Arthur Hayes ties $1M Bitcoin-by-2030 call to late-2027 AI credit stress
He argues data-center debt losses could force a policy backstop that turns into a liquidity bid for BTC.
Maelstrom CIO Arthur Hayes reiterated his forecast that Bitcoin could reach $1 million by 2030, and he put the strongest leg of that move in late 2027 or early 2028. His timing hinges on a debt-funded AI infrastructure buildout running into credit stress and prompting a liquidity-expanding policy response.
Key Takeaways
- Arthur Hayes reiterated a $1 million Bitcoin target for 2030 and anchored the strongest rally to late 2027 or early 2028.
- The trigger he’s pointing to is credit stress in debt-funded AI infrastructure if data-center revenues fail to cover construction and compute-equipment commitments.
- Hayes framed the AI buildout as a “credit story like 2008 and not an earnings story like 2000,” putting the focus on leverage and refinancing mechanics.
- Apollo’s Torsten Slok put the potential scale at more than $2 trillion of additional investment-grade AI debt, with AI-related borrowing near 40% of longer-duration investment-grade corporate bond supply (data through July).
Hayes Repeats $1M Bitcoin Target, Pins the Big Move to Late 2027/Early 2028
Arthur Hayes is keeping his $1 million Bitcoin-by-2030 forecast on the table, but the more actionable part of the call is the calendar. He expects the strongest advance to land in late 2027 or early 2028, not as an extension of today’s cycle but as a second-order effect of stress in AI infrastructure financing.
The setup he describes is straightforward: data centers and compute buildouts are being financed with large amounts of debt, and the cash flows have to arrive on schedule. If data-center earnings do not cover the sums committed to construction and computing equipment, losses move from the projects into the balance sheets that funded them.
Hayes has also been explicit about what he does not know. In an earlier scenario referenced alongside the thesis, he put Bitcoin in a $60,000 to $70,000 trading range with possible downside toward $50,000 before any eventual move toward $1 million, and he said he could not identify the borrower that would trigger a crisis or Bitcoin’s exact bottom.
The Causal Chain: AI Data-Center Debt Stress → Policy Backstop → Liquidity Bid for BTC
Hayes’s mechanism is a credit-to-liquidity story, not a “AI stocks crash, Bitcoin pumps” slogan. He argues the AI buildout boom is a “credit story like 2008 and not an earnings story like 2000.” The distinction matters because it shifts the failure mode from public-market multiples to repayment schedules, collateral values, and who is warehousing the risk.
On the asset side, the buildout includes land, buildings, power and cooling connections, and processors that can lose value as newer equipment becomes cheaper and more efficient. On the liability side, the repayment schedules are longer. Hayes’s timing window comes from that mismatch: hardware can age and depreciate while the debt remains priced off earlier, higher revenue expectations.
From there, the chain runs through the institutions that hold the paper. Hayes’s scenario is not limited to tech equities falling. He has argued that banks, insurers, private lenders, and infrastructure investors could take losses if projects fail to generate enough cash flow to meet interest payments, leases, and other obligations, even if the strongest technology companies remain profitable.
The final link is policy. In his “Safety First” essay, Hayes outlined two U.S. backstop paths: Washington becomes a “compute buyer of last resort,” purchasing computing capacity to support the industry, or Washington provides financial assistance to insurers facing losses on AI-linked debt. Either route, in his framing, increases the supply of money and becomes a liquidity tailwind for Bitcoin. As of the Sep. 22 reference in the packet, neither measure had been announced.
Credit-Market Scale Check: Apollo’s $2T+ AI Debt Math and the Private-Credit/Insurer Angle
The reason this narrative keeps resurfacing is that the credit numbers are large enough to matter outside a single sector. In an Aug. 14 note, Apollo chief economist Torsten Slok estimated the AI ecosystem could support more than $2 trillion of additional investment-grade debt, meaning bonds or loans rated as relatively low default risk and typically eligible for conservative institutional mandates.
Apollo also argued the public investment-grade market might absorb less than $1 trillion through 2030 due to issuer concentration and credit-rating limits. The implication is that more than $1 trillion of the financing demand could shift into private placements, infrastructure lending, equipment financing, and project-specific structures. Those channels sit closer to private credit, meaning loans made outside public bond markets with less transparency and different liquidity terms.
Apollo’s snapshot of current flow adds a near-term anchor: using data through July, it said AI-related borrowing already represented nearly 40% of longer-duration investment-grade corporate bond supply. That does not prove a blow-up, but it does quantify how much of the marginal duration in credit has been tied to the AI buildout.
The insurer angle is where Hayes’s macro story meets a real reporting timeline. The National Association of Insurance Commissioners (NAIC), a U.S. standard-setting body coordinating state insurance regulators, has flagged liquidity, pricing, and transparency concerns in private credit. It has pointed to valuation and lending-standard questions and noted that some retail private credit funds have faced withdrawal requests and used withdrawal limits, while also saying those developments do not necessarily establish deterioration across private credit markets or insurers’ holdings.
Two NAIC rule changes matter for this thesis because they can surface exposures that are otherwise hard to map. Under amendments adopted in 2025, NAIC requires private rating rationale reports within 90 days of an annual update or rating change, and it says those reports must contain “analytical substance.” Separately, NAIC’s Statutory Accounting Principles Working Group adopted changes effective at year-end 2026 to improve reporting of insurers’ private credit holdings.
What Would Make This Tradeable: The 2026–2028 Signposts to Track in AI Credit and BTC
The first hard waypoint is year-end 2026, when NAIC’s statutory accounting and reporting changes for insurers’ private credit holdings take effect for annual financial filings. If those disclosures begin to break out AI and data-center exposure more cleanly, it gives the market a better map of where losses could land and who might need liquidity.
The second is the flow data in credit itself. Apollo’s framing makes two metrics worth tracking: growth in AI-related investment-grade issuance and whether AI continues to represent an outsized share of longer-duration investment-grade corporate bond supply. If that share stays elevated while revenue assumptions soften, the refinancing and downgrade path gets tighter.
The third is policy signaling that resembles Hayes’s backstops. Any move toward the government acting as a “compute buyer of last resort,” or any discussion of financial assistance tied to insurer losses on AI-linked debt, would be a concrete step from narrative to mechanism.
The last signpost is Bitcoin’s own behavior versus Hayes’s earlier range markers. He previously sketched BTC between $60,000 and $70,000 with downside risk toward $50,000 before an eventual advance, so deviations from that band matter mainly as a check on whether the market is front-running a liquidity story years early.
My Read: This Is a Liquidity-Timing Thesis, Not a Near-Term Price Call
The part that makes Hayes’s framing useful is that it is time-anchored. Late 2027 and early 2028 are not vibes, they are his window for when the asset-liability mismatch in AI infrastructure financing could start forcing decisions, and when a policy response would be most likely to show up as liquidity rather than as a clean, contained default cycle.
The threshold that matters is whether AI financing stress migrates from project-level misses into institutions that are politically and systemically hard to let fail, like insurers holding opaque private credit. If the NAIC’s year-end 2026 reporting changes surface concentrated AI exposure and the policy conversation shifts toward explicit backstops, the setup starts to look structural rather than narrative-driven.
