a16z's text-to-web-app market map highlights Bolt at $20M ARR and Lovable at $10M in two months

Summary

Andreessen Horowitz released a market map on text-to-web-app generation that surfaces two standout growth numbers: Bolt.new at $20M ARR and Lovable at $10M ARR in just two months of monetization. The map, authored by a16z partner Justine Moore, catalogs a sprawling category of tools that let users build websites and web applications from natural language prompts—a space experiencing what the firm describes as an "explosion of products."

The core mechanism is straightforward: most of these tools use LLMs to generate code based on prompts, run that code through middleware logic to handle file tracking and API calls, then push the result to a browser execution environment that streams the display back to the user. But capability has hard limits. The tools excel at simple builds and feel "like magic" to non-technical users, but they struggle with integrations, accumulate bugs, and can generate code that becomes unwieldy as projects scale.

The market map frames customer segments along three primary vectors: consumers building one-off personalized apps (bedtime story creators, niche tools), developers using these tools to accelerate prototyping and shipping, and consultants and agencies—traditionally reliant on Squarespace and Wix—now using AI generation to serve SMBs and solopreneurs faster. Choice of tool depends on friction points: whether a user wants to code or only type, whether they're building a website or web application, and how much design control they need. That segmentation reflects a deeper reality: these tools are not replacing developers building production systems; they're lowering the floor for everything below that.

The incumbent problem runs deep. Webflow, a once-hot no-code platform, does not appear on the map—likely because it has not yet launched a text-interface product. Legacy website builders like Squarespace and Wix face a branding challenge: they can add AI generation features, but they cannot easily reposition from "website builder" to "AI-first." Newcomers like Bolt.new and Lovable own that narrative from launch. GoDaddy's experience is instructive—the company aggressively surfaces AI website generation to domain buyers, but the user experience feels bolted-on rather than native.

The distribution problem shadows the category. Instantiation is now trivial: spin up an idea, generate a site in hours, launch it live without coding. But discovery is broken. Without a mechanism to surface these creations—the way StumbleUpon once surfaced random web pages—most of these sites languish unseen. The bedtime story creator mentioned in the market map is novel, but few will ever find it. Monetization and distribution, not product, are the harder problems ahead.