The tools that builders use to build.
“A 20% productivity gain across 30 million developers would unlock roughly $600 billion in latent economic output -- the GDP of Sweden.”
Thirty million developers generate approximately $3 trillion in annual economic value, according to a16z estimates. The AI-native shift is rebuilding the entire development workflow. In 2025, 41% of code was AI-generated. Goldman Sachs projects AI agents will expand the software market by 20% or more by 2030. Every layer of the toolchain has an incumbent optimized for the pre-AI world and a challenger built for the new one.
The velocity of change in developer tools is unprecedented. Cursor reached $1.2 billion in annual recurring revenue with 1,100% year-over-year growth. Replit's valuation tripled from $3 billion to $9 billion. Google acquired Windsurf for $2.4 billion along with Cognition. These are not incremental improvements to existing tools. They are category-creating products that change how software is built.
We invest in the infrastructure layer of the developer stack: the IDEs, deployment platforms, observability tools, and security systems that every developer touches every day. The application layer of software will be built faster and by more people. The tools they use to build it will capture the margin.
01
Cursor reached $1.2 billion ARR with 1,100% year-over-year growth and a $29.3 billion valuation. Windsurf was acquired by Google for $2.4 billion. GitHub Copilot has 20 million users and $2 billion in ARR. The IDE is being rebuilt from first principles around AI, and the market is repricing what a developer tool company can be worth.
02
85% of developers now use AI tools. Replit's valuation moved from $3 billion to $9 billion on the strength of Agent 3, which can sustain 200 minutes of autonomous development. The trajectory is clear: from code suggestion to code generation to autonomous software development. Each step expands the addressable market.
03
Vercel reached $9.3 billion valuation with $200 million ARR. Railway, Render, and Fly.io are competing to abstract away infrastructure complexity. Humanitec is building Internal Developer Platforms. The trend is clear: developers want to ship code, not manage infrastructure. The platforms that make deployment invisible will own the deployment layer.
04
Google acquired Snyk for $32 billion, the largest security acquisition in history. Gartner projects $213 billion in security spend. AI-generated code is increasing the attack surface. The security layer is moving from production monitoring to development-time prevention. Snyk Evo introduces agentic security that operates within the development workflow.
The center of gravity for developer productivity. GitHub Copilot holds 42% market share with $2 billion ARR and 20 million users. Cursor's $1.2 billion ARR and $29.3 billion valuation demonstrate the market's willingness to pay for AI-native development environments. Replit's trajectory from $3 billion to $9 billion reflects the shift toward autonomous development agents.
Vercel at $9.3 billion and $200 million ARR leads the deployment platform category. Railway, Render, and Fly.io compete on developer experience and pricing. Humanitec builds Internal Developer Platforms for enterprises. The common thread: make infrastructure invisible so developers can focus on application logic.
Datadog serves 30,500+ customers and has defined the category. The Grafana LGTM stack provides the open-source alternative. Sentry monitors errors for 4 million developers across 140,000 organizations. OpenTelemetry is becoming the standard instrumentation layer. As systems grow more complex, observability becomes more essential.
GitLab generates roughly $500 million ARR. GitHub Actions has become the default CI layer. AI-native CI is emerging as a category: automated test generation, intelligent PR review, and predictive build optimization. The continuous integration pipeline is the choke point between code and production.
The most underdeveloped sub-vertical in developer tools, which a16z identifies as a key unsolved problem. As AI generates more code, the need for automated testing grows proportionally. Test generation, visual regression testing, and AI-powered QA represent significant greenfield opportunities.
Snyk's $32 billion exit to Google validates the category. Semgrep, Socket, Aikido, and Endor Labs compete in the static analysis and supply chain security space. Snyk Evo introduces agentic security that operates autonomously within the development workflow. Security is becoming a development-time concern, not a post-deployment audit.
“Developer love is the most reliable leading indicator in software.”
Organic pull with paid individual accounts before enterprise budgets. The best developer tools spread bottom-up through individual adoption, not top-down through procurement.
IDE, deployment gate, or incident response loop. The tools that sit at choke points in the development workflow have pricing power and retention that tools at the periphery cannot match.
Built from first principles for AI-assisted development, not AI bolted onto a legacy codebase. The architectural difference determines the ceiling for what the product can become.
GitHub stars, Discord activity, and developer survey mentions are 12-18 month leading indicators of commercial demand. Community momentum predicts enterprise adoption with remarkable consistency.
Usage-based and compute-linked pricing models that align with how AI tools create value. Per-seat pricing from the SaaS era does not capture the economics of AI-augmented development.
Team-level value proposition without killing individual simplicity. The best developer tools maintain the simplicity that drives adoption while adding the controls and visibility that enterprises require.
Building developer
tools?