Illustration for: Cloudflare's Prince: Bots Now Outnumber Humans Online

Cloudflare's Prince: Bots Now Outnumber Humans Online

Cloudflare CEO Matthew Prince says bots now account for 57.4% of web traffic versus 42.6% from humans, and argues the internet's future is 'pay to crawl' for AI companies that scrape content to train models.

By the Numbers

57.4%
Bot share of traffic
42.6%
Human share of traffic
1,000:1
5-yr bot:human projection
TC
By the Markets Desk
Edited by Trace Cohen · Early-stage VC & angel · Founder, New York Venture Partners
2 min read
ShareXLinkedInEmail

THE RUNDOWN

1

Cloudflare Radar's own data shows bots crossed the majority threshold of global HTTP requests earlier and faster than Prince previously projected -- he had forecast bot traffic overtaking human traffic by 2027, and it happened this year instead.

2

Prince's 'pay to crawl' framing, backed by a new digital wallet letting agents pay fractions of a cent per request, is a direct monetization play for Cloudflare sitting at the center of a huge share of global web infrastructure -- not just a policy opinion.

3

Apple, Google and Microsoft agreeing to Cloudflare's 'Disallow AI Training' setting, which lets sites stay indexed for search while blocking AI-training scraping, shows the largest AI labs are willing to accept some limits on training-data access when a dominant infrastructure provider sets the terms.

4

At the current growth rate, Prince projects automated traffic could reach 1,000 times human traffic within five years -- a trajectory that, if directionally correct, changes the basic economics of running any content-driven website.

TC

The VC Read · Trace's Take

Trace Cohen

Cloudflare has an obvious commercial incentive to declare the pay-to-crawl era arrived right as it ships the wallet to collect on it -- the real test is whether AI labs without Big Tech's legal budget and negotiating leverage adopt the same rails or just route around Cloudflare-protected sites entirely. Watch which mid-tier labs sign on, not just Apple, Google and Microsoft.

Analysis

Cloudflare CEO Matthew Prince said bot traffic has overtaken human traffic online faster than he expected, with Cloudflare Radar showing bots now account for 57.4% of HTTP requests globally versus 42.6% from humans, according to Fortune. Prince's conclusion for where the web goes from here: "Clearly it's going to be pay to crawl."

A Forecast That Arrived Early

Prince had previously projected bot traffic would exceed human traffic by 2027; Cloudflare's own data shows it crossed that threshold in May 2026, well ahead of schedule. Much of the acceleration comes from AI crawlers ingesting web content to train models -- traffic that, as Prince frames it, consumes a site's bandwidth and compute without sending a reader, or any revenue, back to the publisher. Pulse has previously covered Cloudflare's earlier pay-per-crawl policy push; this week's disclosure is the clearest numeric evidence yet that the underlying trend it responded to has only accelerated.

The Infrastructure Play Behind The Framing

Cloudflare's "Disallow AI Training" setting lets a website remain indexed for search while blocking AI crawlers from using its content for model training -- a middle path between total openness and blocking search entirely. Apple, Google and Microsoft have all agreed to respect the setting, a meaningful concession from three of the largest AI-model operators to a policy set by an infrastructure provider rather than negotiated bilaterally with individual publishers. Cloudflare has gone further, unveiling a digital wallet designed to let AI agents pay a fraction of a penny per request -- infrastructure meant to make "pay to crawl" an actual functioning payment rail rather than just a licensing negotiation.

Why This Matters Beyond Cloudflare

Cloudflare sits in front of a large share of global web traffic, which gives Prince's numbers unusual authority relative to any single publisher's anecdotal complaints about AI scraping. If bots really do reach 1,000 times human traffic within five years, as Prince projects at the current growth rate, the basic economics of running a content business -- built for decades on the assumption that a human reader eventually sees an ad or converts -- shift meaningfully toward a model where AI-agent traffic itself has to become a direct revenue source, not just an operating cost to absorb.

Cloudflare's Own Incentive

The bear case: Cloudflare has an obvious commercial incentive to frame the problem in a way that positions its own new payment infrastructure as the solution, and "pay to crawl" as a widely adopted standard requires AI labs broadly agreeing to pay for access they can currently take for free in many cases -- something Apple, Google and Microsoft doing so for search-indexing exceptions doesn't yet prove will extend to full training-data licensing at scale. Whether smaller AI labs and startups, without Big Tech's negotiating leverage or legal budgets, adopt the same payment rails or simply route around Cloudflare-protected sites is the open question that determines whether this becomes an actual internet-wide norm or stays a policy Cloudflare enforces only where it has direct infrastructure control.

ShareXLinkedInEmail

Key Sources

2 sources

Reported by Fortune · Analysis by Value Add Pulse.

← Back to Pulse

THE WIRE in your inbox— Tech, startup & VC news with Trace's take. Free, no spam.