Data current as of August 20, 2026. Numbers on this page snapshot the JobShift dataset on the date above; the AI hiring view and AI layoffs view refresh continuously as new job postings and WARN filings land.
The 13-company overlap
Year to date through August 20, 2026, the JobShift dataset records 145 US companies with at least one AI-attributed hiring event, and 29 US companies with at least one AI-attributed layoff. The intersection is thirteen firms:
Amazon, Microsoft, Snowflake, Oracle America, Cloudflare, Pinterest, ServiceNow, GitLab, Robinhood, Snap, Coinbase, Atlassian US, and Uber.
Between them, these thirteen firms account for 2,977 of the 21,921 US AI-attributed hiring events on file for 2026 (roughly 13 percent), and 13 of the 31 US AI-attributed layoff clusters for the same period (roughly 42 percent).
How overlap is counted
Each of the thirteen firms shows the same underlying pattern in the data: a substantial volume of AI-role job postings across the year alongside exactly one AI-attributed layoff cluster. A cluster is a set of layoff filings that share the same company, the same restructuring incident, and a bounded window of filing dates. The cluster count is more informative than raw WARN-event counts because a single restructuring at a large employer often produces dozens of per-site filings when state WARN portals require separate rows for each affected location.
AI-role hiring postings, layoff-cluster counts, estimated workers affected where disclosed, and the date of each cluster's first filing:
| Company | AI hiring postings (YTD) | AI layoff clusters | Workers affected | First cluster filing | |---|---:|---:|---:|---| | Amazon | 1,528 | 1 | ~12,100 | 2026-01-01 | | Microsoft | 563 | 1 | ~6,105 | 2026-07-01 | | Snowflake | 266 | 1 | not disclosed | 2026-03-19 | | Oracle America | 232 | 1 | ~1,750 | 2026-03-31 | | Cloudflare | 151 | 1 | ~1,324 | 2026-05-07 | | Pinterest | 81 | 1 | ~118 | 2026-01-27 | | ServiceNow | 46 | 1 | ~287 | 2026-07-28 | | GitLab | 42 | 1 | not disclosed | 2026-05-11 | | Robinhood | 32 | 1 | ~353 | 2026-06-16 | | Snap | 25 | 1 | ~415 | 2026-04-15 | | Coinbase | 7 | 1 | ~700 | 2026-05-05 | | Atlassian US | 3 | 1 | ~315 | 2026-03-11 | | Uber | 1 | 1 | ~41 | 2026-07-22 |
Two firms with cluster-level filings but no disclosed worker counts (Snowflake, GitLab) appear in the count of clusters but not in any workers-affected aggregate. Filings that omit a worker count are common outside the largest states with strict WARN reporting requirements and are a known limitation of the WARN dataset.
Aggregate estimated workers affected across the disclosed clusters: approximately 23,500 across 2026 YTD. Amazon and Microsoft alone account for roughly 77 percent of that total.
Two patterns in the thirteen
The thirteen firms split into two hiring-volume tiers.
Ten firms — Amazon, Microsoft, Snowflake, Oracle, Cloudflare, Pinterest, ServiceNow, GitLab, Robinhood, and Snap — each posted at least twenty AI-role positions in 2026. The comparison between hiring intent (postings, a leading signal that may or may not convert to headcount) and workers separated (a hard WARN count) is not a direct like-for-like: 1,528 postings at Amazon do not imply 1,528 hires, and no inference about net headcount change follows from the two numbers alone. What the data supports is that AI-role hiring activity at these ten firms is substantial and ongoing across 2026 while an AI-attributed restructuring cluster ran in the same period.
The remaining three firms — Coinbase, Atlassian US, and Uber — posted between one and seven AI-role positions in 2026. The classification depends on very small hiring samples, and small-sample noise means the overlap for these three is suggestive, not confirmatory. A single missed AI-role posting for any of them would shift the picture.
The pattern this data does and does not confirm
The pattern in the data is a coincidence of AI hiring and AI-attributed restructuring inside the same firms during 2026. Several inferences that might look natural from that framing are not supported by the dataset.
Same-role churn is not established. The dataset does not confirm that the roles being hired are the same roles being cut. An AI-role hiring push could concentrate in one function (research, applied ML, infrastructure) while a layoff cluster affects a different function (support, sales, non-technical operations) inside the same firm. Media coverage of the individual restructuring incidents sometimes clarifies which function was cut; the JobShift classifier does not attempt that finer-grained matching.
Causation is not established. A layoff coinciding with AI adoption at the same company is not proof that AI caused the layoff. JobShift's classifier requires causal language in corroborating press coverage before applying the AI label to a layoff, which reduces the false-positive rate for the label itself, but a valid AI-attributed label describes the causal claim in the coverage, not a proven mechanism inside the firm. The classifier deep-dive documents the full attribution logic.
Firm-level volatility ranking is not established. Firms not on the thirteen-company list are not therefore stable AI employers. A firm hiring AI talent in 2026 without a corresponding layoff on file may lay off next quarter; a firm with a 2026 layoff but no AI hiring on record may add AI roles later in the year. The overlap is a snapshot of the current 2026 dataset.
What the overlap does confirm
Roughly 45 percent of the firms with an AI-attributed layoff in 2026 (13 of 29) also appear on the AI hiring leaderboard for the same year. The alternative narrative — that AI layoffs and AI hiring happen at different companies in different segments of the economy — is not consistent with the JobShift dataset for 2026 US events.
The pattern is dominated by a small number of very large employers. Amazon and Microsoft alone contribute 70 percent of the AI hiring postings across the thirteen-company set and 77 percent of the workers affected. The public discourse framing "AI is causing tech layoffs" and the framing "AI is creating tech jobs" both point at the same firms as evidence. Both framings are describing the same subset of the 2026 US labor market.
Live per-company data
The company-level snapshot on this page reflects the JobShift dataset on August 20, 2026. For real-time hiring and layoff activity as new postings and WARN filings land:
- The AI hiring view shows current AI-role postings with company detail and monthly totals.
- The AI layoffs view shows AI-attributed layoffs with monthly totals and per-company detail.
- The methodology page documents the AI-attribution logic applied to both hiring and layoff events.