Meta's AA- Rating Survives, but S&P's Negative Outlook Asks Who Pays for the AI Buildout

Meta Platforms still holds an AA- issuer credit rating from S&P Global Ratings, one of the strongest grades in corporate America. On Friday, 30 October 2026, S&P nonetheless revised its outlook on that rating to negative from stable. Nothing about the advertising business prompted the change. The agency cited significant uncertainty around Meta's capital spending and free operating cash flow, and rising leverage. The company has forecast annual capital expenditure of $135bn-$145bn after reporting a free cash flow deficit of $2.12bn in the third quarter of 2026, and is planning to tap public markets with a $97.5bn stock sale to fund the AI buildout. S&P expects capital expenditure above $225bn in 2027 and $252bn in 2028, with free operating cash flow deficits of $22bn in 2026, $71bn in 2027 and $39bn in 2028 — a year in which it had previously expected Meta to return to positive cash flow. It also warned that leases signed for giant AI data centres could push Meta's debt measure above the agency's downgrade threshold as early as 2027.

That is the shift worth registering. For two decades, the credit question about a company like Meta was whether its core business could keep generating cash. The answer was never seriously in doubt. Now the question is how the buildout is paid for, and whether returns arrive before the bills do. A negative outlook is an early step toward a possible downgrade within roughly two years. It does not mean a downgrade is certain, and it does not mean the rating has been cut. But it does mean that even a AA- balance sheet is now being judged on the financing structure of artificial intelligence rather than on the quality of the franchise behind it.

The mechanics matter more than the headline. Data-centre leases are the quiet instrument here. Long-dated commitments for computing capacity can be treated by rating agencies in a way that resembles debt, and S&P has said explicitly that leases signed for giant AI data centres could push Meta's debt measure above its downgrade threshold as early as 2027. That is a financing constraint dressed as an operating expense. Add a planned $97.5bn equity sale — not a completed transaction, but a stated intention — and the picture becomes clearer still. Meta is not simply spending its own cash; it is assembling a capital structure to sustain the spending. S&P's judgement that Meta's ability to monetise its AI investments is more limited than that of its peers sharpens the point. If the revenue arrives later or thinner than peers', the financing has to carry more of the weight for longer.

This is why the AI race should now be read as a credit story. The competition is no longer only about model quality or talent. It is about who can fund the buildout without straining the balance sheet, and who can demonstrate a path from capital expenditure to cash return. When a company with Meta's rating and advertising engine draws a negative outlook on cash-flow uncertainty, the signal is not that Meta is weak. It is that the market's assessment of AI ambition increasingly runs through the language of leverage, lease obligations and free cash flow — the vocabulary of credit analysts, not product launches.

The strongest case against this reading deserves a fair hearing. Meta's AA- rating was affirmed, not lowered. Its advertising business remains strong, and the outlook change may be routine caution rather than a verdict that the AI financing model is broadly unsustainable. Rating agencies frequently move outlooks to negative without following through, and a two-year horizon gives Meta ample room to demonstrate that its spending converts into revenue. On this view, S&P is flagging uncertainty, not pronouncing on viability, and the core business still underwrites the whole enterprise.

That is a reasonable position, and it should not be dismissed. But it does not answer the specific concern. The affirmation of the rating and the negative outlook are not in tension; they are two halves of the same judgement. S&P is saying the current grade still holds while the trajectory has become uncertain. The uncertainty is not about whether Meta can sell advertising. It is about whether the cash-flow deficits S&P projects — $22bn in 2026, $71bn in 2027, $39bn in 2028 — close before the leverage threshold is breached. Routine caution would not normally come with a named mechanism and a named year. The warning that data-centre leases could push the debt measure above the downgrade threshold as early as 2027 is specific, and specificity is the opposite of boilerplate.

What remains genuinely open is how close Meta is to that threshold, and what precise leverage level would trigger an actual downgrade. Officials have not said. Nor is it known whether the reported $97.5bn equity raise will proceed, or how much dilution it would mean for existing shareholders. Whether Meta's AI investments can generate revenue fast enough to close the projected deficits by 2028 is the central question, and it is not yet answered. How Meta's capital expenditure plans and monetisation compare with peers S&P considers better positioned is likewise unresolved.

Those gaps are the story. The AI buildout has matured to the point where it is assessed not by what it promises but by how it is financed. Meta's advertising business remains formidable. Its rating remains strong. But the terms of judgement have changed, and S&P's negative outlook is the clearest sign yet that in this phase of the AI race, the balance sheet is the battlefield.