Software development lifecycle (SDLC) automation startup Factory has released details on how it scales its cloud backend on Vercel, as of August 3, 2026. According to Vercel's report, Factory serves tens of millions of daily API requests while maintaining a sub-350ms p95 response time, all without a dedicated infrastructure team.
Background & Causes
Founded in 2023 by Matan Grinberg and Eno Reyes, Factory aims to automate repetitive tasks in software development using autonomous agents called Droids. As their user base expanded from individual developers to enterprises with tens of thousands of engineers, Factory's backend faced immense scaling pressure. Instead of migrating to complex traditional infrastructure, the company chose to scale directly on Vercel to keep its engineering team lean and agile.
Technical Analysis & Technology
A notable aspect of Factory's architecture is using Next.js as a full-stack cloud backend for its SaaS product, rather than just a frontend framework. Next.js API routes handle customer-facing traffic, while middleware manages authentication and routing logic across all surfaces. Webhook handlers and log drain pipelines run concurrently, streaming analytics downstream.
To optimize performance, Factory focused on tuning memory limits and function duration. According to Vercel, their Fluid Compute feature keeps functions warm across requests, eliminating the "cold start" penalty that traditionally plagues serverless setups. This keeps the p95 latency at 350ms or below, ensuring reliable performance for enterprise customers.
Expert Opinions & Insights
Factory's engineering team noted that deploying their own Droids internally freed up significant developer resources. Previously, non-technical teams had to request analytics dashboards or custom demos from engineers. Now, Droids programmatically interact with Vercel's APIs to build and deploy these internal tools autonomously, with engineers only setting guardrails and reviewing major decisions.
Impact & Future
Transitioning to a self-serve model exposed Factory's backend to automated bot traffic and fraudulent signups. However, by using Vercel Observability to track traffic patterns and the Web Application Firewall (WAF) to enforce route-level rules, Factory successfully mitigated these threats with minimal security overhead. Moving forward, this continuous automation loop will likely set a new standard for lean software development operations globally.