At a recent AI conference in San Francisco, over 300 founders and investors were asked a provocative question: which billion-dollar AI startup would you bet against? The answer was surprising. Perplexity AI topped the list, with OpenAI coming in second. While the OpenAI vote raised eyebrows given its market dominance, the Perplexity verdict reveals something deeper about the AI search landscape in 2025.

The $20 Billion Question

Perplexity Founded in 2022, the company hit a $20 billion valuation by September 2025, processing 780 million queries monthly with over 30 million active users. Impressive on paper, but the company has raised nearly $1.5 billion in funding, with valuations jumping from $500 million to $20 billion in just 18 months. Fundraising rounds roughly every two months suggests either extraordinary growth or growing desperation to prove the business model works.

The Wrapper Problem

Here’s the uncomfortable truth: Perplexity is increasingly looking like what Silicon Valley dreads most a wrapper. The company initially had a competitive edge when it pioneered AI powered web search with real time information. But that advantage has evaporated faster than anyone expected.

ChatGPT now has comprehensive web search capabilities, available to all users since February 2025. OpenAI didn’t stop there they launched ChatGPT Atlas, a full browser with integrated AI search, directly competing with Perplexity’s own Comet browser.

The Distribution Moat That Perplexity Doesn’t Have

Distribution is everything in tech, and this is where Perplexity’s vulnerability becomes existential. OpenAI has hundreds of millions of users who already trust ChatGPT for everything from coding to creative writing. Adding web search wasn’t building a new product; it was enhancing an existing habit.

Google’s advantage is even more crushing. With billions of daily users, Chrome integration, and deep Android ties, Google can push AI search capabilities to more people in a day than Perplexity might reach in a year.

Perplexity, meanwhile, is stuck acquiring users one at a time and burning investor capital. Their attempt to buy Google Chrome for $34.5 billion a bid that went nowhere revealed just how desperate they are for distribution leverage.

The India Problem: Manufacturing Metrics

In July 2025, Perplexity announced a partnership with Bharti Airtel to offer free Perplexity Pro subscriptions to all 360 million Airtel customers for one year. The Pro subscription, normally ₹17,000 annually, would be free through the Airtel Thanks app. On paper, this seemed like a distribution masterstroke.

The reality? It’s a textbook case of vanity metrics disguising fundamental problems.

India is one of the world’s most price sensitive markets. Netflix slashed its basic plan by 60% from ₹499 to ₹149 just to compete. Amazon Prime raised its fee from ₹179 to ₹299 and faced significant pushback. Even global giants struggle with conversions at these price points.

Now imagine Perplexity’s challenge: converting free users who got the service bundled with their telecom plan into paying subscribers at ₹17,000 per year. Even if 10% of Airtel customers claim the offer 36 million users the numbers look great in pitch decks. But when those free trials expire in 2026, how many will renew?

The harsh reality: when the free year ends, most users won’t convert. They’ll go back to ChatGPT (free), or Google (free). Perplexity will be left with vanity metrics that evaporate the moment payment is required.

The Model Advantage

Perhaps most critically, OpenAI and Google own their foundational models. OpenAI’s GPT-4o and o3, Google’s Gemini 2.5 these represent complete control over the technology stack, from training to inference to continuous improvement.

Perplexity builds on third party models. They’re always playing catch-up, always dependent on others’ innovation cycles, and always vulnerable to their suppliers becoming their competitors. When your core technology is rented, not owned, you’re not building a sustainable competitive advantage you’re building a temporary arbitrage opportunity that’s about to close.

My View

The Silicon Valley founders and investors who voted Perplexity as most likely to fail weren’t being cynical they were being realistic. The company proved the market existed but lacked the resources to defend it once the giants woke up.

Perplexity’s early lead has vanished. ChatGPT does web search just as well. Google has integrated AI into its core product. Both companies have deeper technology, wider distribution, stronger brands, and sustainable business models. Perplexity has a $20 billion valuation built on a crumbling foundation.

The India strategy reveals the core problem. When you’re giving away your premium product for free to hundreds of millions of users who historically won’t pay for similar services, you’re not building a business you’re manufacturing metrics for the next funding round. The Airtel partnership isn’t a distribution victory; it’s a Hail Mary from a company that knows its window is closing.

The AI bubble will eventually deflate. When it does, wrappers built on vanity metrics and unsustainable unit economics will be the first to go. Perplexity’s 360 million “free users” in India won’t save them when those users discover that ChatGPT and Google do the same thing for free and they don’t need to pay ₹17,000 for the privilege.