Google and Accel picked 5 Indian startups from 4,000 applications — and none of them is an AI wrapper

Google and Accel reviewed more than four thousand applications from India-linked startups during intake for the Atoms accelerator, and came away with a sobering conclusion: about 70% of applicants were building what the industry calls AI wrappers. Five teams made the final cut, and none of them simply exploits other companies' models without adding value of its own.

Over the past two years, "wrapper" has become a fixture of venture capital vocabulary. The term describes startups that take a large language model's API and build an interface on top of it with minimal in-house development. These products are easy to copy and have no technological moat, and the slightest change in the provider's terms can put the entire business at risk. That is why investors increasingly turn wrappers down, even when their early-stage revenue looks convincing.

Atoms is a joint initiative of Google and the venture firm Accel, aimed at India-focused startups. This year the program drew a record number of applications. The selection team reviewed products, talked to founders, and assessed whether a polished pitch deck was backed by real technical differentiation. The result was telling: seven out of ten teams were building products that were nearly indistinguishable from one another.

The five finalists are described as startups with genuine technology at their core. The organizers plan to disclose details on each company as part of the program. Still, the fact that the selectors are openly talking about wrapper culture in India's AI sector says a lot about the market's maturity.

India is now one of the most active markets by number of AI startup launches. Cheap engineering talent, a large domestic market, and the growing presence of tech giants make it fertile ground for experimentation. But the same set of factors also fuels a wave of clones: it is easier to take an off-the-shelf model, package it into a product for local needs, and try to raise money than to invest in fundamental R&D. Accelerators in the US and Europe report a similar picture, but in India it stands out because of the sheer volume of applications.

For Google, Atoms is about more than supporting the ecosystem. The company is looking for partners who will build on its infrastructure: Gemini API, Google Cloud, Vertex AI. Startups with real technological depth are potential long-term customers and partners, not one-off API users. Accel, for its part, has been investing in the Indian market for more than twenty years and wants to shape the next generation of companies with global potential.

Picking five startups out of four thousand sends a signal to the market: the era of easy money for wrapper projects is ending. Institutional investors, including the largest tech companies, are drawing an ever sharper line between real AI companies and companies that simply rent someone else's intelligence. For founders, the takeaway is simple: if you can't explain your product without referring to the underlying model, you'll have a serious positioning problem in your next round.