Many synthetic intelligence startup concepts are nonetheless little greater than superficial “wrappers” constructed on prime of current fashions. However because the AI mannequin makers add extra options, traders are cautious of startups that would turn into so simply pointless.
Living proof: when reviewing extra 4,000 purposes for the joint AI accelerator for India startups run by Google and enterprise agency Accel, “wrapper” concepts dominated. However none of them have been among the many 5 startups for the most recent cohort, Accel companion Prayank Swaroop advised TechCrunch (pictured above).
Introduced in November, the AI-focused Atoms program by Google and Accel goals to again early-stage startups constructing AI merchandise linked to India. Startups chosen for the most recent cohort will obtain as much as $2 million in funding from Accel and Google’s AI Futures Fund, together with as much as $350,000 in cloud and AI compute credit from Google, the companies stated.
Roughly 70% of the rejected purposes have been “wrappers” — startups that layered AI options equivalent to chatbots on prime of current software program however “weren’t reimagining new workflows utilizing AI,” Swaroop stated.
Lots of the remaining purposes that have been denied, Swaroop stated, fell into crowded classes equivalent to advertising automation and AI recruitment instruments, areas the place traders noticed little novelty. Startups in these sectors usually wrestle to distinguish themselves, he stated.
This isn’t, maybe, shocking. This yr’s program obtained almost 4 occasions the purposes than earlier Accel’s Atoms cohorts — with many first-time founders.
India’s rising AI ecosystem stays largely centered on enterprise purposes and Swaroop stated the purposes mirrored that. About 62% of the submissions centered on productiveness instruments and one other 13% on software program growth and coding, that means round three-quarters of the purposes have been enterprise software program concepts quite than client merchandise. (Swaroop had hoped to see extra concepts for healthcare and training.)
Techcrunch occasion
San Francisco, CA
|
October 13-15, 2026
Jonathan Silber, co-founder and director of Google’s AI Futures Fund, stated the 5 startups chosen aligned carefully with areas the place Google expects AI to see deeper real-world adoption.
This system doesn’t require startups to make use of Google’s fashions solely, Silber stated, noting that many firms mix a number of fashions relying on the workflow. The purpose, he stated, is to assemble suggestions from startups on how Google’s fashions carry out in real-world purposes.
Insights from these startups can then be fed again to Google DeepMind groups to assist enhance future fashions, creating what Silber described as a “flywheel” between startup experimentation and AI growth. “If an organization is utilizing another mannequin, meaning Google has work to do to construct one of the best mannequin out there,” he advised TechCrunch.
This yr’s startups chosen are:
- K-Dense, which is constructing an AI “co-scientist” to speed up analysis in fields equivalent to life sciences and chemistry;
- Dodge.ai, which develops autonomous brokers for enterprise ERP programs;
- Persistence Labs, which focuses on voice AI for name centre operations;
- Zingroll, which is constructing a platform for AI-generated movies and exhibits;
- Level Plane, which applies AI to industrial automation in automotive and aerospace manufacturing.
