Turns out Dead Internet Theory was right: AI agents are eating the Web, growing by nearly 8,000% and rewiring the Internet’s business model

Bot traffic now dominates online browsing: Cloudflare pegs June crossover at 57.5% of webpage requests from bots (Thales dates it to 2023 with 53% in its report). Agentic web activity is accelerating sharply—HUMAN Security cites +7,851% YoY for agents that take actions (vs. +597% for scraper traffic)—but measurement and detection remain inconsistent, creating analytics and security risk for businesses and developers. Pitchbook estimates only ~1% of the ~$20T addressable “agent work” is flowing via agents today (with “agent GDP” run-rate estimated at $36B), implying adoption is early while monetization, payments, and liability infrastructure are still lagging.

Analysis

The cleanest winner is infrastructure that sits between agents and the open web: bot traffic forces more spend on verification, rate limiting, and authenticated access, which is structurally favorable for NET and, to a lesser extent, other edge-security vendors. This is not a one-quarter story; over the next 1-3 months the market should start rewarding commentary around AI-driven traffic mix, but the bigger 6-18 month effect is higher attach rates for bot management and API security as enterprises redesign around machine-to-machine usage.

The less obvious loser is the measurement/monetization stack built on human-behavior assumptions. If session quality, bounce rate, and conversion signals get noisier, performance marketing ROI becomes harder to trust, which can pressure spend by consumer internet and ad-tech names even before top-line damage shows up. That argues for caution on businesses whose unit economics depend on clean attribution; the first-order problem is not lost traffic, it is mispriced traffic.

RAMP and V are longer-dated beneficiaries if agents become embedded in payments and enterprise workflows, but the near-term monetization is limited by liability and authorization plumbing. RAMP can gain from automated expense/payment initiation, yet the revenue delta will likely be incremental until agent identity and controls are standardized. V is more of a call option on machine-authenticated payments; the market should not pay up for that optionality until there is visible adoption in issued-card or tokenized rails.

Contrarian view: the consensus may be overestimating how fast this becomes a revenue story and underestimating how quickly it becomes a security and data-quality story. The visible traffic explosion is real, but economic conversion remains tiny; that means the immediate trade is in picks-and-shovels security, not in sweeping AI-commerce beneficiaries. The thesis would be falsified if next earnings show no incremental spend on bot mitigation or if managements say AI traffic is still immaterial to margins and CAC.

Key Decisions for Investors

• Long NET on a 1-3 month horizon; use pullbacks to add. Risk/reward is attractive because higher bot volume should translate into higher WAF/bot-management attach, while downside is limited unless management signals traffic growth is not monetizing into billings.
• Small tactical long RAMP vs. a basket of consumer internet names that rely on attribution (e.g., DASH/other demand-gen-heavy platforms) over 3-6 months. Thesis: agentic workflows increase workflow automation spend, but size small because the revenue conversion is still early.
• Do not chase V on this headline alone; treat it as a 6-18 month watchlist name. Re-rate only if card-present or tokenized machine payments show up in processed volume commentary; otherwise the optionality is too far out to underwrite.
• Pair idea: long NET / short an ad-tech or analytics proxy on any rally in the next 2-4 weeks. The market is likely to overprice the infrastructure beneficiaries before it fully discounts deteriorating signal quality in ad measurement.
• Set alerts for earnings language on 'AI traffic,' 'bot mitigation,' and 'API authentication.' Falsify the bullish infra thesis if those keywords do not translate into higher ARR, usage, or billings within the next two reporting cycles.