The Economic Survey specifically notes India's constraints in compute infrastructure and funding, reinforcing a strategy focused on smaller or alternative AI approaches.

Vembu points to the USD 50-100 billion cost range, rising energy demands, and GPU shortages as reasons to avoid a direct chase of large models.

He argues that the investment and energy needs for big LLMs, along with GPU constraints, make emulation impractical for India and suggests alternative directions instead.

The discussion comes ahead of the India AI Impact Summit, touted as the largest global AI gathering to date.

The summit is expected to draw attendees from over 100 countries, with more than 500 leaders and 500 startups, and significant participation from tech giants and government figures.

Zoho is actively pursuing smaller models and energy-efficient AI strategies, prioritizing research and development over large-scale, energy-intensive expansion.

He suggests staying slightly behind the global giants in this race can be advantageous and emphasizes building from India’s strengths in research rather than emulation.

The event features high-profile attendees and speakers from NVIDIA, Anthropic, Microsoft, Google DeepMind, and Accenture.

International figures such as Jensen Huang, Dario Amodei, Brad Smith, Demis Hassabis, and Julie Sweet are anticipated to attend.

The stance mirrors the Economic Survey’s view that limited access to cutting-edge compute and capital makes large foundational models impractical for India, advocating a bottom-up AI development approach.

Zoho founder argues India should avoid racing head-to-head with massive LLMs and instead invest in smaller, energy- and cost-efficient AI models or alternative approaches, leveraging India’s brain power for R&D.

Pre-event registration already exceeds 35,000, signaling strong international interest ahead of the summit.