Attention is narrowing, and it has changed subject.
The five most-covered subjects now take 23.6% of all named-subject events, against 33.4% in the preceding ninety days. Nvidia alone moved from 1.6% to 3.4%. Over the same span, Product gave up 7.1pp of the cycle while Funding took +7.6pp.
- Events counted
- 5,573
- Named subjects
- 1,992
- Categories
- 17
- Months
- 19
Subjects
Share of named-subject events, trailing 90 days against the 90 before.
| Subject | Then | Now | Movement | ±pp |
|---|---|---|---|---|
| Nvidia | 1.6 | 3.4 | +1.8 | |
| United States | 0.9 | 1.8 | +0.9 | |
| UK Government | 0.0 | 0.7 | +0.7 | |
| Amazon | 0.7 | 1.3 | +0.6 | |
| CXMT | 0.0 | 0.5 | +0.5 | |
| Australian Government | 0.0 | 0.5 | +0.5 | |
| Losing ground | ||||
| 8.1 | 4.0 | −4.1 | ||
| OpenAI | 10.9 | 7.6 | −3.3 | |
| Anthropic | 9.3 | 6.6 | −2.7 | |
| Microsoft | 3.2 | 1.7 | −1.5 | |
| Apple | 1.9 | 0.9 | −1.1 | |
Categories
Share of all events. Rotation shows here before it shows in headlines.
| Category | Then | Now | Movement | ±pp |
|---|---|---|---|---|
| Funding | 8.0 | 15.6 | +7.6 | |
| Product | 20.0 | 12.9 | −7.1 | |
| Investment | 6.3 | 13.3 | +7.1 | |
| Release | 9.6 | 3.1 | −6.5 | |
| Research | 14.6 | 9.6 | −5.0 | |
| Policy | 7.5 | 11.6 | +4.1 | |
| Legislation | 2.4 | 4.7 | +2.3 | |
| OpenSource | 4.0 | 2.3 | −1.7 |
Gone quiet · Cohere, Trump administration, TikTok — each carried eight or more events in the archive and none in the last ninety days.
Listed vs not · 47.1% of coverage we can classify by listing status goes to companies that are not publicly traded — effectively unchanged from 49.6% the quarter before. OpenAI and Anthropic alone account for 14.2% of every named-subject event in the archive. Listing status is a maintained list, not archive data; it covers 35.1% of named-subject events.
Method
Every figure is a full count of the archive, not a sample. Coverage runs 25 Mar 2025 to 23 Sept 2026.
Shares are reported rather than totals because publishing volume is not constant — 938 events in the current window against 574 before it. Comparing raw counts across those windows would show a decline everywhere and mean nothing.
Subject labels are normalised onto head entities, so Google DeepMind and Google Cloud count as Google. Generic labels such as “AI” or “Big Tech” are excluded.
The same archive, read once a week by a human.
Dearer Memory, a $278bn Plan, and ZCode's Silent Snapshots. Two insights, each with the case against it. Tuesdays, 7:30am UK.
Read it on LinkedIn →