Payrolls
Change in nonfarm employment on the previous month
+29k
−61k vs consensus +90k
Sep 2026
Monitor
Hiring, unemployment and pay in the US and the largest economies.
Payrolls, unemployment, job openings and wage growth from the BLS, jobs by sector, job cuts linked to AI, AI exposure by occupation, and the US against Europe: unemployment, hiring, migration and AI use.
Latest reading: 23 October 2026
Labour market
Hiring, unemployment, job openings and pay, job cuts that companies link to AI, and how AI adoption relates to job growth.
Unemployment rate
Share of the labour force without a job
4.2%
+0.1 pp vs consensus 4.1%
Sep 2026
Job openings
Open positions on the last business day of the month
7.08M
−256k vs prior
Aug 2026
Wage growth
Average hourly earnings, private sector, change on a year earlier
3.0%
−0.2 pp vs consensus 3.2%
After inflation −0.7%
Sep 2026
Jobs added each month
Change in US nonfarm payrolls on the month before · October 2024 to September 2026
Jobs addedJobs lostConsensus for Oct 2026: +40k–65k
Seasonally adjusted. The two latest months are first estimates. October 2025 has no bar: the survey did not run that month.
Jobs added each month
Change in US nonfarm payrolls on the month before · October 2021 to September 2026
Jobs addedJobs lostConsensus for Oct 2026: +40k–65k
Seasonally adjusted. The two latest months are first estimates. October 2025 has no bar: the survey did not run that month.
Most jobs added, by sector
Year to date: December 2025 to September 2026
Change in jobs, thousands
All nonfarm jobs: +612k. Seasonally adjusted, December to the latest month. The two latest months are first estimates. 5 sectors with smaller changes are not shown.
Most jobs lost, by sector
Year to date: December 2025 to September 2026
Change in jobs, thousands
All nonfarm jobs: +612k. Seasonally adjusted, December to the latest month. The two latest months are first estimates. 5 sectors with smaller changes are not shown.
Job cuts companies linked to AI
166 announcements since January 2023: 338,537 jobs stated, about 16,621 more estimated from shares of staff
Use Tab to move through the list. Press Enter to show a company's details.
Microsoft
Tech · MSFT · USA
- Announced
- Oct 26, 2026
- Jobs cut
- Not stated
- Share of staff
- –
- Reason given
- Possibly linked to AI
A voluntary exit offer, not a layoff: open to staff at level 70 and below in Canada, India, the UK and the US (press estimates about 60,000 eligible); take-up not yet known. Microsoft's post does not mention AI. Possibly linked: the offer reshapes the workforce while Microsoft spends record sums on AI infrastructure; the link is our judgement, not Microsoft's words.
Source: blogs.microsoft.com ↗“It also gives participating organizations another way to adapt their workforce to changing needs.”
Cuts that management linked to AI in a statement, filing, earnings call or internal memo. Cuts only the press linked to AI are left out, except a few large ones marked "Possibly linked to AI" (our judgement). Numbers are announced plans. Click a company for details.
AI adoption against job growth, by industry
AI adoption in Sep 2026 against job growth vs the 2015–19 trend
IndustryAll businesses / all jobs
Job growth vs 2015–19 trend, pp a year
Use Tab to move through the industries.
Each dot is an industry. Further right, more firms pay for AI. Lower, jobs grow more slowly than before the pandemic. Interest rates also affect these seven industries.
Source: Ramp AI Index ↗Headcount after firms adopt AI, by sector
Headcount over 24 months vs similar firms that adopt AI later · study of June 2026
Heavy AI spendersLight AI spenders
Change in headcount, %
Use Tab to move through the sectors.
* Statistically significant. Heavy: top third by AI spend per employee. In information, AI-heavy firms grew while the industry shrank. These firms already grew faster before.
What markets expect
AI and work
How much US work and pay AI touches, which jobs it reaches, and how fast AI agents take on work.
Share of US work AI touches
From Claude use. 24.1M work in jobs where AI handles 25%+
11.1%
Est. Nov 2026: ~12.7% (11.1%–14.5%)
May 2026
Pay for work AI touchesEach occupation's workers times its average wage (BLS, May 2025) times its AI exposure, summed. The estimate for today uses the same growth as the share-of-work card.
Per year at May 2025 wages, 13% of all US pay
$1.3T
Est. Nov 2026: ~$1.5T ($1.3T–$1.7T)
May 2026
Agent hours per human hour
All Anthropic staff; research teams 19.1×. Anthropic makes the models
6.2×
+30.7% vs 4 weeks ago
Week to Aug 24
Workers AI agents could addFull-time-worker equivalents that AI chips shipped worldwide could run at full use, in Epoch's central case.
Full-time workers AI chips shipped by 2026 could run (Epoch)
84–168M
Through 2027: 211–423M
Oct 2026
Which jobs AI and robots already reach
Each square is a US occupation, shaded by the share of its work AI handled in practice in April and May 2026
0%91%None
Bar under each family: its workers, shaded by average exposureClick a family to see its occupations
328 of 756 occupations (85.1M of 140.5M workers) show AI use. Data Entry Keyers lead at 91%. Pay for work AI touches: $1.3T of $10T a year. Our rebuild of Anthropic's measure from Claude.ai use in April–May 2026: the share of a job's task time done with AI.
OpenAI puts 18% of US workers in jobs at higher risk of automation
Share of US workers (May 2025) in each OpenAI group: all workers, then each job family, highest automation risk first
Grow with AIWill reorganizeHigher automation riskLess immediate changeNot in OpenAI's table
Share of workers
OpenAI's own classification of 923 occupations (data download, June 2026). It is a judgement framework, not a forecast, and it does not come from Claude use. We weight each group by BLS workers (May 2025); the occupations OpenAI covers hold 99.9% of those workers. The number at each bar's end is the share at higher automation risk.
ChatGPT work share by plan
Share of each plan's messages likely related to work, to Sep 2026
FreeGoPlusPro
Consumer plans only: no Enterprise, Edu or Codex, so business use is understated. Uses OpenAI's label for messages likely related to work. Shares of messages, not counts. OpenAI restates past months with each release.
US work hours assisted by AI
Share of work hours, employed US adults aged 18–64, quarterly, to Q3 2026
Self-reported in the Real-Time Population Survey (Bick, Blandin and Deming). Non-users count as zero; each figure is the midpoint of a reported range.
Source: St. Louis Fed ↗Agent hours overtake human hours at Anthropic
Hours AI agents worked per human work hour, weekly, log scale
All employeesResearch organizationsParity
Agent hours per human hour
Use the left and right arrow keys to inspect weeks. Press Escape to close the readout.
One company, the developer of the models. Agent hours passed human hours in April (research) and May (all staff) and doubled about every five weeks. Read from Anthropic's chart.
How many workers could AI agents add?
Agents the AI chips shipped through 2026 and 2027 could run, in full-time-worker equivalents · Oct 2026 estimate
Central caseAll scenariosUS workers
Full-time-worker equivalents (log scale)
Use Tab to move through the rows.
Agents the chips shipped worldwide could run. An agent works 168 hours a week, as much as 4.2 full-time workers. The 20% case is Epoch's estimate of modest use.
International
The US against Europe: unemployment on the common international definition, firms' use of AI, hiring, who fills the jobs, software hiring and AI use. Eurostat publishes the map's unemployment rates about a month after the month ends; the UK's come later.
Hiring, euro area
Job postings on Indeed against a year earlier
−11%
United States +3%
Prior month −11%
Week to Oct 23Source: Indeed Hiring Lab
Workers born abroad, EU
Share of people employed who were born abroad
15.3%
United States 19.2%
+0.5 pp vs prior
2025
Working-age people, EU
People aged 15 to 64
285.8M
2035, UN projection 267.0M
−400k vs prior
2025Source: UN
Software job postings, Germany
Indeed postings for software development, February 2020 = 100
48
United States 79
−16.5% vs a year ago
Oct 2026Source: Indeed Hiring Lab
Unemployment across Europe
Unemployment rate, seasonally adjusted, latest month
2.8%11.0%no data
Click a country for its figures. Eurostat's harmonised rates, ages 15 to 74, as a share of the labour force; the UK from the OECD on the same definition. Countries report with different lags, so the latest month can differ.
Source: OECD ↗Job postings, month by month
Indeed job postings against the same month a year earlier, January 2022 to Oct 2026
Fewer postingsMore postings(−30% … +30%)
Use the arrow keys to move between months and countries. Press Escape to close the readout.
One cell per month. Postings on Indeed only, seasonally adjusted. Indeed's share of each job market differs, so each country is compared with its own past. Postings tend to move before unemployment. The last column compares the latest week with the same week a year earlier.
Source: Indeed Hiring Lab ↗Who filled the new jobs
Change in people employed by country of birth, 2022 to 2025
Born in the countryBorn abroadOne square = 100,000 jobs
Use Tab to move through the countries.
Outlined squares left of the line are jobs lost. In Germany, fewer people born there work than in 2022, so people born abroad took all the growth. Labour force surveys; UK left out.
Who leaves working age, who arrives
People reaching 65 and people reaching working age, 2025 to 2035
Reach 65Reach working ageShortfall
Use Tab to move through the countries.
Each grey figure is 10 of every 100 people who reach 65 by 2035; green figures, how many reach working age (15) in the same years. Before migration; UN estimates from 2024.
Source: UN World Population Prospects 2024 ↗Software developer job postings
Indeed postings for software development, monthly average, 1 Feb 2020 = 100, to Oct 2026
United StatesGermanyFranceUnited KingdomFeb 2020
Index
Use the left and right arrow keys to inspect months. Press Escape to close the readout.
Software development is among the jobs most exposed to AI. Interest rates and the end of the 2021–22 tech hiring boom also weigh on these postings.
AI use by people and by firms
Claude use per working-age person (April and May 2026) against the share of firms using AI (2025)
European countryLargest economiesUnited States
Claude use per person
Use Tab to move through the countries.
Claude use: share of Claude.ai use over share of working-age people (1 = in line). Firms: share with 10+ staff using AI. Eurostat has no US firm figure, so the US is a line.
Figures are shown as published by the named source and are revised by their publishers. This page is for information only and is not investment advice or a recommendation to buy or sell any security. This dashboard is maintained and updated with the help of AI. Errors can occur; we apologise for any.
Polymarket ↗