Perplexity AI — Voted "Most Likely to Fail" and Facing Existential Legal Risk

• Sources: TSG Invest, Reuters, Reuters
• Verdict: keep
• Why: The industry already voted on this one.

The L-takes

The industry verdict. At industry events, Perplexity was voted "most likely to fail" by founders and investors. That's not a blog post take — that's the people who fund startups telling you they don't see it.

The valuation. ~$20B private valuation after multiple rounds, $1.5B+ raised from Accel, SoftBank, Nvidia, Bezos, IVP. Valuations ballooned from hundreds of millions to tens of billions in under 2 years. Classic startup hypergrowth pattern that doesn't necessarily reflect profits or unit economics.

The legal wall. NYT, Chicago Tribune, and others have sued over unauthorized use and summarization of content. Perplexity's defense — claiming it doesn't have "volition" — is viewed as a long shot. A federal judge also issued a preliminary injunction blocking Perplexity's AI from accessing Amazon accounts, calling the behavior unauthorized. These aren't noise — they could curb core product features.

The margins. Infrastructure cost per query vs. revenue per query suggests margins are thin or negative, especially on free users. Revenue sharing with publishers is a cost, not a moat. Reddit users report marketing spend and feature development have slowed — classic cash crunch signals.

The competition. Google (Gemini/AI search), OpenAI (ChatGPT + search), Microsoft, Anthropic — all with cash reserves, distribution channels, and user bases orders of magnitude larger. Unless Perplexity finds a real niche or huge product differentiation, it's easy to see being squeezed out or commoditized.

The verdict: "A classic VC hype bubble with real tech but weak defensibility."