AI (Artificial Intelligence) letters and robot hand are placed on computer motherboard in this illustration taken, June 23, 2023. REUTERS/Dado Ruvic/Illustration/ File Photo
(Reuters) - AI software developer platform Replit said on Wednesday it has raised $250 million in a funding round that values the San Francisco-based company at $3 billion, as investors continue to back AI startups that help people write code.
Valuations for code-generation startups, or "code-gen," have climbed as companies look to use AI to assist software engineers or help non-technical teams to write their own software.
Replit was valued at just over $1 billion in 2023 after it raised $97.4 million. The company said its annualized revenue rose to $150 million from $2.8 million in less than a year.
The latest financing was led by Prysm Capital and includes Google's AI Futures Fund and Amex Ventures as strategic investors. Existing backers including Andreessen Horowitz — also known as a16z — and Coatue increased their investments.
The startup plans to use the funding to invest in research and development as well as sales and marketing.
In the crowded field of code-gen startups, Replit said its product is differentiated because it builds vibe-coding tools for non-technical users in enterprises. Cognition, another startup focusing on building AI coders, raised over $400 million at a $10.2 billion valuation earlier this week.
"There's a lot of competition for software engineers, but we're seeing that people who are able to use Replit are from every part of the enterprise, from sales, HR, to operations, and that has helped them with product development cycles," said Replit CEO Amjad Masad.
Companies including Duolingo and Zillow use Replit to build applications. Replit on Wednesday launched Agent 3, an autonomous tool that can test and fix code and build custom agents and workflows.
In May, Cursor, another San Francisco-based code-gen startup raised $900 million at a $10 billion valuation. The firm helps users by suggesting and completing lines of code and can write sections of code autonomously.
(Reporting by Juby Babu in Mexico City and Krystal Hu in San Francisco; Editing by Tasim Zahid and Alan Barona)
