Goldman Sachs Projects $7.6 Trillion AI Capital Spending Through 2031, Shift to Physical Economy Accelerates

According to Goldman Sachs' latest report released on July 7, artificial intelligence investment is shifting from computing infrastructure toward broader physical industries including manufacturing, energy, logistics, defense, life sciences, and robotics. The bank projects global AI capital expenditure on computing, data centers, and power will reach approximately $7.6 trillion from 2026 through 2031, with annual spending rising from $765 billion in 2026 to $1.64 trillion by 2031. Hyperscale cloud providers' AI investments alone could exceed $6 trillion by 2030.

Global data center capacity has expanded from 30 gigawatts in 2019 to 57 gigawatts in 2024, with an additional 65 gigawatts expected before 2030. However, power is emerging as a critical bottleneck; data center grid interconnection queues now extend 8 to 12 years in key markets. Goldman Sachs cited Google's June deal with SpaceX—paying approximately $920 million monthly for roughly 110,000 Nvidia GPUs through mid-2029, totaling about $30 billion—as evidence that even major technology firms struggle to meet AI compute demand through internal capacity alone.