Startups Use AI to Write Most Production Code

Business Insider reports a rapid shift toward AI-generated code among startups, based on a survey of more than two dozen founders and venture capitalists. The article quotes Rami Alhamad, cofounder and CEO of Alma, saying, "Nearly everything we ship now is AI-generated." Business Insider also quotes Volodymyr Giginiak, CTO of Wordsmith AI, as saying the company's code is "nearly 100%" AI-generated. Dan Lorenc, cofounder and CEO of Chainguard, told Business Insider that his team now produces 100% of its code via Claude Code, up from 60% a year earlier. Business Insider reports founders' common concern that the speed of generation introduces "slop" and low-quality outputs.

Editorial analysis: Industry observers note the practical enablers behind this shift are improved base models, richer tool-call integrations, and tighter prompt/harness workflows that let developers generate end-to-end code faster than before. Practitioners undertaking similar high-volume code generation typically confront three technical risks: increased technical debt from autogenerated scaffolding, brittle or incorrect dependency and API usage that passes cursory tests, and a greater surface for supply-chain and licensing issues when third-party snippets are incorporated.

Editorial analysis: Business Insider reports venture capital interest in AI coding infrastructure and startups, citing firms such as Lovable, Replit, and Cursor as beneficiaries of that funding trend. For practitioners, the rise of near-total AI code authorship shifts priorities from raw implementation speed toward systems for validation: automated testing, provenance tracking, and dependency vetting become higher-leverage investments at scale. This pattern also raises security and maintenance questions that engineering and security teams will need to operationalize.

For observers and engineers: monitor adoption of LLM-integrated CI/CD controls, emergence of provenance and reproducibility tools for generated code, spikes in supply-chain or licensing incidents tied to autogenerated snippets, and whether vendor ecosystems (models, IDE plugins, testing tools) standardize interfaces for safe code generation. Business Insider's reporting highlights adoption and concern; forthcoming indicators will be tooling adoption and measurable incidents tied to generated code quality.