Goldman Sachs GS SUSTAIN: The AI Innovation Cycle — Investment Lessons from the Shale Innovation Cycle

Goldman Sachs draws on the 2003–2020 US shale innovation cycle to assess the current AI innovation cycle and its investment implications. The firm argues that AI currently remains in the evaluation phase—characterized by optimism and ambition—which historically favors infrastructure investment and elevated equity valuation multiples.

The transition from evaluation to execution during the shale era was driven by three catalysts: deteriorating corporate returns, constrained financial flexibility, and oversupply. By comparing hyperscalers' current financial flexibility and projected returns against these historical benchmarks, Goldman Sachs concludes the AI cycle has not yet reached a turning point. However, financial flexibility is emerging as a key differentiator for investors evaluating companies.

During the evaluation phase, efficiency gains typically accompany flat or rising corporate budgets. In contrast, the execution phase pressures budgets downward, shifting focus toward returns, balance sheets, and market share. The shale cycle demonstrated that corporate valuations (EV/GCI) expanded when returns were high but collapsed once returns declined, even as growth inventories expanded. Currently, hyperscaler corporate returns remain near the upper end of their historical 24%–31% range.

Goldman Sachs maintains a bullish outlook on AI and data center power supply chain stocks. Global data center power consumption is projected to surge roughly 175% by the end of this decade, driving unprecedented electricity demand growth across the US and Europe. The firm highlights a reliability theme for 2026 and identifies companies across the Americas, EMEA, and Asia-Pacific positioned to benefit from sustained power infrastructure investment, decarbonization, and grid modernization.