The public narrative surrounding Anthropic’s October 26 S-1 filing has shifted from valuation debates to a more pressing concern: credit risk. The Financial Times reported that Blackstone, the anchor investor in a $63 billion financing deal for Anthropic’s chips, has revised its expectations. While the firm previously predicted an investment-grade rating before 2027, it now anticipates a high junk rating, specifically in the double-B range. This forecast, though preliminary, signals how the institutions closest to the deal are pricing the risk of Anthropic’s massive infrastructure commitments.

The Shift from Investment Grade to Speculative

Blackstone’s revision follows preliminary guidance from major rating agencies like Moody’s, S&P, and Fitch. The change is significant because it reflects a structural objection rather than a critique of product quality. Anthropic plans to spend over $800 billion on computing infrastructure through 2045. Rating agencies are reluctant to grant investment-grade status to a company with such enormous committed future spending and no through-cycle history. They are questioning whether a business with this level of leverage has earned the right to be treated as a low-risk borrower, regardless of its rapid revenue growth.

Anatomy of the $63 Billion Deal

The rating question is inextricably linked to the financing structure for Broadcom AI accelerators. Executed through a special purpose vehicle, the deal is layered. The senior tranche, worth about $43.5 billion, is backed by Broadcom’s residual value support, effectively leveraging the chipmaker’s investment-grade profile. The junior tranche, roughly $18 billion, relies on Anthropic’s lease payments. Blackstone has committed $9 billion to this junior layer, targeting a yield above 10%. This structure highlights that while senior lenders are betting on Broadcom, junior lenders are taking direct exposure to Anthropic’s credit, priced accordingly at the top tier of speculative grade.

Why the $80.7 Billion Loss Is Accounting Noise

The headline loss of $80.7 billion in Anthropic’s S-1 is largely a non-cash accounting event. Approximately $72.4 billion stems from fair value changes on convertible notes and derivative liabilities as the company’s private valuation climbed. Under GAAP, as the unicorn’s valuation rises, the conversion features on earlier notes become more valuable, recorded as a fair value loss. However, the operating business is improving. Anthropic reported $16.29 billion in revenue for the first half of 2026, up from $1.16 billion a year earlier, and generated positive operating cash flow of roughly $3.9 billion. The rating agencies focus on cash generation and debt service, not these mark-to-market adjustments.

The Bull Case and Structural Objections

Bankers argue that Anthropic’s 1,300% year-over-year revenue growth and 900% net revenue retention justify an investment-grade rating. Customers spending over $100,000 annually grew from 950 to over 11,900 in a year. Yet, structural objections remain. AI accelerators depreciate quickly, and their resale value is uncertain. The concentration of compute contracts with a few hyperscale providers creates counterparty risk. Furthermore, special purpose vehicles obscure aggregate leverage. Moody’s warned that the aggregate volume of machine-readable obligations from multiple AI labs financing datacenter capacity simultaneously is a systemic concern that agencies cannot ignore.

Key Takeaways

  • Blackstone now expects Anthropic to receive a double-B rating, the top tier of speculative grade, reversing prior predictions of investment-grade status.
  • The $63 billion financing deal is structured with a senior tranche backed by Broadcom’s credit and a junior tranche exposed to Anthropic’s lease payments.
  • Anthropic’s $80.7 billion net loss is primarily a non-cash accounting adjustment, while the company generated $3.9 billion in positive operating cash flow in H1 2026.
  • Rating agencies cite the lack of through-cycle history and the massive $800 billion infrastructure commitment through 2045 as key barriers to investment-grade status.

The Bottom Line

This isn't a verdict on Anthropic's technology; it's a price tag on its ambition. The shift to a junk rating forecast signals that the bond market is demanding a premium for the capital intensity of the AI buildout, a cost that will inevitably trickle down to API prices and infrastructure availability.