Federal Reserve Chair Kevin Warsh is scheduled to host a three-hour public forum in San Francisco on Monday on how technological change could affect the economy and workers. The meeting comes as AI investment grows, while estimates of its economic contribution, productivity effects and employment consequences remain uncertain. Reuters reported that the forum will feature three speaker panels, with remarks from Warsh at either end.
Investment and electricity are different measures
The scale of AI-related spending depends on what is counted. Capital Economics estimated that hyperscaler spending would reach about $800 billion in 2026 and $1 trillion in 2027, and that AI investment added roughly 0.3 percentage points to real GDP growth in 2025, according to CBS News. Epoch AI estimated that U.S. investment in AI-related data-center construction, computing hardware and networking was about 0.8% of GDP in the first quarter of 2026; computing infrastructure overall was about 1.5%. Those figures cover different categories and are not directly comparable.
For California, the immediate planning question includes electricity demand. The California Energy Commission said data centers accounted for about 1,000 megawatts, or 2% of California ISO peak electricity demand, in early 2026. That is a measure of demand at that time, not a projection of future needs. The commission’s 2025 IEPR planning methodology uses a mid data-center scenario, applying moderate-confidence assumptions to signed agreements and active applications, rather than treating every proposed facility as certain to be built. At a January 28, 2026, Assembly hearing, the California ISO said large loads are incorporated into Energy Commission forecasts and their resource needs inform CPUC resource-planning requirements, according to the commission’s methodology memo and the California ISO hearing presentation.
Local opposition reflects concerns about how development affects communities, not a settled estimate of its statewide costs. Reuters reported that Oakland Mayor Barbara Lee joined calls to restrict local data-center construction, citing responsible AI use and possible harms including bias, discrimination and displacement. As a separate example of local planning policy, Loudoun County, Virginia, approved a March 18, 2025, zoning change requiring special-exception approval for data centers in specified districts where they had previously been allowed by right, according to the county. These local actions illustrate different planning responses; they do not establish a broader trend or predict the outcome of California’s planning process.
Productivity and employment effects remain unsettled
Research cited in the forum debate points to uneven gains so far. A survey of corporate executives in an Atlanta Fed working paper found productivity improvements that differed by sector and little evidence of aggregate employment declines in the near term. Larger firms anticipated AI-related workforce reductions, while smaller firms expected modest employment gains. A Minneapolis Fed article described current effects as moderate and the wider arrival of productivity gains as uneven. A Federal Reserve research note, while cautioning that its findings were lower confidence, attributed a meaningful contribution to quarterly GDP growth from proposed AI-related components between 2025 and the first quarter of 2026.
Public anxiety is not evidence that job losses have already occurred. A Reuters/Ipsos poll reported by USA Today found that 53% of Americans feared AI could cost them or someone in their household a job. Meanwhile, Reuters reported that Fed Governor Lisa Cook warned of a possible temporary rise in unemployment and potential increase in underlying joblessness. Vice Chair for Supervision Michael Barr urged preparation for serious short-term disruption, with outcomes depending partly on adoption speed and access to the technology; Richmond Fed President Tom Barkin warned that the scale of investment could pose risks if companies later cut back.
Warsh appointed venture capitalist Marc Andreessen to lead the task force focused on AI, which is charged with assessing the economic effects of general-purpose technologies to inform Fed policy judgments, according to CNBC. Reuters reported that task-force members come from the technology industry and that their recommendations would not determine policy on their own; decisions remain with Fed officials. Reuters also reported that Warsh’s earlier expectation that AI-driven productivity could ease inflation and create room for lower rates has become more qualified since he joined the Fed in May, and that he sold AI investments under Fed ethics rules after taking the post. The forum is set to examine the competing possibilities of productivity gains, employment disruption and price pressures from the infrastructure investment required to build AI systems.
