One email each morning — yesterday's AI, sortedGet it in your inbox

METAL LAB

Perplexity's India Downloads Drop 90% After Free Promo Ends, Revenue Up 60%

A year of free subscriptions distributed via Airtel partnership shrank the user base but boosted spending per user

스마트폰 화면에 Perplexity 앱 설치 페이지가 표시되어 있다

이미지: TechCrunch Startups

Summary

  • As Perplexity's 12-month free subscription deal with Indian telecom carrier Airtel, launched in July 2025, winds down, downloads have plunged sharply while in-app revenue has actually risen
  • According to Sensor Tower data, downloads during the promotion period were nine times higher than in the preceding seven months, but fell more than 90% after the promotion ended; monthly active users (MAU) dropped 37% from their peak but remained about five times higher than before the promotion
  • Similar free-distribution strategies targeting the Indian market are continuing, including OpenAI's ChatGPT Go and Google's partnership with Jio, making Perplexity's case an early test for the industry
제휴 시작
2025년 7월, 인도 2위 통신사 에어텔(가입자 3억6000만명)
무료 혜택
페럴렉시티 프로 12개월 무료 구독(정가 약 200달러)
신규 신청 마감
2026년 1월 16일
프로모션 기간 다운로드
7개월간 5600만 건(직전 7개월의 9배, 센서타워)
종료 후 다운로드 감소
2026년 2~7월 330만 건, 직전 6개월 대비 90% 이상 감소
MAU 변화
2025년 10월 2200만명 정점 → 2026년 7월 1400만명(37% 감소했지만 프로모션 이전 평균의 5배 이상)
매출 변화
2026년 2월~8월 중순 인앱 매출, 프로모션 제공 기간 대비 약 60% 증가(센서타워)
월 순모바일매출(Appfigures)
2025년 1월 약 3만4000달러 → 2025년 12월 약 7만달러 → 2026년 7월 약 15만6000달러

The number of people downloading the Perplexity app in India fell by more than 90% over the past year. Yet during the same period, revenue earned in India actually rose by nearly 60%. These figures come from data provided to TechCrunch by market research firm Sensor Tower.

이미지: TechCrunch Startups

What happened with the 12-month free subscription deal with Airtel

In July 2025, Perplexity partnered with Airtel, India's second-largest telecom carrier, to offer its Perplexity Pro subscription—normally worth about $200 a year—free for 12 months to Airtel's 360 million subscribers. New sign-ups closed on January 16, 2026, but each subscription itself ran for one year from the date it was activated. As a result, the free period for the earliest sign-ups began expiring in sequence starting last month, July 2026. Users who didn't want to continue had to cancel their subscription before the automatic renewal billing date.

이미지: TechCrunch Startups

Downloads went on a rollercoaster ride

The reaction to the promotion's announcement was immediate. According to Sensor Tower, Perplexity recorded 5.9 million downloads in India in July 2025 alone—a 625% increase from the previous month, and more than the cumulative downloads for the first six months of that year combined (5.4 million). The surge didn't stop after the first month. Over the seven months during which new sign-ups were open, cumulative downloads reached 56 million, more than nine times the total of the preceding seven months. MAU also more than doubled to 8.9 million in July 2025, then surged further to 22 million by October of that year.

Daily figures compiled by app analytics firm Appfigures show the swing even more starkly. In the week before the offer launched, Perplexity's average daily downloads in India were about 11,200. In the first week after launch, that jumped twentyfold to about 223,000, and between mid-September and mid-October it climbed to an average of about 305,000 per day. Downloads of ChatGPT and Claude showed no significant change during the same period, which Appfigures co-founder and CEO Ariel Michaeli said was a deliberate comparison "to confirm this wasn't a broader surge in demand for AI apps."

But the trend reversed sharply once new sign-ups closed on January 16. From February to July, India downloads totaled 3.3 million, down more than 90% from the preceding six months. However, the users already acquired didn't leave as quickly. July MAU stood at about 14 million, down 37% from the October peak, but still more than five times the average for the first half of 2025 (about 2.6 million).

PeriodDownloadsMAU
Pre-offer (H1 2025, 6 months)5.4 millionAverage 2.6 million
Offer period (Jul 2025–Jan 2026, 7 months)56 millionPeak 22 million (October)
Post-offer (Feb–Jul 2026)3.3 million14 million (July)

Downloads fell, but revenue rose

The more notable point is revenue. Sensor Tower estimated that Perplexity's India in-app purchase and subscription revenue from February through mid-August rose about 60% compared with the promotion period when new sign-ups were still open—even as downloads were declining. From July 18 to August 12, when the earliest sign-ups' free periods began expiring, average daily in-app revenue was 9% higher than in the preceding 30 days, and 27% higher than the average for the first half of 2026.

Sensor Tower senior insights analyst Abe Yousef said that "given the nature of the promotion, some decline in usage after it ends is natural, but actual usage held up robustly."

Monthly net mobile revenue figures compiled by Appfigures show a similar trend. Revenue, which was about $34,000 in January 2025, grew to about $70,000 by December of that year, and reached about $156,000 by July 2026. Cumulative revenue from January to July 2026 totaled about $878,000, 16% higher than all of 2025's revenue.

PeriodMonthly net mobile revenue (Appfigures)
January 2025About $34,000
December 2025About $70,000
July 2026About $156,000

However, neither Sensor Tower nor Appfigures conclude that this revenue growth came entirely from Airtel subscribers voluntarily converting to paid plans. Because the free subscription was set to auto-renew, some users may have been billed inadvertently by failing to cancel before the renewal date. Michaeli also noted that the heavy exposure generated by the promotion may have attracted new paying users unrelated to the original free offer.

A shared bet on the Indian market

Perplexity's experiment is part of a broader trend. In August 2025, OpenAI made its lower-cost ChatGPT Go plan free for a year in India, and Google struck a partnership with Reliance Jio to offer its AI Pro subscription free for 18 months to Jio users. Because Perplexity's Airtel deal launched a few months before both of these, it stands as an early test case for whether paid conversion actually follows free distribution.

OpenAI is also considering a similar move separately in the U.S. market.

India, with more than a billion internet users and relatively cheap mobile data, is considered the world's largest market by generative AI app downloads, but it also has a persistent reputation as a market that is difficult to convert into revenue. With these results, Perplexity has gained some evidence that it can retain a meaningful share of both users and revenue even after free distribution ends, but it remains unconfirmed whether the same pattern will hold for subsequent cohorts of subscribers as their renewal dates arrive. Perplexity did not respond to TechCrunch's request for comment.

Editor's View

What makes this experiment interesting is that the numbers point in opposite directions. Downloads collapsed once the free offer ended, but revenue actually rose afterward. Conventional wisdom holds that both users and revenue decline together once a free promotion ends, but in Perplexity's case, the rise in payments stood out more than the drop in users. That could mean the increase isn't necessarily a sign the service is well-liked—it may simply be an artifact of the auto-renewal billing design itself. Both Sensor Tower and Appfigures flagged this possibility with some caution.

Based on experience watching this kind of growth experiment play out, the download curves generated by free distribution are always dramatic—and fade just as quickly. The real test lies in how many users remain after that curve disappears, and whether they actually open their wallets. Perplexity's July MAU of 14 million, though down 37% from its peak, is still more than five times what it was before the promotion—a sign that at least some of the "habit of using it for free" hasn't entirely disappeared.

For teams in Korea running AI services or considering telecom partnerships, two points are worth noting. First, the billing design at the end of a free period—auto-renewal versus opt-in—can dramatically skew conversion numbers on their own. Second, download figures alone shouldn't be used to judge a campaign's success. What actually needs to be tracked is retention and continued payment over the months following the promotion's end.

In the coming months, the free periods for OpenAI's ChatGPT Go and Google's Jio partnership will also expire in sequence. If the pattern Perplexity experienced first—a sharp drop in downloads alongside modest revenue growth—plays out similarly for its two competitors, the "free distribution followed by paid conversion" strategy in the Indian market could well become a standardized playbook.