VC
Value Add VC
โšกHomePulseโšกHelpful Apps๐Ÿ“Blog
โ† Value Add PulseAI

It's an AI Web, We're Rats in the Walls

A Register column argues AI crawlers, agents and bot traffic have quietly become the majority presence on much of the web, reshaping how sites are built, defended and monetized around non-human visitors.

TC
Trace Cohen
Early-stage VC & angel ยท Founder, New York Venture Partners
July 12, 2026
2 min read
ShareXLinkedInEmail
THE RUNDOWN
1

The Register's July 12 column argues that AI crawlers, scraping bots and autonomous agents have become such a dominant share of web traffic on many sites that human visitors are increasingly the minority presence, inverting the assumption the web was built to serve people first

2

The piece describes site operators reworking infrastructure, rate-limiting and content-access strategy specifically around managing AI bot traffic -- from training-data scrapers to agent-driven browsing tools like OpenAI's now-discontinued Atlas -- rather than treating it as marginal background load

3

The framing connects directly to VentureBeat's slopsquatting research and CISA's own incident-response gaps, reinforcing a broader theme that the infrastructure and security assumptions built for a human-centric web are increasingly mismatched to actual current traffic patterns

4

For publishers and content businesses, the shift raises a genuinely unresolved commercial question: if a majority of a site's traffic is non-human and non-monetizable through traditional advertising, the underlying business model of large parts of the open web needs to be rethought, not just defended

TC
The VC Read ยท Trace's TakeTrace Cohen

If most of your site's traffic is already non-human, then ad-impression-based business models are quietly dying underneath publishers who haven't noticed yet. The founders who figure out how to charge AI agents and crawlers directly for access -- rather than just trying to block them -- are sitting on a real, underbuilt category right now.

The Register's July 12 column makes an argument that's more structural than it first sounds: AI crawlers, scraping bots and increasingly autonomous browsing agents have become such a dominant share of traffic on many websites that human visitors are, in a meaningful sense, becoming the minority presence -- rats in the walls of a house built for someone else.

The practical consequence is that site operators are reworking core infrastructure decisions specifically around managing non-human traffic: rate-limiting strategies, content-access gating, and increasingly aggressive bot-detection systems designed to distinguish training-data scrapers, search-indexing crawlers and agent-driven browsing tools -- like the agentic capability OpenAI folded into ChatGPT after discontinuing its standalone Atlas browser -- from genuine human visitors.

The theme connects directly to two other stories in this same news cycle: VentureBeat's slopsquatting research, which documents a new AI-specific software supply-chain attack vector, and CISA's own disclosed gap in incident-response readiness for cloud and code-repository exposures. All three point toward the same underlying pattern -- infrastructure, security and content-access assumptions built for a predominantly human web are increasingly mismatched to how the web actually gets used and abused today.

The commercial question underneath all of this is the one most publishers haven't fully answered yet: if a majority of a site's traffic is non-human and largely non-monetizable through conventional display advertising, the fundamental business model of large parts of the open web -- built on human eyeballs and ad impressions -- needs active rethinking, not just better bot-defense tooling bolted onto the existing model.

For founders building content, publishing or media businesses, the shift argues for treating AI-traffic monetization -- licensing content access to AI labs directly, building agent-friendly paid APIs, or gating premium content behind authentication that filters bot traffic by design -- as a first-order product and business-model decision, not an afterthought defensive measure. For infrastructure and security founders, bot-traffic management for an AI-dominated web is a genuinely underserved category with real, growing demand from any site operator dealing with this shift in real time.

The bear case: framing this as an existential crisis for the human web risks overstating a trend that may simply require better technical tooling to manage, rather than a wholesale rethink of how the internet works -- bot traffic has always existed, and the industry has adapted to prior waves of automated traffic before. What to watch next: whether major publishers begin disclosing bot-versus-human traffic splits more transparently, and whether AI labs move toward paying for content access at scale rather than relying on scraping that operators are increasingly motivated to block.

ShareXLinkedInEmail

Originally reported by The Register. Analysis and editorial commentary by Value Add Pulse.

โ† Back to Pulse

THE WIRE in your inboxโ€” Tech, startup & VC news with Trace's take. Free, no spam.

Read Next

AI

AI Demand Stays 'Almost Unlimited' as Buyers Get Pickier

Executives tell CNBC AI compute and chip demand remains "almost unlimited" even as enterprises shift to "valuemaxxing" -- squeezing more return from existing AI spend rather than simply buying more capacity.

AI

Memory Chipmakers Ride AI's Wildest Boom-Bust Cycle Yet

The Register argues memory chipmakers have always lived on a boom-bust cycle, but the AI-driven demand for HBM is producing the most extreme swing yet, with pricing and capacity decisions made years ahead of actual demand.

AI

While Neuralink Drills, China's BrainCo Bets on Wearables

As Musk's Neuralink pursues implanted brain-computer interfaces, China's BrainCo is betting the more scalable future of brain tech is noninvasive, wearable devices that read neural signals without surgery.

@Trace_Cohenยทt@nyvp.com