Perplexity ran one of the boldest growth bets in AI: giving its premium tier away free to tens of millions of telecom customers in India. Now that the earliest free subscriptions are expiring, the numbers are starting to show whether a giveaway can actually convert into a lasting, paying business.

The arrangement was simple in concept and massive in scale: Perplexity teamed up with Airtel, one of India's largest telecom carriers, to hand every one of its hundreds of millions of subscribers a full year of premium access, a plan that would normally run about $200. New sign-ups for the free tier closed in mid-January, but because each activation locked in twelve months of access from that point, the earliest adopters only recently reached their renewal date — and had to actively cancel if they didn't want to start paying.

The launch effect was immediate and dramatic. App downloads in India jumped more than sixfold in the very first month compared to the month before, and that single month alone brought in more new installs than the company had accumulated across the entire first half of the year. The surge didn't fade quickly either — download volume over the full seven-month promotional window came in at roughly nine times what the app had generated over the previous seven months.

Once new users could no longer claim the free year, the download curve fell off a cliff — new installs in India dropped more than 90% compared to the run-up period. But the people who had already installed the app stuck around far longer than the download numbers suggested. Active usage did ease off its peak, yet even after that decline, the app was still pulling in roughly five times the monthly active users it had before the promotion ever existed.

The more surprising twist is what happened to spending. Even as new downloads dried up, in-app purchase and subscription revenue in India climbed by roughly 60% compared to the period when the free offer was still live. That pattern held up even as the earliest free subscribers began hitting their renewal dates: daily revenue in the weeks after those renewals started outpaced both the prior month and the first-half average for the year.

It's an encouraging signal that some users are willing to start paying once they've gotten used to having the premium tool for free — but it isn't proof of that on its own. Because the free plans were set to auto-renew by default, some of that revenue bump could simply reflect people who forgot to cancel rather than users who actively chose to become paying customers. Analysts note there's currently no clean way to separate those two groups from the data alone.

India has become the largest single market for generative AI app downloads worldwide, driven by its enormous population, over a billion internet users, and cheap mobile data. But turning that scale into revenue has historically been difficult, which is exactly why AI companies keep experimenting with subsidized access as a way to buy habit formation before asking users to pay.

This particular giveaway started well before similar moves from other major AI players, which means its users are among the very first large group in the country to hit the point where free premium access turns into a real payment decision. The early data suggests the strategy can leave a company with a meaningfully bigger, more monetizable footprint than it had before — but it still can't confirm how many of the giveaway's original recipients are the ones actually choosing to pay, versus simply not noticing the charge. That answer will only become clear as later waves of users hit their own renewal dates in the months ahead.