Google and Accel Atoms Open a Pre-Seed AI Cohort in India

What Google and Accel are actually offering

The Accel Atoms x AI Futures Fund cohort targets a specific stage: pre-seed founders who have not yet built out full companies. That's the earliest point at which a founder makes foundational technical bets, and it's where Google is choosing to intervene.

The package has four named components. Selected startups get early usage of Google DeepMind models — Gemini for text, Imagen for images, Veo for video. They get mentorship from Google and Accel engineering, product and go-to-market staff. They get Google Cloud credits with dedicated support. And they get the possibility of direct equity investment from both Accel and Google.

Applications run through the Accel India portal, and the cohort begins in February 2026. The source frames the goal plainly as kickstarting 'the next wave of AI innovation in India.'

Why the earliest stage is the point

Most model-access programs court companies that already have a product and a workload to migrate. This one deliberately reaches founders before those decisions are locked in. A pre-seed team that builds its first prototype on Gemini, Imagen and Veo tends to keep building on them.

The three named models map to three distinct application categories — conversational and reasoning apps, image generation, and video generation. Offering all three signals that Google isn't recruiting for one narrow use case; it's inviting founders across the generative stack to standardize on its stack early.

Pairing that access with cloud credits reinforces the same pull. A startup that develops on Google Cloud with dedicated support inherits an infrastructure default that persists well past the credit balance.

The equity and mentorship layer

The involvement of Accel — an established venture firm — alongside Google changes the incentive structure from a typical developer-credits promotion. Direct equity investment means both parties take financial positions in the companies, not just usage relationships.

Mentorship from go-to-market specialists at both organizations is notable given the pre-seed focus. Founders at this stage usually lack distribution help more than they lack technology, and GTM support addresses the gap that credits alone cannot fill.

What a February 2026 cohort tells builders in India

For applied AI teams in India, the concrete takeaway is timing and eligibility: the window is open now, the cohort is pre-seed, and it starts in February 2026. Founders who already intend to build on Gemini, Imagen or Veo lose little by applying and gain early access plus a shot at capital.

The broader signal is that Google is treating the pre-seed layer in India as a place worth seeding directly — combining model access, infrastructure, mentorship and money into a single funnel. Teams should weigh the value of that bundle against the standardization it encourages, since the program's design rewards building on Google's models from the first line of code.