BitMEX co-founder Arthur Hayes has warned that heavy share selling following major artificial intelligence IPOs could weaken stock markets, restrict dollar liquidity and put pressure on Bitcoin before any potential central bank intervention.

  • BitMEX co-founder Arthur Hayes warns that large share unlocks following major AI IPOs could put pressure on technology stocks and financing costs.
  • Hayes estimates that Anthropic’s early public investors could face what he calls 83% dilution by June 2027, although this is his interpretation of the expected share supply.
  • He expects falling AI stocks to make lenders more cautious, potentially slowing bank credit growth and reducing dollar liquidity available to Bitcoin.
  • His investment firm, Maelstrom, plans to remain on the sidelines until liquidity conditions improve.
  • Hayes believes a severe AI financing crisis could eventually prompt central banks to ease monetary policy, potentially benefiting Bitcoin over the longer term.

His latest essay, D is for Dilution, examines the shares that could become available for trading after the public listings of SpaceX, Anthropic and OpenAI. Hayes argues that investors may struggle to absorb the additional supply, particularly as these companies continue raising money for expensive AI infrastructure.

He expects falling share prices to make borrowing more expensive for technology companies, potentially affecting the banks financing their expansion. His investment firm, Maelstrom, plans to remain on the sidelines until conditions improve.

Arthur Hayes sees 83% dilution risk for Anthropic investors

Anthropic features prominently in Hayes’s analysis because of its planned stock market debut and the amount of money required to develop its AI models.

Using the company’s updated registration statement, Arthur Hayes estimates that early public investors could face what he describes as 83% dilution by June 2027.

The forecast concerns the large number of shares that could become available after restrictions on employees and existing investors expire. However, previously issued shares becoming tradable do not necessarily represent new equity issuance.

Anthropic’s November 2 filing outlines several share-release dates, with certain restrictions lasting 180 days and others extending to 365 days after the offering.

Hayes argues that selling by existing shareholders could leave new investors facing losses, even if the company continues reporting revenue growth.

Assuming an initial valuation of $1.5 trillion, he estimates that approximately $1.245 trillion in additional buying would be needed to absorb the potential share supply without a decline in price.

That calculation is Hayes’s hypothetical estimate, not a financing requirement disclosed by Anthropic.

The company’s filing separately reports approximately $895.4 billion in AI infrastructure commitments, demonstrating the scale of its future spending obligations.

SpaceX and OpenAI face similar concerns over share supply

SpaceX provides another example in Hayes’s argument, with the investor claiming its publicly traded share supply has grown considerably since its IPO.

He estimates that another 8.6 billion shares could become eligible for trading by June 2027, equivalent to roughly half of the existing share count.

Arthur Hayes believes additional supply could keep the stock under pressure unless SpaceX generates enough cash to support shareholder returns.

He raises the possibility of further dilution from a rumored merger with Tesla, although no completed transaction establishes that outcome.

OpenAI faces a less certain timetable because it has not publicly filed an IPO registration statement.

Hayes expects the company could initially offer approximately 5% of its shares and subsequently allow employees and early investors to sell additional holdings.

His projections assume a January listing and a valuation approaching $1.5 trillion. Neither assumption represents an announced offering plan.

The investor argues that public listings would give the companies more direct access to corporate debt markets as they fund their computing needs.

AI borrowing costs could reduce dollar liquidity

Arthur Hayes’s main concern is what happens when falling technology shares begin affecting the companies’ ability to borrow.

Investors buying corporate bonds may demand higher yields if they become less confident in AI companies’ future earnings and share prices.

Higher borrowing costs would make financing data centers and other computing infrastructure more expensive, potentially forcing companies to reconsider planned spending.

As crypto.news previously reported, Hayes has warned that weaker AI demand could create problems in credit markets before governments respond with financial support.

His latest argument focuses more closely on share unlocks as a possible trigger for those financing difficulties.

Banks currently provide loans to AI businesses, finance stock purchases and hold securities connected to the technology sector.

Hayes believes that declining confidence in AI investments could slow those activities, limiting the amount of new credit entering the financial system.

He pointed to a recent weekly decline in U.S. commercial bank balance sheets, while acknowledging that one reading does not establish a lasting trend.

Under his scenario, weaker bank lending would reduce dollar liquidity and create difficult trading conditions for Bitcoin and other cryptocurrencies.

Bitcoin outlook depends on central bank response

Despite his concerns about the immediate market environment, Hayes remains optimistic about Bitcoin’s longer-term prospects.

Arthur Hayes expects authorities to intervene if problems financing AI infrastructure develop into a more serious financial crisis.

Such intervention could involve monetary easing or other measures that increase the supply of money available throughout financial markets.

Hayes believes that additional liquidity would eventually support Bitcoin, although the timing and form of any policy response remain uncertain.

His previous analysis linked AI credit risks to a longer-term Bitcoin forecast of $1 million by 2030.

The investor has not identified a date when central banks might intervene or established that a financial rescue will occur.

Meanwhile, Maelstrom intends to wait for clearer evidence of improving liquidity before increasing its market exposure.

Beyond Bitcoin, Hayes identified FLOP Network and Genlayer as projects he expects to benefit from cheaper computing power. He disclosed that Maelstrom advises Genlayer, which is preparing for a token generation event.