GF Securities lowered its forecasts for capex growth at the five largest US cloud providers in 2027 and 2028: AI demand and tight computing capacity support near-term investment, but free-cash-flow and credit-rating pressures will limit subsequent expansion.
Bernstein expects the Texas data center moratorium to continue and has lowered its price targets for miners transitioning to AI infrastructure amid approval uncertainty and rising financing costs; the scarcity value of approved power coexists with project financing and delivery pressures.
Bernstein estimates that OpenAI’s infrastructure spending through 2030 will total 614 billion dollars if all contracts are delivered on time and none are canceled; whether revenue can continue to grow rapidly will affect its ability to pay and suppliers’ long-term demand.
Jefferies sees cloud providers’ backlogs and business growth continuing to rise, but expects capital spending to keep depressing free cash flow; borrowing costs for new cloud providers are also climbing.
Based on guidance from Alphabet, Microsoft and Meta, JPMorgan estimates that the three companies’ combined capital spending in the fourth quarter of 2026 will be approximately 174 billion dollars, up approximately 87 billion dollars year over year. Delays in third-party capacity delivery have affected Microsoft’s spending in the quarter, but all three companies have signaled continued investment.
J.P. Morgan’s tracking shows that Token spending on OpenRouter rose 40% month over month in October, but usage growth slowed; prices for different GPUs and memory products followed divergent paths over the same period.
Results from three cloud giants will test demand for AI computing capacity and returns on investment. Goldman Sachs expects the AI boom’s boost to economic growth to peak between the end of 2026 and early 2027, while energy prices and interest rates add downside pressure to growth.
Large financing deals and demand for computing capacity continue to support AI expansion, but companies differ markedly in cash flow, contract structures and the conditions for bringing capacity online. Meanwhile, model usage is rising as providers cut prices to win customers.
A compilation by Cailian Press shows that enterprise AI has entered finance, auditing and manufacturing, while tight AWS capacity supports a positive assessment of the computing-power supply chain. The extent to which domestic suppliers benefit depends on specific products and deliveries, and the market’s rebound on higher volume coincided with a decline in the number of companies hitting new highs.
