OpenAI and Google push free AI plans in India to harvest users and data

The India “freebies war” reveals how central user behavior and data have become to the next phase of the AI race. OpenAI making ChatGPT Go free only in India, Google bundling Gemini AI Pro with 500 million Jio customers, and Perplexity comping Pro via Airtel are not just growth hacks—they’re bets that multilingual, code‑switched, mobile‑first usage will generate training signals Western datasets lack. India’s 730 million smartphones and ultra‑cheap data make it a uniquely dense testbed for stress‑testing models against messy, real‑world language use.

Strategically, this tilts the race toward players with the balance sheet and telecom partnerships to subsidize large‑scale usage in high‑growth markets. It’s also an early illustration of how competition for AGI‑relevant data will play out: not just by crawling the open web harder, but by shaping how billions of people work, learn and communicate through closed platforms.

That raises uncomfortable questions about consent and bargaining power. Even with opt‑outs, the asymmetry between global AI labs and price‑sensitive users is stark. But from a purely technical lens, the more diverse and high‑volume the interactions, the faster models will improve at handling code‑mixed queries, non‑Western knowledge, and resource‑constrained devices—capabilities central to any credible claim of “general” intelligence.