OpenAI cut its latest fundraising target to $20 billion from $30 billion as S&P Global shifted Meta’s credit outlook to negative, first reported by Fortune. Together, the developments put fundraising ambitions and Meta’s AI spending under scrutiny as investors weigh rising investment against cash generation and questions about when its AI-related revenue will become material.
OpenAI’s expected valuation also fell, to $1.25 trillion from $1.4 trillion. The reduction followed a selloff in technology stocks and came as quarterly results from hyperscalers heightened concern about heavy investment.
Anthropic’s upcoming IPO helped draw attention away from OpenAI’s round. Its Monday S-1 filing showed revenue had soared and cash flow turned positive in the first half of this year.
S&P kept Meta’s credit rating at AA- but changed its outlook from stable to negative, citing negative cash flow and uncertainty around capital expenditure. The agency said leverage could top 1.0x as soon as 2027, the downgrade threshold for the current rating.
That compares with about 0.1x in 2025, with the increase primarily reflecting Meta’s strategy of leasing data-center capacity rather than debt issuance. S&P expects Meta’s capital expenditure to reach $140 billion this year, top $225 billion in 2027 and hit $252 billion in 2028, then flatten out.
The agency projects free operating cash flow of -$22 billion this year, -$71 billion in 2027 and -$39 billion in 2028. S&P’s concern is not only the size of Meta’s spending, but when AI-related revenue will make material contributions above rising costs.
S&P said Meta’s ability to monetize AI investments is more limited than its peers and that it is unclear when new revenue streams will make material contributions that exceed rising costs over the next few years. That puts the question of returns, not just spending totals, at the center of the warning.
S&P believes Meta intends to seek returns through improved advertising tools and new revenue from agentic apps such as Muse. Its rivals, it said, monetize data centers by leasing compute capacity to frontier model labs and enterprise customers.
Meta disputed the assessment, saying its AI investments were already delivering stronger engagement, better recommendations and greater returns for businesses.
Market veteran Ed Yardeni said financial, physical and political constraints could be healthy for the AI boom’s durability. He expects hyperscalers to retain access to financing because of their strong balance sheets and to keep spending.