$1.3T Hyperscaler Capex by 2027 Leaves Only Microsoft FCF-Positive

For investors and credit analysts, the S&P Global forecast is a stark capital-allocation warning: a 50% jump in hyperscaler capex to $1.3 trillion in 2027 will likely pressure free cash flow, debt metrics, and credit spreads. With only Microsoft expected to stay FCF-positive, the market may need to reprice how much AI infrastructure growth is worth.

The central data point in the latest S&P Global report is bracing for investors: six hyperscalers driving the AI revolution—Microsoft, Alphabet, Amazon, Meta Platforms, Oracle, and SpaceX—are projected to spend a combined $1.3 trillion on capital expenditures in 2027. That figure follows $470 billion in actual 2025 capex and a forecast of $870 billion for 2026, implying another roughly 50% year-over-year increase in 2027. Just as striking, only Microsoft is expected to generate positive free cash flow next year, meaning five of the six companies are forecast to be cash-flow negative while they fund data centers, accelerators, memory, and server infrastructure for AI workloads.

“

That figure follows $470 billion in actual 2025 capex and a forecast of $870 billion for 2026, implying another roughly 50% year-over-year increase in 2027.

The spending trajectory reflects an insatiable demand for AI compute, from consumer LLM usage to enterprise AI integration. S&P Global, which assigns credit grades on debt issued by most major companies, is effectively flagging a credit and capital-allocation problem: hyperscaler capex is accelerating even as the incremental uplift to revenue and cash flow may lag by years. The 2025-to-2026 increase was much larger in percentage terms—$470 billion to $870 billion is roughly 85%—but the 2027 jump from an already elevated base is arguably more significant for balance sheets. A 50% increase on $870 billion adds about $430 billion in annual spending, equivalent to the entire 2025 base of the group.

The market impact extends well beyond the hyperscalers' own shares. Investors and credit analysts are likely to scrutinize free cash flow coverage, interest expense, and debt-to-EBITDA ratios as the spending cycle matures. If S&P's forecast is correct, only Microsoft's diversified software and cloud mix, strong margins, and lower apparent capex intensity will keep its free cash flow in positive territory. Alphabet, Amazon, Meta, Oracle, and SpaceX may need to fund the buildout with additional debt, suppress buybacks, or slow other investments. Even well-capitalized tech giants could see credit spreads widen if S&P signals negative outlooks or downgrade pressure. For equity markets, the trade-off is clear: the AI capex cycle is acting as a tax on free cash flow, and valuations tied to discounted future cash flow may need to adjust.

Company-level dynamics are already shifting. Most CEOs of these companies said on recent earnings calls they expect strong capex growth in 2027, suggesting management teams are willing to prioritize long-term AI capacity over near-term profitability. That behavior is rational in a land-grab context—hyperscalers that underinvest risk losing AI customers and developer mindshare to rivals. Yet the article notes that the incredible spending has begun to deteriorate the balance sheets of the group. Higher debt loads and reduced liquidity buffers could make these companies more vulnerable to an economic downturn or a period of AI demand normalization. S&P's grades will be a key early-warning signal; if the rating agency moves from stable to negative outlooks, borrowing costs could rise even before cash flow turns negative.