The question of whether the AI funding boom is starting to crack has moved to the forefront after the Financial Times reported two new setbacks for the sector. OpenAI has cut its funding target by $10 billion and now seeks $20 billion at a $1.25 trillion valuation, down from $1.4 trillion. Separately, Blackstone has warned that Anthropic may not receive an investment-grade credit rating after its planned initial public offering.

The developments land in a market where lending to AI companies has expanded rapidly. Goldman Sachs estimates that investors had lent AI-related companies about $500 billion in 2026 by early October, and the cost of insuring several AI-linked companies' debt against default rose sharply in October. Together, they amount to an early test of whether credit markets will keep underwriting the sector's expansion at its current scale.

Two Setbacks Raise Questions About AI Funding

Coin Bureau raised the question in a post on X on October 31, pointing to the OpenAI reduction and the Blackstone warning:

Is the AI funding boom starting to crack? The FT reports two new setbacks. OpenAI has SLASHED its funding target by $10 BILLION. It now seeks $20 BILLION at a $1.25 TRILLION valuation, down from $1.4 TRILLION. Blackstone, an anchor investor in Anthropic's $63 BILLION chip… pic.twitter.com/9We0OGKTDI — Coin Bureau (@coinbureau) October 31, 2026

(Post: Coin Bureau on X)

The second setback involves Anthropic. Blackstone is an anchor investor in the company's $63 billion chip financing deal, but it cautioned that Anthropic may not get an investment-grade rating following its IPO. The label matters because many institutional funds are restricted to holding investment-grade debt, so a rating below that threshold typically narrows a company's buyer base and raises its borrowing costs.

Credit Markets Show Early Stress

Credit default swap spreads, a measure of what it costs lenders to insure debt against default, illustrate how the AI funding boom is being priced. A basis point is one-hundredth of a percentage point. CoreWeave's five-year spread topped 1,100 basis points in mid-October — more than 11 percentage points of annual insurance on the covered debt. Oracle's spread reached roughly 300 basis points in October, while Meta's hit a record 114.79 basis points on October 22, after its earnings.

CoreWeave's bonds have also come under pressure. They fell below 80% of face value, and the yield approached 15% on Thursday, according to SiliconANGLE and Naked Capitalism. The company already carries roughly $52 billion in debt and lease obligations.

Oracle's Project Jupiter adds another example. The 1,400-acre campus sits in Doña Ana County, New Mexico. Permit denials, environmental litigation and construction delays pushed the timeline back by at least seven months, and about $18 billion in related loans trade at 89 to 91 cents on the dollar. As a result, Oracle reportedly invoked force majeure — a contract provision that suspends obligations during events beyond a party's control — and its credit rating slipped to BBB-, one notch above junk.

Borrowing Pace and Spending Forecasts

The FT reported that America's roughly $10 trillion AI infrastructure buildout is starting to fray. Hyperscaler debt issuance — the term covers the largest cloud operators — reached $121 billion in 2025, four times the historical average, and Bank of America expects another $100 billion in 2026. The pace resembles the dot-com era telecom buildout, when carriers borrowed heavily to lay fiber networks and many were left insolvent early in the following decade.

PIMCO estimates that hyperscaler capital expenditure absorbs 93% to 94% of operating cash flow. Private credit — loans arranged outside bank and public bond markets — is filling part of the gap. Apollo and Blackstone finalized a $35 billion package in June, called Project Big Sky, which funds Anthropic's lease of Google-designed tensor processing units. Blackstone has since pitched a follow-up of roughly $36 billion.

Spending forecasts keep rising despite these signals. Citigroup now expects tech giants to spend more than $2.8 trillion on AI infrastructure through 2029. Morgan Stanley, meanwhile, estimates a $1.5 trillion funding shortfall against $2.9 trillion in data center capital expenditure between 2025 and 2028 — a gap that private credit lenders have been moving to address.

Wall Street has also turned more cautious on parts of the hardware supply chain. Barclays cut Dell, Hewlett Packard Enterprise and NetApp to hold on October 15. Those stocks, together with Oracle and CoreWeave, lost about $80 billion in one session. Barclays found that businesses put only about 5% of IT budgets toward AI, and a quarter of that goes to hardware. Near-term markers for the sector include whether OpenAI's revised round closes at its new target, how Anthropic's IPO and rating process unfold, and the next readings on AI-linked bond and loan prices.