and why he bolted nine days later, you first need his macro framework. This thread runs through at least three long posts in 2026:

• "Situationship," 5 August: He argues that AI data-centre capex is essentially "real estate, not technology" — asset-heavy, debt-financed and slow to pay back. He expects capex growth to slow in the second half of 2027 and to decelerate markedly in 2028; and once the AI bubble bursts, both China and the US will be forced to bail things out through bank credit, government guarantees and balance-sheet expansion. That liquidity spillover, he suggests, could push BTC toward his extreme target of $1 million.
• "Safety First," 22 September: He questions the motives behind the "safety first, slow down on AGI" rhetoric from big US AI companies, arguing the real drivers are compute costs, Chinese competition and softening demand; and he notes that data-centre debt is financed through funds and even insurance companies, so the risk could transmit into the traditional financial system.
• "Reality Test," 9 June: This is his earlier version — the AI bubble bursts first, dragging crypto down, BTC follows a "dump then pump" path and finally spikes on bailout liquidity. Notably, in this piece he openly admits that his earlier call that "BTC won't drop below 60,000" was wrong.

String the three together and the logic is coherent: first a decline → policy bailout → liquidity spillover → crypto assets benefit. The 15 October post about "trading the eventual trillion-dollar bailout" is the execution layer of that framework.

His ETH price target: $10,000 — but the number isn't consistent

During Korea Blockchain Week 2026 (KBW 2026), held from 29 September to 1 October 2026, Hayes gave a concrete number in an interview with CoinDesk: ETH at $10,000 by year-end, on the grounds that ETH's market cap and security are enough to carry large capital.

A caveat: that number is reported inconsistently across outlets — some wrote $5,000, others a figure well above $10,000. We only credit the $10,000 figure that traces back to the interview itself, and leave the other discrepancies out. That is also the first thing to do with any KOL's view: go back to the original words, not the retelling.

How to read Hayes: three rules of use

What makes Hayes unusual is that his views can be verified on-chain — because he's a public whale, his moves get monitored. That's both his value and his risk.

1. Separate the "macro framework" from the "trade." His August framework was "down then up," but his October action was "long first, flat after." A framework can last years; a position can reverse in nine days. If you only watch the price levels he calls, you'll miss what he's actually doing.
2. Pay attention to the mistakes he's admitted. Beyond conceding in "Reality Test" that his call that BTC wouldn't drop below 60,000 was wrong, there's a documented reversal in August 2025: he sold out when ETH was around $3,507 and bought back nine days later at about $4,252, leaving copy traders roughly 21% worse off. Both are publicly verifiable.
3. Factor in his "position" as information too. He is both a BitMEX co-founder and Maelstrom's CIO, and his views are tightly bound to his own fund's directional book. Separately, BitMEX permanently ceased operations on 23 September 2026; on 23 July 2026, two former customers filed a proposed class action in the Southern District of New York alleging the exchange ran an undisclosed "Insider Trading Desk" and froze servers during sharp volatility to trigger liquidations, naming Hayes and co-founders Samuel Reed, Benjamin Delo and Gregory Dwyer among the defendants. These are background facts — they don't affect his macro read, but they affect how much weight you should give him.

As for the new concept in his latest piece, "The Trifecta," on 27 October — building a spot plus derivatives market for AI compute (FLOPS) on top of his not-yet-launched Flop Network (FCU forwards, the FRCR index, Quanto quarterly futures and perpetuals), which he calls his third bet after "perpetual contracts" and "synthetic dollar stablecoins" — this is an early-stage idea with no live product, and is best treated simply as a window into where he's looking.

To read "institutional narratives" against "real on-chain data," see the institutional-product moves in BlackRock's filing for a bitcoin "downside protection" ETF, and the real usage of AI-plus-crypto in AI agents made 138 million on-chain payments in a year — those two threads are exactly the two verifiable sides of Hayes's macro framework.

FAQ

Where are Arthur Hayes's official website and blog?

The official blog is on Substack at cryptohayes.substack.com, titled "Crypto Trader Digest"; the Archive page lists every post with its date. His X account is @CryptoHayes. One warning: hayes.co is not his official site — the domain currently shows a for-sale page, so don't treat it as a source.

What exactly is his ETH price target?

The figure traceable to the interview itself is "$10,000 by the end of 2026" (around 2026-09-30, with CoinDesk during KBW). The $5,000 or higher numbers relayed by some outlets could not be traced to an original source, so they aren't used here.

Can I just copy his positions?

Not advisable. His disclosures rely on on-chain monitoring and therefore lag — the AERO, PENDLE and ZAMA he bought on 15 October were all cleared out by 24 October; and that August 2025 "sell, then buy back" left copy traders about 21% worse off. The right way to use on-chain data is to understand what he's thinking, not to replicate his moves.

Is BitMEX's shutdown related to his views?

Not directly. BitMEX announced in July 2026 that it would permanently cease operations on 23 September; the class action filed around the same time concerns allegations about how the exchange was run. His current public identity centres on being Maelstrom's CIO, and according to media reports the family office is raising capital for its private equity fund. That's background, separate from his calls on macro and crypto prices.

Is his "BTC to $1 million after the AI bubble bursts" credible?

It is a conditional call: the premise is that the AI bubble really does burst and that China and the US really do bail things out at scale through bank credit and government guarantees. If the premise doesn't hold, the conclusion is moot. A more practical reading is to treat it as a "scenario exercise" rather than a "price forecast," focusing on whether his chain of reasoning holds rather than on the number itself.

Risk notice: Cryptoasset prices are highly volatile and you may lose your entire principal; this article does not constitute investment advice. The Hayes views, positions and price targets discussed here are relayed from his public blog, X posts and public interviews; the on-chain position data comes via the third-party monitoring firm Lookonchain and may differ by methodology. Details of the BitMEX shutdown and related litigation are subject to the exchange's announcements and court filings. Acting on any celebrity's positions can cause significant losses; judge for yourself.