One of these was vibe coding platform Emergent, based in India and the US, which raised $70 million in a funding round led by Khosla Ventures and SoftBank Vision Fund 2 on 20 January. Barely four months ago it raised a $23 million Series A funding from investors such as Lightspeed Venture Partners, Prosus, Together Fund and Y Combinator.
However, difficult questions loom just beyond this hype: Is the excitement among investors and founders justified? Will vibe coding fundamentally change how code is written and startups are built? What are the limitations and potential pitfalls of vibe coding? Mint explains.
What is vibe coding?
At its simplest, vibe coding is an AI-driven approach to software development that prioritizes high-level intent over technical syntax. It shifts the focus from how to write a program to what the program should actually do.
Instead of requiring years of training in specific programming languages, a ‘vibe coder’ uses natural language to describe a desired product or feature. An AI agent then translates that description into a functional codebase, handling everything from the user interface to the backend logic.
Fundamentally, vibe coding lowers the bar for quickly creating new software. The AI is able to simplify workflows, generate working code and let users tweak things on the fly, or at least that’s what vibe coding startups say their technology can do.
The core promise of vibe coding is to lower the barrier to entry for software creation. By simplifying complex workflows and generating working code instantly, it allows users to iterate and ‘vibe-check’ their ideas in real time.
Vibe coding aims to lower the barrier to software creation by using AI to generate functional code from high-level inputs, allowing users to iterate ideas in real time.
Why is venture capital excited about vibe coding?
Many venture capital firms believe vibe coding represents a shift in how tech startups will be built. While most people use vibe coding platforms for small projects such as building simple apps for themselves, these investors are betting that some will find ways to harness vibe coding to create software and business models that haven’t existed before.
While investors aren’t expecting vibe coding to be a magic pill that turns a layman into a coding expert, so do expect it to bring in fundamental changes. A startup’s competitive advantage or ‘moat’ has traditionally been its complex proprietary code. Today, some VCs assume code is a commodity. They are betting on founders who can use vibe coding to pivot in hours rather than months, increasing their chances of finding ‘product-market fit’ before their funding runs out.
Where does India stand?
Vibe coding is taking off in India, albeit slowly. While Emergent has raised the most funds so far, it isn’t the only one making moves in the space.
Rocket, formerly known as DhiWise, raised $15 million in a seed round last September, co-led by Accel and Salesforce Ventures. It had previously raised $9.5 million from Accel and India Quotient.
Vibe coding agentic AI company Composio raised $25 million in July 2025 in a round led by Lightspeed, with participation from Elevation Capital and Together Fund.
There were smaller cheques written as well. CodeKarma received $2.5 million from Prosus Ventures and Accel, while Stellaris Venture Partners wrote a $2.7 million cheque to Drizz, and Titan Capital backed Haizen with $500,000.
India’s broader startup ecosystem has grown rapidly over the past few years. According to a KPMG report from 2024, it’s the third largest globally, with more than 100 unicorns (companies valued at or more than $1 billion). The number of startups in India has increased dramatically, from 500 in 2016 to more than 159,000 as of January 2025, according to the Press Information Bureau.
While the country does have a large and rapidly growing startup ecosystem, it often takes cues from the US, which has the most frenzied and aggressive startup culture in the world. To that end, several tech trends born in the US make their way to India with a 9- to 24-month lag. However, advances in artificial intelligence are narrowing this gap, experts say.
What’s going on globally?
The US’s vibe coding market is on a completely different level, primarily because of deeper pockets and quicker adoption of new technologies. There are already vibe coding unicorns in the US.
Replit was the first of these. The company started out by offering a development environment supporting various programming languages that could be accessed through a web browser. It became a unicorn in 2023, and in 2024 it launched the popular Replit Agent, an autonomous AI developer that converts natural language descriptions into fully functional, production-ready applications. Unlike traditional coding assistants that only suggest snippets of text, it manages the entire development lifecycle, including setting up the server, configuring databases, writing both frontend and backend code, and handling one-click deployment.
Foundation model company Anthropic, which became a unicorn in 2023, launched its own coding agent Claude Code in 2025. It is widely considered the most versatile and useful such platform by developers worldwide.
Cursor and Lovable are two of the fastest-growing vibe coding startups in the world. Cursor, a US startup, was valued at $23.3 billion after a massive $2.3 billion funding round last November, in which investors such Coatue, Nvidia and Google joined the company’s cap table alongside Accel, Thrive, and Andreessen Horowitz. This was barely six months after its $900 million funding in June from Thrive, Accel, Andressen Horowitz and DST, which valued the company at $9.9 billion.
Lovable, a vibe coding startup based in Sweden, raised $330 million at a $6.6 billion valuation in December 2025, in a round led by CapitalG and Menlo Ventures’ Anthology fund.
What are the limitations and potential pitfalls of vibe coding?
Vibe coding excels for prosumers (individuals who are both producers and consumers) and for small and medium businesses, where software requirements are relatively streamlined. It also allowed enterprise teams to independently build custom internal tools. However, its primary challenge, and greatest market opportunity, lies in integrating with complex, high-scale corporate environments that demand rigorous security and architectural standards.
That isn’t to say that the companies listed above have no enterprise ambitions. Cursor, Replit, Lovable and even Claude have enterprise plans and sales teams that can work with large businesses to ensure seamless deployment.
Another issue with vibe coding is that it can lead to what one Gartner analyst described as ‘vibe debt’. This is similar to technical debt, in which a developer chooses a solution with a short turnaround time but faces problems in the long run such as maintenance issues, random bugs, broken workflow logic, and in extreme cases, outages. “Generating code is easy; validating that it is secure, accurate and maintainable is hard,” said Ashish Banerjee, senior principal analyst at Gartner.
Also, as a system scales, the security of AI-generated code tends to drop. A paper published on arXiv, a curated research-sharing platform, showed that code modified up to 40 times by OpenAI’s GPT-4o contained errors. Researchers found that after just five iterations by the large language model, the code had 37% more critical vulnerabilities than the initial version.
