In July 2025, Bharti Airtel gave all 360 million of its customers a free year of Perplexity Pro — India's first generative AI telecom partnership, and the single most effective customer acquisition move in the company's history. India's Perplexity user base grew 640% year-on-year within a quarter. India became Perplexity's largest market in the world. Then, in January 2026, as the free year quietly expired, Airtel did not renew. It signed with Adobe instead. Perplexity has not explained why. This letter covers both halves of the story: the genuinely extraordinary growth — a $9 billion company, then $18 billion, then $20 billion in under two years, a $34.5 billion bid to buy Chrome from Google, a federal court fight with Amazon over what an AI agent is allowed to do with your account — and the quieter, harder question of what happens to a growth strategy built on free distribution once the free year ends.
Not investment advice. Data sourced from Perplexity AI official disclosures, Wikipedia company record (continuously updated), Airtel official announcements July 2025, Whalesbook February 2026, TechTimes June 2026, No Hacks Agentic Browser Landscape March 2026, DemandSage Statistics February 2026. All figures current as of June 2026.
On July 17, 2025, Bharti Airtel — one of India's two dominant telecom operators, serving roughly 360 million prepaid, postpaid, Wi-Fi, and DTH customers — announced something no telecom operator anywhere in the world had done before: a free, twelve-month Perplexity Pro subscription, worth ₹17,000 (roughly $200) annually, bundled into every customer's existing plan at no additional cost. This was India's first generative AI partnership between a telecom operator and an AI company, and it triggered the single fastest growth spurt in Perplexity's history. India's Perplexity user base grew 640% year-on-year in the quarter immediately following the launch. App downloads jumped 600% year-on-year, reaching 2.8 million in a single quarter. By late 2025, India had become Perplexity's largest market by traffic anywhere in the world — ahead of the United States, the company's home market.
The mechanism behind this growth deserves to be stated plainly, because it is genuinely instructive: Perplexity did not buy this growth through advertising. It bought distribution at hardware and infrastructure scale — the same playbook the company used separately with Samsung, bundling free Pro subscriptions into every 2025 Samsung TV sold globally, and again with Snapchat's nearly one billion monthly active users in January 2026. Nearly three-quarters of Airtel's entire 360-million customer base reportedly activated the free Perplexity offer — an adoption rate that would be considered extraordinary for any consumer technology promotion in any market, in any country, at any point in the industry's history.
The Airtel offer's claim window ran from July 17, 2025 to January 17, 2026. When the twelve months ended, Airtel did not renew or extend the Perplexity partnership. Instead, the telecom operator pivoted to a new partnership with Adobe — despite the Perplexity offer's documented, extraordinary engagement during its active year. No joint statement, press release, or public explanation has been issued by either company describing why the partnership ended rather than continued or evolved into a paid or revenue-shared arrangement. Perplexity's own public communications made no comment on the transition, and the company's leadership was notably absent from at least one major industry event where the question might reasonably have been addressed.
Independent analysis has offered plausible, though unconfirmed, explanations: the cost of providing premium AI access free to a customer base as large as 360 million people at scale represents a genuinely significant ongoing financial commitment for a company of Perplexity's size, and Adobe's more conventional, established subscription-software revenue model may have offered Airtel a more predictable, more easily monetisable commercial framework going forward than a continued AI research-tool giveaway. Whether the decision originated from Airtel's own strategic recalculation, from commercial terms Perplexity was unwilling or unable to renew at scale, or from some combination of both, remains genuinely unknown from public information available as of this letter's publication.
From India's largest generative AI partnership to its quiet, unexplained end
January 2024 to June 2026 · Company disclosures, TechCrunch, Forbes
Perplexity's most audacious single move came on August 12, 2025, when the company submitted an unsolicited $34.5 billion bid to buy Google Chrome — timed deliberately to coincide with active US antitrust litigation in which a federal judge was weighing whether to force Google to divest the browser entirely. Perplexity explicitly framed its bid as a remedy to that litigation. The bid was not accepted, but it generated enormous attention for a company with, at the time, a fraction of Chrome's market value in total funding raised — and it foreshadowed the company's subsequent strategy with its own browser, Comet, launched in July 2025 and made free worldwide in October 2025, backed by a fresh $200 million raise in June 2026 explicitly earmarked to win what the company's own framing describes as "the surface where an AI agent starts a task, and increasingly finishes a purchase, on your behalf."
Comet is the centre of Perplexity's entire current strategy, and the logic is explicit in the company's own public statements: owning the browser means owning the moment an AI agent begins a task and, increasingly, completes a purchase on a user's behalf. Built on Chromium, Comet integrates Perplexity's search and reasoning directly into browsing — summarising pages, comparing options across tabs, filling forms, and executing multi-step tasks autonomously. The free-worldwide pivot in October 2025, followed by full cross-platform rollout completing with iOS in March 2026, mirrors the distribution-over-advertising playbook that made the Airtel partnership so effective domestically in India.
A direct, structural response to the mounting publisher lawsuits examined below: a $5-a-month tier that creates a dedicated revenue pool, with Perplexity retaining only 20% and routing the remaining 80% to participating publishers — including CNN, Condé Nast, The Washington Post, Fortune, Time, and Der Spiegel — based on direct traffic, AI citation, and agent-driven content use. "Perplexity only succeeds if journalism succeeds," the company's head of publisher partnerships told Digiday — a notably different posture from several AI labs currently facing similar litigation without an equivalent revenue-sharing response in place.
Independent of the consumer growth narrative, Perplexity's enterprise push — Comet Enterprise (launched March 2026, with customers including AWS, Fortune, AlixPartners, and Gunderson Dettmer), Perplexity Computer, and a three-year, $750 million GPU capacity commitment with Microsoft Azure signed in January 2026 — represents the company's most durable long-term revenue foundation. Operating with roughly 90 to 100 employees against revenue estimates approaching $450-650 million annualised, Perplexity's revenue-per-employee ratio is among the highest of any company in the AI sector — a genuine structural efficiency advantage distinct from, and arguably more durable than, the headline-grabbing consumer distribution deals.
Perplexity's most consequential legal fight is not a publisher copyright dispute but a structural question about agentic AI itself. Amazon sued Perplexity in November 2025, and in March 2026 won a preliminary federal injunction blocking Comet from accessing password-protected Amazon accounts — Senior US District Judge Maxine Chesney found Amazon likely to succeed on claims that Comet violated the Computer Fraud and Abuse Act, drawing a sharp distinction between a user granting Comet permission to act and Amazon itself authorising that access. Perplexity has appealed, framing Comet in its 96-page brief as functionally analogous to Safari — the user's own tool, never directly connecting to Amazon's servers. The Ninth Circuit stayed the injunction pending appeal, with oral arguments held June 11, 2026. The eventual ruling will set the first significant federal precedent on whether buyer-side AI agents can access third-party platforms at a user's direction against the platform owner's own terms of service — a question with consequences reaching far beyond Perplexity or Amazon specifically.
Copyright infringement; also alleges hallucinated quotes harmed brand reputation.
Alleges unauthorised scraping of NYT content for AI training and outputs.
Demands content deletion and compensation for alleged unauthorised scraping.
Three major Japanese publishers allege "free-riding" use of 120,000+ articles.
Alleges unauthorised scraping of 17,000+ articles after revenue-share talks collapsed.
The structurally most significant case — defines what agentic browsers may legally do.
The Airtel partnership's quiet, unexplained ending is the single most important unresolved question in this entire letter, and it deserves more weight than the headline growth figures it produced. A 640% surge in user growth driven by a free, subsidised product giveaway demonstrates genuine product appeal once people try it, but it does not, on its own, demonstrate durable monetisable demand once the free period ends — and the fact that Airtel, having seen extraordinary engagement from three-quarters of its customer base, chose not to renew or extend the arrangement in any visible form is a data point a more cautious reading of this letter would weigh heavily. Whether India's Perplexity user base and revenue contribution have held up since January 2026, declined sharply, or settled at some sustainable smaller level is not publicly disclosed information as of this letter's publication — a genuine gap in the available evidence that this letter cannot responsibly paper over.
Perplexity's distribution-partnership strategy — Airtel, Samsung, Snapchat, more than 25 telecom operators globally — has been extraordinarily effective at generating user growth at near-zero customer acquisition cost, but the Airtel outcome raises a genuine structural question about whether that strategy converts reliably into durable, monetised relationships once the free period inevitably ends. The Samsung TV integration and the Snapchat partnership (reportedly worth roughly $400 million over its term, according to industry reporting) both appear structurally different from the Airtel arrangement — built around longer-term, more clearly defined commercial terms from inception, rather than a fixed twelve-month free giveaway. Whether the Airtel outcome reflects something specific to that one partnership, or a more general weakness in the free-distribution-first model when it transitions toward monetisation, is a question this letter's available evidence cannot fully resolve.
The Amazon CFAA litigation carries risk that extends well beyond Perplexity's own balance sheet, and a ruling against the company could meaningfully constrain Comet's core functionality regardless of the browser's user growth or revenue trajectory. An adverse Ninth Circuit ruling would not merely cost Perplexity a legal fee — it would establish that agentic browsers cannot access password-protected third-party accounts without the platform owner's explicit authorisation, a constraint that would apply to every competing agentic browser examined in this letter series, including OpenAI's Atlas, but that would hit Perplexity hardest given how central agentic account access is to Comet's core value proposition of completing tasks (bookings, purchases, form-filling) on a user's behalf.
Long-horizon thinking on capital, technology, and the forces shaping the next decade of wealth creation. Written from first principles. Not consensus. Not noise.