Big Tech AI Capex Won't Peak Until 2028, Goldman Sachs Warns

The massive capital expenditure announcements from technology giants building AI infrastructure represent just the beginning of a multi-year spending wave that won't balance until 2028, according to Goldman Sachs.

Eric Sheridan, co-head of TMT research at Goldman Sachs, told Yahoo Finance that the supply-demand imbalance in AI infrastructure will persist for years, driving continued increases in capital spending, revenue growth, and supply chain constraints. Memory pricing is rising, chip costs are higher than six and 12 months ago, and companies are racing to acquire land and build data center shells years in advance of when components will be available.

Record spending levels across the sector

The second quarter of 2026 demonstrated the scale of investment underway. Alphabet reported capital expenditures of $44.9 billion for the quarter and raised its full-year guidance to $195 billion to $205 billion, up from $180 billion to $190 billion. Company executives indicated a "significant" increase is expected in 2027.

Tesla announced it will commit $25 billion in capital expenditures for 2026—roughly three times its historical spending levels—as CEO Elon Musk accelerates production of Optimus robots and robotaxis. The company expects another substantial increase in 2027.

SpaceX posted second-quarter capital expenditures of $18.4 billion, far exceeding analyst estimates of approximately $6 billion. Commentary from executives suggested third and fourth quarter spending could remain at similar levels, implying full-year capex of around $65 billion versus Wall Street's $50 billion estimate. JPMorgan has warned that SpaceX could allocate $200 billion to AI spending in each of the next two years.

Why it matters

The extended timeline for AI infrastructure buildout has direct implications for technology investors and the broader supply chain. With Goldman Sachs not expecting supply and demand to balance until the first half of 2028, companies face years of elevated input costs and capital allocation decisions that will reshape balance sheets. Several Big Tech stocks, including Tesla, declined during the recent earnings season as investors grappled with the magnitude of spending commitments. The question of whether current valuations adequately reflect this multi-year capex cycle remains unresolved, particularly as rising costs for memory, chips, and data center construction compress margins even as companies race to secure AI infrastructure capacity.

Supply chain under pressure

The constrained supply chain is driving up costs across the entire technology stack. When components are scarce, pricing power shifts to suppliers, resulting in higher input costs for the hyperscalers building AI infrastructure. Companies are responding by securing data center sites and building shells years before they expect to install equipment, adding to near-term capital requirements.

The details were first reported by Yahoo Finance in an interview with Goldman Sachs' Eric Sheridan.