Anthropic's expected 'junk' rating shows the AI race is now about the cost of capital as much as model quality

Anthropic will probably receive a high "junk" rating from credit agencies after its initial public offering, according to a Financial Times report cited by TechCrunch. The forecast comes from Blackstone, the anchor investor in a recent $63 billion deal to finance Anthropic's chips. Blackstone reportedly told investors in that deal it now expects a rating in the double-B (BB) range.

First, what the label means. Credit agencies split borrowers into two groups. Investment grade covers BBB- and above and signals low risk to lenders. Speculative grade, known as "junk," covers everything below that. BB is the top tier of junk, one step below investment grade. It is not a distress rating. It does mean lenders will ask for higher interest rates and stricter terms.

The change matters as much as the rating itself. According to the FT, Blackstone previously said the Claude maker would get an investment-grade rating before 2027. It now expects the opposite. Anthropic and Blackstone both declined to comment to TechCrunch. This is still a prediction from one investor, not a decision by a rating agency. The timing is notable, though. Anthropic formally filed to go public earlier this week, and TechCrunch notes that its compute deals over the past 12 months have worried investors.

The practical impact is simple. A junk rating would make it harder and more expensive for Anthropic to raise the money behind its heavy spending. Frontier AI labs are no longer only software companies. They are becoming infrastructure buyers, committing to chips and data centres years ahead of the revenue those assets are meant to generate. Credit analysts don't score benchmarks. They look at cash flow, debt load and how predictable future income is.

**Our reading:** this is a real tension, not a verdict on Anthropic's technology. On our own Zendoric Quality index, Anthropic's Fable 5 leads the frontier at 90, ahead of OpenAI's GPT-5.6 Sol at 79. Credit markets, however, are pricing something else: the risk of financing a huge buildout before the business model has fully proven it can pay for it. Both things can be true at once. Technical leadership doesn't automatically mean financial strength, and the IPO will put that gap in public view. Our thesis is that the AI race has a new axis. Beyond who builds the smartest model, it now matters who can fund compute most cheaply. A lab that borrows at junk rates pays more for every GPU than a rival backed by a hyperscaler's balance sheet. That could quietly tilt the field toward the companies with the deepest pockets, extending the concentration-of-power concern we have flagged before.

We should avoid two lazy conclusions. A BB forecast is not a sign that the AI boom is collapsing. Many capital-intensive companies have grown into investment grade over time. Nor should it be brushed aside: higher borrowing costs during a spending spree is exactly the kind of short-term friction that comes with every major infrastructure transition. In the long run, the compute being financed today is what could power the drug discovery, scientific acceleration and productivity gains we expect from AI. In the short run, the real test is whether Anthropic's revenue grows fast enough to convince the agencies, rather than Blackstone's forecast, that it deserves a better rating. Watch the IPO filing's disclosures on debt and compute commitments. They will say more than any single forecast.