Many AI startup ideas are still just superficial “shells” built on existing models. But as AI model makers add more features, investors are wary of startups that could so easily become irrelevant.
Case in point: When reviewing more than 4,000 applications for the joint AI accelerator for Indian startups run by Google and venture capital firm Accel, “wrapper” ideas dominated. But none of them were among the five startups in the latest cohort, Accel partner Prayank Swaroop (pictured above) told TechCrunch.
Announcement This November, Google and Accel’s AI-focused Atoms program aims to support early-stage startups building India-related AI products. Startups selected for the latest cohort will receive up to $2 million in funding from Accel and Google AI Futures Fundas well as up to $350,000 in cloud computing and AI credits from Google, the companies said.
About 70% of rejected applications were “wrappers” — startups that layered AI features such as chatbots on top of existing software but “were not reimagining new workflows using AI,” Swaroop said.
Most of the remaining applications that were declined, Swaroop said, fell into crowded categories such as marketing automation and AI recruiting tools, areas where investors saw little novelty. Startups in these sectors often struggle to differentiate themselves, he said.
This is perhaps not surprising. This year’s program received nearly four times as many applications as Accel’s previous Atoms cohorts – with many new founders.
India’s growing AI ecosystem remains largely focused on enterprise applications and Swaroop said the applications reflect that. About 62% of submissions were for productivity tools and 13% were for software development and coding, meaning about three-quarters of the applications were enterprise software ideas rather than consumer products. (Swaroop had hoped to see more ideas in health care and education.)
Techcrunch event
San Francisco, California | October 13-15, 2026
Jonathan Silber, co-founder and director of Google’s AI Futures Fund, said the five selected startups closely match the areas where Google expects AI to see greater adoption in the real world.
The program does not require startups to use Google’s templates exclusively, Silber said, noting that many companies combine multiple templates depending on workflow. The goal, he said, is to gather feedback from startups on how Google’s models work in real-world applications.
Information from these startups can then be fed back to Google DeepMind teams to help improve future models, creating what Silber described as a “flywheel” between startup experimentation and AI development. “If a company uses an alternative model, that means Google has work to do to build the best model on the market,” he told TechCrunch.
The startups selected this year are:
- which builds an AI “co-scientist” to accelerate research in areas such as life sciences and chemistry;
- which develops autonomous agents for enterprise ERP systems;
- Persistence Labswhich focuses on voice AI for call center operations;
- Zingrollwhich is building a platform for AI-generated movies and shows;
- Level planwhich applies AI to industrial automation in automotive and aerospace manufacturing.
Jagmeet covers startups, technology policy updates and all other major technology developments in India for TechCrunch. He previously worked as a senior correspondent at NDTV.
You can contact or check Jagmeet’s outreach by sending an email mail@journalistjagmeet.com.