After nine years of grinding, Replit finally found its market. Can it keep it? (techcrunch.com)

After a decade-and-a-half of fits and starts, Replit announced a $250 million round that nearly tripled its valuation to about $3 billion, on the back of a dramatic revenue jump from roughly $2.8M to an annualized >$150M in under a year. The company, which once hovered at the same modest ARR and cut half its staff to survive, credits that turnaround to Replit Agent — an agent-based coding assistant that can write, debug, deploy and provision databases — and a bold repositioning away from professional dev tooling toward enabling “knowledge workers” with no formal technical background. The move has won enterprise customers (Zillow, Duolingo, Coinbase), reported gross-margin-positive economics (Masad cites 80–90% margins on enterprise deals) and a #3 spot on a16z’s AI spending list. Technically and strategically, Replit’s edge is twofold: large-scale cloud development infrastructure (multiplayer editing, deployment and DB management) and productized agents that act like software-engineering partners. A high-profile safety failure (an agent wiped a production DB) prompted a rapid fix that segments practice vs. production environments — an example of hardening safety as a moat. Risks remain: foundation-model owners like OpenAI and Anthropic can subsidize competing tools and tune models end-to-end. Replit’s playbook now is to scale, invest in product/vertical agents, and use its cash cushion and infrastructure differentiators to defend against commoditization.