OpenAI and Anthropic are lobbying S&P for a BBB- rating: their near-trillion-dollar valuation is hitting the credit threshold of continuous losses.

Cutting-edge AI labs like OpenAI and Anthropic are seeking investment-grade credit ratings equivalent to Oracle's, a demand that tests whether rating agencies can establish a consistent set of evaluation standards for such drastically different business models.

On September 9th, the Financial Times, citing sources familiar with the matter, reported that banking advisors for these soon-to-be-listed AI labs have begun lobbying major credit rating agencies such as S&P and Moody's to secure investment-grade ratings for their clients. Specifically, this refers to Oracle's current S&P BBB- rating.

Investment-grade ratings are of great significance to AI labs: once obtained, these companies will be able to sell bonds to insurance companies, banks, and other asset management institutions that are constrained by conservative investment strategies, significantly broadening their financing channels.

However, the financial profiles of these two types of companies are fundamentally different. AI Labs has a private market valuation of nearly a trillion dollars, but it is still operating at a loss; Oracle, on the other hand, is a mature company with stable profits, although it also carries a heavy debt burden from data centers.

This comparison is pushing the pricing logic of the bond market into a new realm where there is no precedent to follow.

Two completely different financial profiles

The gap in credit quality between AI Labs and Oracle is readily apparent from their financial data.

The AI lab has already achieved an annualized revenue of over $40 billion, and its private equity valuation is close to $1 trillion. This substantial equity buffer can provide a safety margin for rating agencies to some extent.

However, the fundamental constraint of their business model lies in the fact that computing resources, R&D investment, and talent salaries constitute huge and continuous costs, which has prevented these companies from achieving profitability to this day.

Oracle, on the other hand, presents a completely different picture. In the fiscal year ending May 2026, Oracle achieved operating profit of over $20 billion, with an operating profit margin exceeding 30%. Its core businesses are enterprise application software and database software. With a market capitalization of nearly $500 billion, it is a well-established and mature company.

However, Oracle itself is not without risks: its massive bets on data center construction have resulted in a debt of $168 billion and negative free cash flow; at the same time, its off-balance-sheet data center lease liabilities amount to $260 billion— it is these factors that prompted S&P to downgrade it to BBB- this summer.

In other words, the investment-grade rating that the AI Labs is pursuing is precisely the lower limit that Oracle has been downgraded to due to financial pressure.

Ecosystem Bonds: OpenAI is Oracle's Largest "Creditor"

The fates of the two types of enterprises are not independent of each other, but are deeply intertwined with the same industrial chain.

According to S&P estimates, about half of Oracle’s current $638 billion “remaining performance obligations”—that is, customer prepayments—come from OpenAI.

This means that the continued high demand for AI models is a core prerequisite for maintaining the operation of the entire ecosystem: only with strong demand can AI labs generate enough revenue to pay the high fees to infrastructure providers such as Oracle; and companies like Oracle continue to increase their investment in data centers based on this expectation.

This interdependent relationship is both the fundamental logic supporting the valuations of both parties and a concentrated source of potential risks . If the growth rate of AI demand falls short of expectations, the ripple effect will simultaneously impact the creditworthiness of both laboratories and infrastructure providers.

Rating agencies face challenges in standardization

For rating agencies, the core challenge of this game is not just how to evaluate AI labs, but also how to establish a consistent evaluation framework between disruptive and established companies.

Stock investors can use more flexible methods when valuing stocks, assigning a high premium to future growth potential; however, bond investors must be more prudent in their judgment, focusing on cash flow repayment capacity and default risk.

Rating agencies face a double pressure on this issue: if they are too lenient with AI labs, they may underestimate the substantial risks faced by bondholders; if they are too strict, they may be questioned for failing to reflect the creditworthiness of new economies in a timely manner.

It is worth noting that after SpaceX's bonds were listed this summer, their trading spreads have been significantly wider than the level implied by the rating— this market signal indicates that even with an investment-grade rating, the bond market still has its own judgment on the true risk pricing of innovative technology companies.

Whether AI labs can gain equal market recognition if they eventually open the door to investment-grade products remains an open question.

Risk warning and disclaimerInvesting involves risk; please exercise caution. This article does not constitute personal investment advice and does not take into account the specific investment objectives, financial situation, or needs of individual users. Users should consider whether any opinions, views, or conclusions in this article are suitable for their specific circumstances. Any investment decisions made based on this information are at your own risk.