Where Does the U.S. Economy Rely on H-1B Labor? Worksite-Level Evidence from Labor Condition Applications, FY2021–FY2026, and Early Responses to the 2025–2026 Policy Changes

U.S. H-1B policy changed sharply in 2025–2026 through a payment on certain new petitions, wage-weighted cap selection, and a proposed rise in prevailing wages. I assemble 3.03 million certified Labor Condition Applications (LCAs) from FY2021 to June 2026, de-duplicate cases, allocate them across listed worksites, and normalize by local QCEW employment. Employers obtained about 525,000 LCAs a year, 3.5 per 1,000 jobs. The ten largest metro markets hold 55% of LCAs but 25% of jobs. San Jose leads, but technology hubs, IT-services hubs and small institutional labor markets have different job structures and policy exposures. Matching FY2025 requests to May 2025 OEWS employment in the same occupation sharpens this distinction: software developers account for 102.5 LCAs per 1,000 software jobs nationally, while general classroom teachers account for 1.8 per 1,000 teaching jobs; Alaska's teacher rate is about twelve times the national rate. Cap-season new-employment filings shift toward higher wage levels in FY2026, a pattern consistent with weighted selection but not an estimate of its causal effect or of changes in selection probabilities. A next-level wage scenario identifies 59% of eligible offers below substitute thresholds; this is not a lower bound on effects of the proposed rule. These indicators measure annual application flows relative to jobs, not the share of workers holding H-1B status.