S&P lowered Meta’s credit outlook from stable to negative, warning that its heavy AI infrastructure spending and weakening free cash flow could pressure its finances, while leaving its AA- rating unchanged.

Meta plans to spend $135 billion to $145 billion on capital expenditure in 2026, and S&P projects its AI-related outlays will climb to more than $225 billion in 2027 and $252 billion in 2028.

S&P said Meta appears less able than competitors to monetize AI investment and expressed concern that the company has not disclosed spending plans beyond 2027; S&P executive Naveen Sarma warned that sustained or rising investment would weaken cash flow and credit quality.

S&P projects Meta’s adjusted leverage could rise from about 0.1x in 2025 to above 1.0x in 2027 and 1.3x in 2028, with $355.7 billion in lease liabilities and $80.92 billion in long-term debt reported at the end of September.

Meta’s bond-funding costs already exceed 7.2% annually and could increase if its rating is downgraded, though S&P stressed that the negative outlook is a warning about the rating’s direction, not an immediate downgrade.

Meta spokesperson Kamran Mumtaz disputed S&P’s view of its AI monetization prospects, saying the company sees substantial product and business opportunities that support a solid credit profile.

The pressure extends across Big Tech: FactSet data showed annual debt growth at Alphabet, Amazon, Meta, Microsoft and Oracle rose from 9% of capital expenditure in fiscal 2024 to 32% over the latest 12 months, while Meta’s planned $97.5 billion equity offering was also viewed against this backdrop.

The five companies’ combined shareholder returns fell to 9.3% of operating cash flow in the first three quarters of the year, from 27.4% a year earlier, as their combined free cash flow registered a $14 billion deficit.

In the third quarter, Alphabet, Amazon and Meta posted free cash flow deficits of about $9.1 billion, $33.5 billion and $2.1 billion, respectively; Microsoft remained in surplus at $18.6 billion while spending $45.2 billion on capital expenditure, including finance leases.

Investors are increasingly asking companies to prove that AI spending can generate profits, particularly as investment grows faster than revenue and buybacks contribute less to earnings per share.

Analysts describe the AI investment cycle as entering a “prove it” phase: companies must show that infrastructure spending produces cash flow, and prolonged pressure to recoup costs could eventually affect investment plans and efficiency efforts.

Financing concerns are also building around Anthropic: Blackstone reportedly expects it could remain in speculative-grade territory after a planned IPO, which may make corporate bond financing more expensive, though major credit agencies have not issued an official rating.