Everything About AI
NextFin News - S&P Global Ratings on Friday lowered the outlook on Meta Platforms' 'AA-' credit rating to negative, the clearest signal yet that Wall Street's tolerance for the artificial-intelligence spending boom is wearing thin. The move stops short of an actual downgrade, but it lands one week after Meta reported negative free cash flow for the first time since its 2012 IPO year and raised its capital expenditure forecast for the third time in twelve months. The question now is whether Meta's advertising engine can generate enough cash to justify a build-out with no clear payoff date.
The action makes Meta the first of the four largest hyperscalers to receive a negative credit outlook tied directly to AI infrastructure spending. Amazon, which carries an 'AA' rating from S&P, and Alphabet remain on stable outlooks. The distinction matters: S&P is not saying Meta's balance sheet is weak today — it is saying the path ahead has become too uncertain to reward with a stable bias.
The Situation: A Rating Under Notice
S&P said Meta's ability to monetize its AI investments is more limited than that of its peers, and that it is unclear when the expenditures will begin producing material revenue that exceeds rising costs. The agency now expects Meta to run free operating cash flow deficits of $22 billion in 2026 and $71 billion in 2027, with a further $39 billion shortfall in 2028. That is a sharp reversal from the agency's prior view that Meta would return to positive free cash flow by 2028.
Meta's ability to monetize its AI investments is more limited than that of its peers, and it is unclear when expenditures will start producing material revenue that exceeds rising costs.
The quarterly numbers behind the downgrade are unforgiving. In the third quarter, Meta generated $32.79 billion of operating cash flow and spent $34.91 billion on capital expenditures, leaving free cash flow at negative $2.12 billion. The company narrowed its full-year 2026 capex guidance to $135 billion to $145 billion — the third upward revision this year, from $115 billion to $135 billion in January. The midpoint of that range is nearly double the $72.22 billion Meta actually spent in 2025.
Meta's balance sheet still looks strong on the surface. Cash, cash equivalents and marketable securities stood at $82.45 billion as of September 30, roughly in line with long-term debt. But the trajectory is what worries creditors. S&P projects adjusted leverage to reach 1.1x by 2027 and 1.3x by 2028, crossing the 1.0x threshold that defines the 'AA-' category. The agency said it could lower the rating if leverage stays above 1.0x on a sustained basis and free operating cash flow remains negative beyond 2028.
The credit markets have already started to reprice the risk. Meta's five-year credit default swaps — the cost of insuring its debt against default — widened to a record 114.79 basis points on October 22, up 7.6 basis points in a single session, according to Markit data. For context, a company with a 'AA-' rating would normally trade in the low double digits; a move past 110 basis points is more typical of single-A issuers. Bond investors are not betting on default — Meta's cash pile makes that remote — but they are demanding more compensation for the uncertainty. The equity market reacted more severely: Meta shares fell 7.0% to $648.85 on October 22 and traded at $614.99 on October 30, down 1.53% on the day and roughly 23% below the 52-week high of $779.82.
Why This Is Different From a Normal Investment Cycle
The immediate trigger is mechanical: capital spending exceeded the cash the business generates. But the deeper issue is that Meta is funding a multi-hundred-billion-dollar infrastructure build-out with a revenue stream whose AI payoff remains unproven. Unlike a factory expansion where capacity and unit economics are known, AI data centers are a bet on future model capability, inference demand, and advertising yield per query — none of which have settled into a predictable pattern.
The spending path itself tells the story. Meta's 2026 capex floor has risen every quarter: $115 billion, then $125 billion, then $130 billion, now $135 billion. Each revision consumes a larger share of operating cash flow. In the second quarter, capex of $31.08 billion absorbed about 98% of the $31.86 billion in operating cash flow, leaving just $784 million of free cash flow. In the third quarter, spending exceeded operating cash entirely. Over the first nine months of 2025, Meta generated $29.5 billion of free cash flow; over the first nine months of 2026, that figure fell to $11.0 billion.
There is also a less visible part of the build-out: commitments that do not appear as debt on the balance sheet. As of September 30, Meta had roughly $825 billion in future lease and contract commitments, up from nearly $700 billion in July. About $356 billion of that is for leases not yet begun, and $121 billion falls due in 2027 alone. These are contractual obligations that do not self-correct if AI revenue disappoints. To meet some of these obligations, Meta transferred $10.80 billion in money market funds into restricted cash equivalents in the third quarter — cash that is locked as escrow until the related purchase obligations are met between 2028 and 2030.
Cyclical Dip or Structural Shift
This is a structural shift in Meta's capital model, not a cyclical dip. A cyclical capex wave has a defined endpoint — you build the capacity, then spending rolls off and cash flow recovers. Meta's commitments extend far beyond the annual guidance and are locked in for years. The company is moving from an asset-light advertising business to an asset-heavy infrastructure owner, and that change does not reverse on its own.
The structural argument is reinforced by the competitive dynamic. The four hyperscalers — Amazon, Alphabet, Meta and Microsoft — collectively plan to spend roughly $725 billion to $730 billion on capital expenditures in 2026, up 77% to 78% from about $410 billion in 2025. This is an arms race: if Meta slows spending, it risks losing model capability and cloud relevance; if it keeps pace, cash flow stays negative. The prisoner's dilemma means no single player can unilaterally revert to the old cash-generative model.
Moody's Ratings made the same point in July, warning that the AI spending surge "threatens credit quality" for the six companies it tracks — Microsoft, Amazon, Alphabet, Meta, Oracle and CoreWeave — and projecting capital expenditures of $785 billion in 2026 across those names, rising toward about $1 trillion in 2027. Moody's noted that the four largest retain among the strongest balance sheets in the world, making an imminent threat to their investment-grade ratings unlikely. But it also said the spending surge is forcing even the most cash-rich corporations to lean on debt, stock sales and off-balance-sheet financing. S&P's action on Meta is the first time that warning has translated into a negative outlook at the top of the group.
The Second-Order Problem: The Funding Model Is Changing
The first-order effect is negative free cash flow. The second-order effect is that Meta is quietly changing how it funds itself. In May, the company completed a $25 billion multi-tranche senior notes offering — one of the largest debt raises in the technology sector — which took long-term debt to about $84 billion by the end of June, up from $59 billion at the end of 2025. And the company is creating a new nonvoting Class C share class, with an amended charter effective January 19, 2027, under which holders of each outstanding Class A and Class B share will receive one Class C share. That structure gives Meta a way to raise equity capital without diluting Mark Zuckerberg's voting control.
That matters for credit because the 'AA-' rating has always rested on Meta's net cash position and its ability to fund growth internally. S&P noted in April, when it rated Meta's proposed notes, that adjusted leverage was 0.1x pro forma for the May debt issuance. The projection of 1.1x by 2027 represents an elevenfold increase in two years. Ratings are not just about the level of leverage; they are about the direction and the durability of the cash engine that services it.
The shift also has a cost. Meta's last bond sale, in October 2025, attracted $125 billion of orders — a sign that investors were eager to lend. But the record CDS spread shows that the same creditors now want more compensation for the same risk. Over time, that translates into higher borrowing costs on every new issuance, which compounds the pressure on margins already squeezed by depreciation on hundreds of billions of dollars of new data-center assets.
The Counter-Thesis: The Ad Business Is Still Firing
The strongest argument against a downgrade is that Meta's core business is not broken. Third-quarter revenue rose 26% to $64.60 billion, with advertising up 25.9%. Excluding a $9.89 billion litigation charge related to a settlement with state attorneys general, income from operations would have been $22.88 billion and diluted earnings per share $7.40 instead of the reported $3.83. The company guided fourth-quarter revenue to $73 billion to $76 billion, implying continued acceleration.
Bulls also note that Wall Street has largely looked through the spending: the consensus 12-month price target of $834.36 across 69 analysts implies roughly 34% upside from current levels. Evercore ISI, while cutting its target to $875 from $930 on October 30, kept an Outperform rating. The metaverse analogy offers further comfort: Reality Labs posted an operating loss of $19.2 billion over the trailing four quarters, yet the stock and the credit held.
But the metaverse comparison cuts the other way. Reality Labs was a contained bet funded by a cash-printing core. The AI build-out is not contained — it is the core, and it is consuming the core's cash. The metaverse never threatened Meta's credit rating. AI spending does. And unlike the metaverse, the AI race is one Meta cannot opt out of without conceding the next computing platform to Microsoft, Alphabet and Amazon.
What Would Prove the Negative-Outlook Call Wrong
The falsifying signal is specific: if Meta's free operating cash flow returns to a sustained positive level by 2028 while adjusted leverage stays below 1.0x, the negative outlook becomes moot and the 'AA-' rating is secure. Concretely, that requires 2026 capex to land at the low end of the $135 billion to $145 billion guidance, 2027 spending to come in materially below the implied trajectory, and AI-driven advertising yield to lift operating margins faster than S&P assumes. A single quarter of positive free cash flow would not be enough — the agency's concern is the multi-year path.
Outlook: Three Scenarios
In the short term, the negative outlook adds a new overhang to a stock already under pressure. Meta closed at $614.99 on October 30, down 1.53% on the day and roughly 23% below its 52-week high of $779.82. The stock trades at about 21 times forward earnings, a discount to the peak multiple it commanded when cash flow was the story.
The medium-term path depends on the fourth-quarter report, when Meta is due to lay out 2027 capital spending. If management signals that the capex floor is stabilizing and free cash flow turns positive in the fourth quarter — helped by the $73 billion to $76 billion revenue guide — the shares can stabilize. If 2027 spending guidance comes in above expectations, the credit story deteriorates further and the rating itself moves into play.
The long-term structural question is whether AI becomes a durable profit engine or a permanent cost center. If AI lifts advertising conversion and engagement enough to expand margins, today's spending will look like a bargain. If not, Meta faces a decade of elevated capex with returns that never fully materialize — and a credit rating that follows.
Base case: Meta holds the 'AA-' rating but carries a negative outlook through 2027, with leverage peaking near 1.1x and free cash flow remaining negative until 2028. Upside case: AI monetization accelerates, capex peaks in 2027, and the outlook is revised back to stable by 2028. Downside case: spending exceeds guidance, leverage breaches 1.0x on a sustained basis, and S&P cuts the rating to 'A+' in 2027 or 2028.
S&P has not downgraded Meta, but it has put the company on notice that the AI spending spree is now a credit issue, not just an investment story. For a company that built its reputation on printing cash, the burden of proof has shifted from investors to management.
Explore more exclusive insights at nextfin.ai.
