GenAI PMGuillermo Rauch
Founder and leader at Vercel, cited for data on bot traffic and agentic usage. His comments reflect how websites and docs are increasingly consumed by agents rather than humans.
Key Highlights
- Guillermo Rauch said 58.18% of recent Vercel network traffic was bot-originated, up from 32% in January 2024.
- He reported that more than 60% of Vercel deployments are now agentic, compared with about 3% in January 2026.
- Rauch said agents account for up to 83% of pageviews on Vercel documentation sites, reinforcing the need to design content for machine readers.
- He announced gdp-ts, a TypeScript library and linter that verifies authorization proofs for sensitive APIs at compile time.
- His recent commentary consistently frames security, credentials, and sandboxing as core requirements for production AI agents.
At a glance
- Role
- Founder
- Company
- Vercel
- X / Twitter
- @rauchg
Guillermo Rauch
Overview
Guillermo Rauch is the founder and leader of Vercel and a prominent voice on how AI agents are changing software creation, deployment, and web consumption. Across product announcements and public commentary, he has highlighted a shift from human-centric workflows toward agentic ones, including data showing that bot-originated traffic and agent-driven deployments now make up a large share of activity across the Vercel network.For AI Product Managers, Rauch matters because his updates offer concrete signals about where agentic product design is heading: infrastructure bought and operated by agents, documentation increasingly consumed by bots, stronger security and authorization controls for AI systems, and developer tooling optimized for autonomous workflows. His comments are especially useful for PMs building AI-native developer products, agent platforms, and workflow automation systems.
Key Developments
- 2026-09-25 — Rauch shared ShellPerfBench and said Opus 5.5 found shell startup optimizations that other models missed, suggesting agents can improve local developer environments such as `.zshrc` and related files.
- 2026-09-27 — He warned that persistent low-quality, unverified AI prose could reduce trust in reading, and he pushed back on a viral claim by noting that a cited performance improvement was attributed in the PR description to algorithms and data structures rather than a compiler change.
- 2026-09-29 — Rauch announced he ported his Mini web browser to Rust and Swift, packaged Chromium via the `cef` crate, and embedded an agent connected to his local `fx` CLI over ACP. He said moving beyond Electron improved startup speed and delivered a more native Mac experience.
- 2026-10-01 — He said Connect helps agents and apps avoid static credentials, framing it as a more secure and easier way to build integrations and distribute services to developers and the agents they ship.
- 2026-10-02 — Rauch commented that tokens are not served when ZDR requirements are set and no suitable providers are available, while noting there are fallback options and sufficient capacity.
- 2026-10-04 — He announced that a KVM zero-day was confirmed through Vercel’s Sandbox bounty program, describing KVM as the industry’s gold-standard Linux virtualization approach. He also argued that security will become an increasingly large function inside software companies, including verification engineering and capital allocation.
- 2026-10-05 — Rauch said `fx.sh` had become much faster, arguing that harness overhead matters more as model latency drops with systems such as Astra ultrafast. He also said the next release would improve session storage, retrieval performance, and cloud durability in `libfx`.
- 2026-10-06 — He announced `gdp-ts`, a TypeScript library, linter, and AI skill for safer API design. The system requires authorization “proofs” for sensitive functions and verifies them at compile time.
- 2026-10-10 — Rauch announced that agents can now purchase domains through a marketplace CLI, extending agent purchasing from infrastructure products and services toward launching online businesses end to end.
- 2026-10-11 — He shared that 58.18% of traffic across the Vercel network over the prior 30 days was bot-originated, up from 32% in January 2024. He also said more than 60% of Vercel deployments are now agentic, up from about 3% in January 2026, and that agents account for up to 83% of pageviews to Vercel documentation sites.
Relevance to AI PMs
1. Use his metrics as product-planning signals for agentic demand. Rauch’s data on bot traffic, agentic deployments, and documentation consumption suggests PMs should design docs, APIs, onboarding, and analytics for machine consumers, not just human users.2. Treat security and authorization as core AI product surfaces. His comments on sandbox security, ZDR constraints, credential handling, and `gdp-ts` point to a practical roadmap: build least-privilege access, verifiable authorization, and secure execution into agent products from the start.
3. Optimize for end-to-end autonomous workflows. From domain purchasing to infrastructure buying to CLI-based integrations, Rauch’s work suggests that successful AI products will let agents complete real business tasks, not just generate suggestions inside a chat box.
Related
- Vercel — The company Rauch leads and the source of the network-level data on bot traffic, agentic deployments, and documentation consumption.
- Next.js / nextjs / nextjs-162 / nextjs-163 — Closely tied to Rauch through the Vercel ecosystem and broader developer-platform strategy.
- AI SDK / vercel-ai-sdk / ai-sdk / aisdk — Related to Vercel’s tooling for building AI applications and agents.
- AI Gateway / vercel-ai-gateway / vercel-ai-gateway-api — Connects to Rauch’s comments around provider choice, capacity, and ZDR requirements.
- Sandbox / vercel-sandbox / sandbox-at-vercel / sandboxes — Important in his security and agent execution narrative, including the KVM zero-day disclosed via the sandbox bounty program.
- Connect / vercel-connect — Linked to his view that agents and apps should avoid static credentials in integrations.
- fx / fxsh / fx.sh / libfx — A recurring area in Rauch’s public experimentation around fast agent harnesses, session durability, and local-agent workflows.
- gdp-ts — A concrete example of his push toward compile-time authorization guarantees for AI-era API design.
- Agents / AI agents / coding-agents / autonomous-agents — Central to his thesis that software usage and deployment are increasingly agent-driven.
- Anthropic, OpenAI, Claude, Grok, Kimi, Sol, Astra ultrafast — Model ecosystem references that appear in his comments on performance, evaluation, and agent tooling.
Frequently asked questions
What is Guillermo Rauch?
Founder and leader at Vercel, cited for data on bot traffic and agentic usage. His comments reflect how websites and docs are increasingly consumed by agents rather than humans.
What should AI product managers know about Guillermo Rauch?
Guillermo Rauch said 58.18% of recent Vercel network traffic was bot-originated, up from 32% in January 2024. He reported that more than 60% of Vercel deployments are now agentic, compared with about 3% in January 2026. Rauch said agents account for up to 83% of pageviews on Vercel documentation sites, reinforcing the need to design content for machine readers.
Newsletter Mentions (131)
“#7 𝕏 Guillermo Rauch shared that 58.18% of traffic across the Vercel network was bot-originated over the last 30 days, versus 32% in January 2024; more than 60% of Vercel deployments are now agentic, up from approximately 3% in January 2026.”
#7 𝕏 Guillermo Rauch shared that 58.18% of traffic across the Vercel network was bot-originated over the last 30 days, versus 32% in January 2024; more than 60% of Vercel deployments are now agentic, up from approximately 3% in January 2026. Agents also account for up to 83% of pageviews to Vercel’s documentation sites, with that share increasing as Vercel optimizes content for agents.
“Guillermo Rauch announced that agents can now purchase domains, extending their existing purchases of infrastructure products and services through a marketplace CLI and enabling them to move from idea to online business.”
Guillermo Rauch announced that agents can now purchase domains, extending their existing purchases of infrastructure products and services through a marketplace CLI and enabling them to move from idea to online business.
“Guillermo Rauch announced gdp-ts, a TypeScript library, linter, and AI skill for safer API design that requires authorization “proofs” for sensitive functions and verifies them at compile time.”
#7 𝕏 Guillermo Rauch announced gdp-ts, a TypeScript library, linter, and AI skill for safer API design that requires authorization “proofs” for sensitive functions and verifies them at compile time. Its README and examples model a Vercel API constraint requiring proof of a certain role and entitlement before changing a Project password. Also covered by: @Guillermo Rauch
“Guillermo Rauch said fx.sh is now much faster as harness overhead becomes increasingly important with faster models such as Astra ultrafast. He announced that the next release will further improve session storage and retrieval performance and deliver a major cloud-durability improvement in libfx.”
#13 𝕏 Guillermo Rauch said fx.sh is now much faster as harness overhead becomes increasingly important with faster models such as Astra ultrafast. He announced that the next release will further improve session storage and retrieval performance and deliver a major cloud-durability improvement in libfx.
“Guillermo Rauch announced that a KVM zero-day was confirmed through Vercel’s Sandbox bounty program, affecting what he called the industry’s gold-standard solution for Linux virtualization.”
GenAI PM Daily - Perplexity Computer now generates interactive 3D learning visualizations October 04, 2026 GenAI PM Daily 🎧 Listen to this brief 3 min listen Today's top 12 insights for PM Builders from X and LinkedIn. Perplexity Computer now generates interactive 3D learning visualizations #5 𝕏 Guillermo Rauch announced that a KVM zero-day was confirmed through Vercel’s Sandbox bounty program, affecting what he called the industry’s gold-standard solution for Linux virtualization. He thanked Paulos and other researchers for helping improve sandbox security for agents and said a full writeup was coming. #9 𝕏 Guillermo Rauch predicted security will become an increasingly large function in software companies, spanning verification engineering and capital allocation.
“Guillermo Rauch commented that tokens aren’t served when ZDR requirements are set and no suitable providers are available, while noting there are plenty of fallbacks and capacity.”
#20 𝕏 Guillermo Rauch commented that tokens aren’t served when ZDR requirements are set and no suitable providers are available, while noting there are plenty of fallbacks and capacity.
“He said Connect helps agents and apps avoid static credentials, making integrations easier and more secure.”
Guillermo Rauch commented that service operators can add their services to Connect to reach 20M+ developers and the billions of agents they’ll ship. He said Connect helps agents and apps avoid static credentials, making integrations easier and more secure.
“in Guillermo Rauch announced that he ported his Mini web browser to Rust and Swift, bundling up-to-date Chromium via the cef crate and embedding an agent that connects to his local fx CLI over ACP.”
GenAI PM Daily - Anthropic announces faster, lower-cost Claude Sonnet 5.5 September 29, 2026 GenAI PM Daily 🎧 Listen to this brief 3 min listen Today's top 20 insights for PM Builders, ranked by relevance from Blogs, X, YouTube, and LinkedIn. Anthropic announces faster, lower-cost Claude Sonnet 5.5 #8 in Guillermo Rauch announced that he ported his Mini web browser to Rust and Swift, bundling up-to-date Chromium via the cef crate and embedding an agent that connects to his local fx CLI over ACP. Built on fx.sh with Opus 5.5 and Sol 6, the project uses fx to manage the browser through an MCP server. Rauch says replacing Electron enabled Liquid Glass, faster startup, and a more Mac-native experience.
“He cited a viral thread attributing a performance gain to a compiler change and said the referenced PR description instead attributed it to changes in algorithms and data structures.”
#13 𝕏 Guillermo Rauch warned that persistent low-quality, unverified AI prose could cause people to discount reading. He cited a viral thread attributing a performance gain to a compiler change and said the referenced PR description instead attributed it to changes in algorithms and data structures.
“In Guillermo Rauch shared ShellPerfBench, saying Opus 5.5 found shell startup optimizations other models missed.”
#7 in Guillermo Rauch shared ShellPerfBench, saying Opus 5.5 found shell startup optimizations other models missed. He suggested asking an agent to optimize `.zshrc` and related files.
Related
An AI company mentioned for releasing a report on model behaviors and for internal tool use in later context. Relevant for PMs tracking safety reporting and frontier model operations.
Anthropic’s coding agent/environment used here to control physical tools in a long-running home experiment.
OpenAI is an AI company building foundation models and consumer products like ChatGPT. In this newsletter it is discussed in connection with GPT-6, Dots, ChatGPT for Teens, and College Planner.
Anthropic’s conversational AI model family, referenced here as handling a task better than the speaker since around Opus 4.8. It is being discussed in terms of agent capability and workflow delegation.
An OpenAI coding model/tool used here as part of coding harnesses turned into reinforcement-learning environments.
An AI model hub and community platform referenced through comments about its usage and download metrics. The newsletter suggests DAUs are more reliable than raw download counts for understanding growth.
A developer platform whose network traffic and deployments are increasingly agentic according to the newsletter. It is also optimizing documentation content for agents.
A creator and product thinker sharing AI usage tracking and product documentation workflows. Relevant to PMs focused on measurement and prompt-to-spec workflows.
Google’s research organization, referenced for a clinical evaluation of its medical AI system AMIE.
Entrepreneur who shared a workflow tool that imports data from LinkedIn, X/Twitter, and Granola.
MCP is a protocol for connecting AI systems to external tools and data sources. Here it is described as the way Granola can use captured context through internal tooling.
A company building GPUs and AI model platforms, referenced for its open-model ecosystem and enterprise model usage. Relevant to AI PMs evaluating open model stacks and deployment cost/performance.
Technology company referenced as the place where John Provine previously built evaluation systems. Relevant here as a source of product experimentation rigor.
An AI infrastructure company focused on agent tooling and harnesses. The newsletter cites its explanation of where typed yes/no calls fit inside a harness.
A major AI company involved here in developing the open Personal Agent Protocol with Sierra. The mention is relevant to interoperability and agent permissions.
A collaborative AI product design/building tool that now supports team visibility and shared workspaces. Useful for PMs collaborating on prototype and product workflows.
A commentator or recapper cited for discussing agentic product design principles. He emphasizes components, signals, assembly rules, and constraints for intent-based interfaces.
Vercel’s model-routing and gateway product, mentioned in relation to spend and token share shifts across providers. Relevant to AI PMs for model traffic management and provider strategy.
Autonomous or semi-autonomous workflows that perform tasks, coordinate tools, and execute actions on behalf of users.
A project management and issue-tracking tool used in agentic software workflows. In this newsletter it is used by Warp’s Wilson to open issues from Slack requests.
A collaborative design platform referenced as an example of broad enterprise SaaS that may remain resilient in the AI era. It is contrasted with niche single-purpose products.
A Claude model version praised for personality and writing style. The newsletter contrasts it with Opus 5 as more concise and friend-like.
A GPT model release referenced as an impressive model by Kevin Weil. For AI PMs, it represents continued frontier-model iteration and user expectation growth.
A system for handling many agent steps as typed questions with calibrated probabilities. It is positioned as a harness component for inexpensive yes/no decisions.
Vercel’s AI infrastructure layer for routing and managing model usage. The newsletter notes strong token volume growth on the gateway.
A 2.8T-parameter open-weight model described as frontier-level by the speaker in the newsletter. It is notable for strong quality and deployment on Nebius Token Factory.
A protocol or capability layer mentioned as part of an open, composable extension philosophy for AI tooling. It is grouped with MCP and Plugins.
A developer tool or coding environment that announced availability of Qwen3.8-Flash. It appears as the platform distributing model access.
A React framework used to build web applications and evaluated here with model-based coding benchmarks. It is referenced in relation to fresh evals and model performance.
Agents used to write, review, and iterate on code as part of software development workflows. The newsletter frames them as shifting developers toward specification, architecture, and evaluation work.
Skills.sh is a site hosting agent skills and tutorials, including frontend best-practices guidance. Here it hosts Vercel Labs' React best-practices tutorial.
Vercel’s SDK for integrating AI features into apps. The newsletter highlights token savings from a single line of code in DeepSeek-powered workflows.
A framework for internal agents that emphasizes instructions, skills, channels, and connectors. It is presented as a default choice for building internal agent systems.
A convention for storing agent instructions in a repository. Here it is cited as one place where preferences can be set for fx.sh.
Google's suite of productivity applications used for email, documents, spreadsheets, and calendaring. It is mentioned here as the environment Cursor agents can now operate across.
Agent Skills are reusable capability modules or instructional patterns for agents. The newsletter references a React best-practices tutorial framed as an agent skill.
Vercel's command-line interface, described here as a self-updating, zero-dependency binary. It is positioned as central to the 'cloud for agents' with usage across agentic coding tools.
A paradigm that treats cloud infrastructure as autonomous coding agents to automate deployment and operations. For AI PMs, it reframes infrastructure as an agentic workflow rather than a static system.
Programmable interfaces that let AI agents and software systems access services and complete tasks. The newsletter positions APIs as one of the means for agents to act on behalf of users.
Vercel Queues is a developer tool for queue-based workflows, designed to simplify background processing and agentic systems.
The latest Next.js release positioned as agent-native, with features intended to help AI agents debug and optimize applications in a specific versioned codebase.
An OpenAI model variant discussed here for its ability to collaborate with HarmonicMath on near-autonomous proof generation. For AI PMs, it highlights stronger reasoning and math capabilities in advanced LLMs.
Stay updated on Guillermo Rauch
Get curated AI PM insights delivered daily — covering this and 1,000+ other sources.
Subscribe Free