Goldman Sachs Maintains U.S. Data Center Growth Forecast Through 2027

Goldman Sachs maintains that U.S. data center expansion remains on track through 2027. This outlook persists despite rising political and community resistance to the infrastructure required to support the artificial intelligence boom.

Although developers face localized pushback, analysts at the firm suggest that current project data supports a steady trajectory for power capacity over the next two years. The firm’s updated outlook indicates that sector growth remains insulated from these social and regulatory headwinds.

"While US data center development faces increasing political and community opposition, project data and our updated capacity estimates suggest that the US data center growth outlook through 2027 remains largely unchanged," according to a note from Goldman Sachs.

Updated Capacity and Power Demand Forecasts

Goldman Sachs has adjusted its capacity projections for the coming years to reflect evolving construction timelines. The firm raised its year-end 2026 U.S. data center capacity forecast by 5 gigawatts (GW) to 64 GW. It reduced its 2027 year-end forecast by 5 GW, bringing that estimate to 90 GW.

The firm projects that U.S. data center power demand will continue to surge, driven by the energy requirements of AI model training and inference. Goldman expects power demand to grow by 38%, which equals 12 GW, in 2026. This will be followed by another 38% increase, or 17 GW, in 2027.

The Growing Gap in Public Sentiment

Despite the technical necessity of these facilities for firms like Meta, Alphabet, Amazon, Microsoft, and xAI, public support for the infrastructure is thin. Goldman Sachs noted that "public sentiment has also turned more negative" as construction activity intensifies in various regions.

Data from a June Reuters/Ipsos poll highlights the scope of this resistance. Only 14% of Americans indicated they would support the construction of a data center in their own community to facilitate AI projects. This disconnect between corporate infrastructure requirements and local community priorities remains a manageable variable for developers.

What Is at Stake

Data centers are the physical foundation of the current AI cycle. They require consistent, high-voltage electricity access that often competes with residential and other industrial needs. As these facilities move from planning to construction, the tension between regional energy grids and the expansion of hyperscale cloud providers will continue to test the limits of local zoning and community approval processes.