Because we cover so much, and Wednesday’s PN headline last week didn’t focus on it, here’s what one of the PDFs inside the PN report read:

That “change vector call” got real this week when NBC picked up the story from an X-post:

Only days after we began asking whether server statistics were pointing toward a Digital Waterloo—a moment when “web traffic” no longer meant “human traffic”—Cloudflare’s CEO announced that the crossover has apparently happened. According to Cloudflare, 57.4 percent of web requests are now machine-generated while only 42.6 percent come from humans. In one sentence, the internet became more machine conversation than human conversation.

The really important sentence, however, wasn’t the percentage. It was Cloudflare CEO Matthew Prince saying that “the web actually shrank” from 2015 through 2025, but during the last six months has exploded again because AI agents are now crawling, reading, comparing, and interacting with thousands of sites at a time. That neatly explains the mystery many independent publishers have been seeing in their own logs: rising bandwidth, changing traffic patterns, collapsing ad economics, and statistics that no longer seem to correspond to actual human readership.

Which means the next economic question isn’t “How many visitors do you have?” The next question is “How many of them are people?” And beyond that lies an even larger question: if bots don’t click ads, then the business model of the web itself must change. That’s why I suggested the Digital Waterloo may already be underway—not because the internet is dying, but because its accounting system is.

The open vector now is how far this newly acknowledged schema shift travels. If the web has crossed from human-majority traffic to bot-majority traffic, then every ad model, analytics dashboard, sales funnel, SEO plan, and media valuation built on “traffic” has to be re-priced. That is not a tech footnote. That is a business-model earthquake.

Which brings us to the larger question: could bot-majority traffic become one of the New ’29 drivers? In 1929, part of the collapse involved discovering that many financial assumptions were not as solid as advertised. In 2026, we may be discovering something similar about the attention economy: the audience numbers may still be there, but the humans behind them may not be.

Stand by. Data’s beginning to roll.

~Ure