The same topics, in state bills
Each row counts the bills and resolutions in state legislatures whose title or summary holds one of a topic’s written phrases, week by week, back to 2017. The count says how often lawmakers named a subject, not what they proposed, and it does not measure what people think.
These counts are kept apart from the Federal Register counts on Trends and from company filings, and are never added to them. The topics are the same fixed list. This source was added on November 6, 2026.
The latest counts
Through the week of October 19, 2026, from the file of November 1, 2026: none of seventeen topics pass the test in state bills.
No growing or fading reading is given here, because bills follow legislative calendars. Change compares each topic’s share of all bills over 52 weeks with the 52 before, because more bills are filed in odd years.
| Topic | 12 weeks | 52 weeks | Change in share | Legislatures | Regions | Weeks held | Test | Windows passed |
|---|---|---|---|---|---|---|---|---|
| Artificial intelligence | 32 | 514 | +92% | 8 | 3 | 10 of 12 | Does not pass | 30 |
| Data centers | 53 | 410 | +352% | 10 | 4 | 9 of 12 | Does not pass | 21 |
| Quantum computing | 0 | 3 | – | 0 | 0 | 0 of 12 | Does not pass | 0 |
| Unmanned aircraft | 2 | 56 | +13% | 1 | 1 | 1 of 12 | Does not pass | 0 |
| Critical minerals | 0 | 4 | – | 0 | 0 | 0 of 12 | Does not pass | 0 |
| Supply chains | 0 | 23 | – | 0 | 0 | 0 of 12 | Does not pass | 0 |
| Foreign adversaries | 0 | 63 | +34% | 0 | 0 | 0 of 12 | Does not pass | 0 |
| Critical infrastructure | 1 | 80 | +22% | 1 | 1 | 1 of 12 | Does not pass | 0 |
| Avian influenza | 0 | 0 | – | 0 | 0 | 0 of 12 | Does not pass | 0 |
| Measles | 0 | 1 | – | 0 | 0 | 0 of 12 | Does not pass | 0 |
| Fentanyl | 0 | 60 | −29% | 0 | 0 | 0 of 12 | Does not pass | 0 |
| PFAS | 2 | 97 | +8% | 2 | 2 | 1 of 12 | Does not pass | 3 |
| Wildfire | 2 | 113 | −30% | 1 | 1 | 2 of 12 | Does not pass | 5 |
| Small modular reactors | 0 | 8 | – | 0 | 0 | 0 of 12 | Does not pass | 0 |
| Digital assets | 2 | 67 | +34% | 2 | 2 | 2 of 12 | Does not pass | 0 |
| Misinformation | 0 | 9 | – | 0 | 0 | 0 of 12 | Does not pass | 0 |
| Bell to bell | 0 | 5 | – | 0 | 0 | 0 of 12 | Does not pass | 0 |
What passing means
The test: no topic passes
A topic passes with a bill in each of the latest 12 weeks, from 10 or more legislatures in at least 3 regions. Windows passed counts the 12-week windows ending in the last 52 weeks that met the same test. In the last year, these passed in at least one window: artificial intelligence, data centers, PFAS and wildfire.
Macro trends: none
Passing here alone is not a macro trend. Each legislature has its own owner, but they are one kind of publisher. A macro trend must also pass in company filings, under Method 1.3. No topic passes in both.
How to check a count
- No single link
- The file has no public search, so a count here cannot be opened with one link, as it can on Trends. It takes the steps below.
- The bills
- Every bill behind the counts, with its date, its topics and a link to the legislature’s own page. Any one can be opened and read.
- The file
- The dated database file from Open States, about 11 GB, written November 1, 2026. The script reads its first 1,167,790,752 bytes, whose SHA-256 is
d83c44408de9cf649aa7c334a8617fbf5b9c8c9b85ada0d7a74ba82c1861a1f7. - The script
- Put the script, topics.json and states-topics.json in an empty folder and run
node fetch-trends-states.mjs --month=2026-11. It needs Node 24 and reads about 1.2 GB. The lastsha256in the file it writes should be9cee57e25b60ac0d810b4bb6aa43f5801953adf9178ce5ab19eadb721487d827. - One state
- Every run compares the file with Washington’s own list of bills. For January 1, 2026 to October 25, 2026, all 1,206 bills on its list were in the file, and phrase matches came to 26 on the list and 26 in the file. No other state is compared.
How this is counted
- Source
- Open States, run by Plural Policy, which gathers bills from each legislature’s site. Its data is offered under a public domain dedication, and it gives no guarantee of the file’s format or availability.
- Unit
- A bill or resolution, by the date of its first action. Bills filed before a session count from that earlier date: in Washington in 2026, 28 percent of bills had a first action one to six weeks before the date the state calls introduced.
- Matching
- A phrase matches whole words in the title or summary, so “fentanyls” does not match “fentanyl”. Full bill text is not used: in a test of 70,385 bills, 26 percent had none in the file, and nearly all of California’s were missing. So these counts are a floor. In that test, full text found about 2 to 30 times as many bills for most topics.
- Summaries
- Only 14 states have a summary for more than half of their bills, so a state with summaries has more words to match.
- Legislatures
- The 50 states, the District of Columbia and Puerto Rico, each with at least one bill in the latest 12 weeks. Congress is left out, because it belongs to the federal government.
- Regions
- The four US Census regions, and Puerto Rico, by legislature. This grouping is proposed.
- Calendar
- Over the last 52 weeks, all bills ran from 114 to 16,640 a week. A topic can pass in spring and fail in autumn with no change in interest, so the 12-week reading and the micro test are not run on this source.
- The 52-week change
- Bills per 1,000 of all bills, the latest 52 weeks against the 52 before, with no percentage under 104 bills. All bills came to 104,334 and 164,547.
- Not checked
- Each bill is counted as the file holds it. No copy check is run and no item has been audited under stage 2.2.
- Updates
- By hand, once a month, because the file is written on the first of the month. This count was retrieved November 6, 2026, and the latest window runs from August 3, 2026 to October 25, 2026. Bills reach the file late, so the latest eight weeks usually rise. Later files have changed 12 past weeks of a topic, listed in the JSON download.