Goldman Sachs: AI capex hits $905B by 2027!

Goldman Sachs is forecasting AI-related capital expenditure to reach $905 billion by 2027, a projection that underscores just how aggressively the technology sector — and the institutions financing it — are committing to the AI buildout. The figure encompasses infrastructure spending across data centres, custom silicon, power systems, and the broader supply chain that feeds large-scale model training and inference.

For macro and markets investors, the number is a directional signal as much as a forecast: Goldman's willingness to put a near-trillion-dollar figure on the cycle suggests the bank sees the current wave of AI investment as structural, not a short-term capex bubble. That framing has direct implications for semiconductor names, hyperscaler earnings, and energy infrastructure plays that sit in the AI spending chain.