
Rocket.new Raises $15M Seed to Scale AI Vibe-Coding Platform
Rocket.new raises $15M in a seed round led by Salesforce Ventures and Accel to expand its AI vibe-coding platform globally.
TL;DR
Rocket.new, a Surat-based AI “vibe-coding” startup, announced on September 23, 2025 that it raised $15M seed funding led by Salesforce Ventures, Accel, and Together Fund. Launched in June, it already has 400K users, 10K paid subscribers, and $4.5M ARR. The funds will drive R&D, global expansion, and a push toward making AI-built apps production-ready, with ambitious revenue goals of $60–70M ARR by mid-2026. The founders aim to transform the platform into a full agentic system supporting the entire product development lifecycle
Indian start up Rocket.new announced it had raised $15 million in a seed funding round, a striking milestone for a company that only launched its beta platform three months earlier in June. The official announcement came via press release on September 23, and within hours the news was covered widely across tech and start up media, reflecting both investor excitement and the growing global attention on the AI-powered coding space.
The seed round was led by Salesforce Ventures, joined by Accel and Together Fund. Structured as an all-equity deal, the raise signals strong investor confidence in Rocket.new’s model and early growth. In just a few months, the start up has gained more than 400,000 users in 180 countries, secured over 10,000 paid subscribers, and built an annual recurring revenue stream of $4.5 million. Few beta-stage companies manage to convert enthusiasm into tangible revenue so quickly, making Rocket.new’s traction unusual.
Founded in Surat, a city better known for its diamond industry than its start up ecosystem, Rocket.new represents an unexpected origin story in India’s tech landscape. The team is led by CEO Vishal Virani, with co-founders Deepak Dhanak and Rahul Shingala. Before launching Rocket.new, the trio built DhiWise, a developer workflow platform whose data and learnings now serve as part of Rocket.new’s foundation. That experience gave them an advantage in understanding both the frustrations of developers and the technical requirements of building a scalable tool.
Virani explained the company’s philosophy in a recent interview. “We are building the first vibe solution platform, which is not just solving the problem of day one, but what we are focusing on is solving the problem of day two,” he said. The comment highlights a core difference between Rocket.new and its competitors. While others might enable users to spin up quick demos in minutes, Rocket.new wants to ensure that the applications created can be maintained, iterated upon, and deployed at scale.
To achieve this, the platform integrates large language models from OpenAI, Anthropic, and Google Gemini with its own proprietary systems trained on datasets drawn partly from DhiWise. The result is a slower but more complete generation process. Rocket.new estimates that its first applications take about 25 minutes to produce, compared to three minutes from rival platforms, but the trade-off is higher quality and production readiness. For users trying to build businesses rather than just test concepts, this difference could be decisive.