BitMEX co‑founder Arthur Hayes warns that the big wave of share sell‑offs expected after major AI IPOs could ripple across markets — squeezing dollar liquidity, tightening corporate borrowing and putting downward pressure on Bitcoin before any central‑bank rescue arrives.
In his new essay, "D is for Dilution," Hayes zooms in on the post‑IPO share unlocks that could hit the market after listings from Anthropic, OpenAI and SpaceX. His thesis: the combination of massive employee and early‑investor stock releases, plus ongoing cash needs for costly AI infrastructure, could swamp demand, push tech share prices lower and make borrowing more expensive — which in turn could tighten credit and reduce liquidity for the whole financial system, including crypto.
Key takeaways and figures from Hayes’s analysis (all estimates or scenarios he lays out, not company disclosures):
- Anthropic: Using the company’s updated registration statement, Hayes estimates early public investors could face what he calls up to 83% dilution by June 2027 due to share unlock schedules (some restrictions lift after 180 days, others after 365). He stresses that unlocks make previously issued shares tradable — they aren’t necessarily new equity — but large sell pressure could still depress prices. Assuming a hypothetical $1.5 trillion valuation, Hayes calculates roughly $1.245 trillion of incremental buying would be needed to soak up that potential supply without price declines. Separately, Anthropic’s filing shows about $895.4 billion in AI infrastructure commitments, underlining its heavy future spending needs.
- SpaceX: Hayes says the company’s publicly tradable share pool has grown materially since its IPO. He estimates roughly 8.6 billion shares could become eligible to trade by June 2027 — about half the current share count — which could keep pressure on the stock unless SpaceX generates substantial cash returns. He also notes speculation about a Tesla merger as a potential additional dilution vector, though no deal is confirmed.
- OpenAI: With no public IPO filing yet, Hayes treats OpenAI’s timetable as uncertain. He models a possible initial float around 5% and subsequent unlocks of employee and early investor stakes, using a hypothetical January listing and a valuation near $1.5 trillion — again, illustrative scenarios rather than announced plans.
How that spills into credit and crypto
Hayes’s core concern is the knock‑on effect from equity weakness to credit markets. If investors grow less confident in AI companies’ earnings and equity values, corporate bond buyers could demand higher yields. Higher borrowing costs would raise the price of financing data centers and other heavy compute investments, potentially forcing AI companies to scale back projects or raise even more capital. Banks that lend to these companies, finance stock purchases or hold tech‑linked securities could slow lending activity if confidence falters.
Hayes points to a recent weekly drop in U.S. commercial bank balance sheets as an early sign that credit could tighten — though he acknowledges one data point doesn’t prove a trend. Under his scenario, reduced bank lending and weaker liquidity would create "difficult trading conditions" for Bitcoin and other cryptocurrencies in the near term.
But Hayes is also long‑term constructive on Bitcoin. He argues that if credit stress from AI financing triggers broader market weakness, authorities may ultimately step in with monetary easing or other liquidity measures. That intervention, whenever it arrives, could flood markets with cash and become a tailwind for Bitcoin. Hayes has previously tied similar dynamics to a long-term Bitcoin projection of $1 million by 2030, though he has not set any date for when central banks might act or guaranteed a rescue.
What Maelstrom is doing
Hayes says his investment firm, Maelstrom, will largely sit on the sidelines until liquidity conditions improve. He also flagged a couple of projects he expects to benefit from cheaper compute — FLOP Network and Genlayer — and disclosed that Maelstrom advises Genlayer, which is planning a token generation event.
Bottom line
Hayes’s warning centers on a plausible chain of events: large post‑IPO share unlocks → equity weakness → higher corporate borrowing costs → tighter bank lending and dollar liquidity → near‑term pain for risk assets including Bitcoin — followed, potentially, by policy easing that restores liquidity and supports crypto later on. Much of his analysis relies on hypothetical valuations and timing assumptions, and several key details (especially around OpenAI) remain unfiled or speculative. Still, his essay highlights how big AI listings could stress both equity and credit markets — a development crypto traders should watch closely.
In his new essay, "D is for Dilution," Hayes zooms in on the post‑IPO share unlocks that could hit the market after listings from Anthropic, OpenAI and SpaceX. His thesis: the combination of massive employee and early‑investor stock releases, plus ongoing cash needs for costly AI infrastructure, could swamp demand, push tech share prices lower and make borrowing more expensive — which in turn could tighten credit and reduce liquidity for the whole financial system, including crypto.
Key takeaways and figures from Hayes’s analysis (all estimates or scenarios he lays out, not company disclosures):
- Anthropic: Using the company’s updated registration statement, Hayes estimates early public investors could face what he calls up to 83% dilution by June 2027 due to share unlock schedules (some restrictions lift after 180 days, others after 365). He stresses that unlocks make previously issued shares tradable — they aren’t necessarily new equity — but large sell pressure could still depress prices. Assuming a hypothetical $1.5 trillion valuation, Hayes calculates roughly $1.245 trillion of incremental buying would be needed to soak up that potential supply without price declines. Separately, Anthropic’s filing shows about $895.4 billion in AI infrastructure commitments, underlining its heavy future spending needs.
- SpaceX: Hayes says the company’s publicly tradable share pool has grown materially since its IPO. He estimates roughly 8.6 billion shares could become eligible to trade by June 2027 — about half the current share count — which could keep pressure on the stock unless SpaceX generates substantial cash returns. He also notes speculation about a Tesla merger as a potential additional dilution vector, though no deal is confirmed.
- OpenAI: With no public IPO filing yet, Hayes treats OpenAI’s timetable as uncertain. He models a possible initial float around 5% and subsequent unlocks of employee and early investor stakes, using a hypothetical January listing and a valuation near $1.5 trillion — again, illustrative scenarios rather than announced plans.
How that spills into credit and crypto
Hayes’s core concern is the knock‑on effect from equity weakness to credit markets. If investors grow less confident in AI companies’ earnings and equity values, corporate bond buyers could demand higher yields. Higher borrowing costs would raise the price of financing data centers and other heavy compute investments, potentially forcing AI companies to scale back projects or raise even more capital. Banks that lend to these companies, finance stock purchases or hold tech‑linked securities could slow lending activity if confidence falters.
Hayes points to a recent weekly drop in U.S. commercial bank balance sheets as an early sign that credit could tighten — though he acknowledges one data point doesn’t prove a trend. Under his scenario, reduced bank lending and weaker liquidity would create "difficult trading conditions" for Bitcoin and other cryptocurrencies in the near term.
But Hayes is also long‑term constructive on Bitcoin. He argues that if credit stress from AI financing triggers broader market weakness, authorities may ultimately step in with monetary easing or other liquidity measures. That intervention, whenever it arrives, could flood markets with cash and become a tailwind for Bitcoin. Hayes has previously tied similar dynamics to a long-term Bitcoin projection of $1 million by 2030, though he has not set any date for when central banks might act or guaranteed a rescue.
What Maelstrom is doing
Hayes says his investment firm, Maelstrom, will largely sit on the sidelines until liquidity conditions improve. He also flagged a couple of projects he expects to benefit from cheaper compute — FLOP Network and Genlayer — and disclosed that Maelstrom advises Genlayer, which is planning a token generation event.
Bottom line
Hayes’s warning centers on a plausible chain of events: large post‑IPO share unlocks → equity weakness → higher corporate borrowing costs → tighter bank lending and dollar liquidity → near‑term pain for risk assets including Bitcoin — followed, potentially, by policy easing that restores liquidity and supports crypto later on. Much of his analysis relies on hypothetical valuations and timing assumptions, and several key details (especially around OpenAI) remain unfiled or speculative. Still, his essay highlights how big AI listings could stress both equity and credit markets — a development crypto traders should watch closely.