AI as a business story quadrupled on our site this week. Here's what we actually wrote about.

Key points

• AI2Day ran 4 stories in the AI business category between October 1 and October 7, 2026, up from 1 the previous week.
• Those 4 stories came from 4 separate outlets, so this is not one newsroom's obsession.
• AI business made up 10.5% of our 38 stories that week; the week before we published 39 stories total.
• Our archive only goes back to July 13, 2026, so we cannot tell you if this is normal or rare over the long run.

So here is what I measured, and only what I measured.

Last week AI2Day published 4 stories we filed under AI business. The week before, we published 1. That is a jump from 1 to 4, which is a 300% rise, out of 38 stories total for the week against 39 the week before. Same size paper, very different mix on one shelf.

Four separate outlets filed the pieces we ended up carrying. That matters more than the raw count. When one outlet suddenly cares about something, it is a hobbyhorse. When four do in the same seven days, something is pulling their attention in the same direction.

What were the four stories actually about?

Two were about money and governance at the big labs. One was about enterprise software doing things on its own. One was a practical column about which Google features might pay you back.

The IPO thread is the loudest. Sam Altman said OpenAI will not go public until its AI is safe enough, published October 3. The next day's news cycle gave us Anthropic's IPO filing, which warns its own AI could try to blackmail people and resist being shut down, published October 2. Two of the biggest AI labs, both talking about public markets in the same week, in very different tones. That alone would move a category.

Then there is Enterprise AI Stopped Giving Advice. Now It's Making the Calls, from October 5. That is about AI agents, software that carries out multi-step work without a human clicking each step, showing up inside companies that are paying for it. And what from Google's September AI pile actually earns you anything, from October 2, is the one closest to what readers of my column usually ask me: does any of this put money in my pocket.

Does a jump from 1 to 4 mean anything?

Honestly, on its own, barely. Four stories is a small number. One quiet week followed by one busy week is exactly what random clumping looks like in a small sample.

What makes me take it seriously is the spread across outlets and the shared subject matter. IPO filings are documents, not vibes. If two frontier labs are talking publicly about going public inside the same seven days, that is a real event in the business of AI, not a mood.

What would confirm this, and what would kill it?

Confirming: another week at 3 or more AI business stories, with at least one more lab or large buyer disclosing financials, pricing or governance terms. If the IPO thread keeps producing filings, the category stays elevated on its own.

Killing it: next week drops back to 1 or 0, no follow-up filings, and the enterprise agent story turns out to be one vendor's launch cycle rather than something customers are paying for at scale.

I will be watching the ratio, not the raw count. If AI business holds above 10% of our weekly mix for two more weeks, I will write that up. If it does not, I will tell you that too.

One honest takeaway, in the spirit of this column: a 300% rise from a base of 1 is a real measurement and a tiny sample at the same time. Both things are true. Do not rearrange your savings around a four-story week.