AI insiders predict Perplexity and OpenAI are “most likely to fail”

• Investor sentiment shifts as legal scrutiny and financial headwinds challenge top AI startups.
• Anthropic now holds a 32% share—outpacing OpenAI’s 25%, according to Menlo Ventures.

At last week’s Cerebral Valley Summit in San Francisco, more than 300 AI founders and investors named Perplexity as the billion-dollar startup most likely to fail, with industry heavyweight OpenAI coming in a close second.

The results, gathered via an anonymous survey by independent journalist Alex Heath, mark a striking change in Silicon Valley’s view of artificial intelligence powerhouses.

Perplexity, long touted for its meteoric valuation—skyrocketing from $14 billion to nearly $50 billion in recent months—has become a focal point for industry scepticism. Analysts warn of a dot-com-era-style frenzy, and the company faces mounting legal turmoil.

Amazon sued Perplexity in November to block its Comet browser from making purchases on users’ behalf, while Reddit and several major Japanese newspapers—including Yomiuri Shimbun, Asahi Shimbun, and Nikkei—have filed copyright infringement claims. The BBC is also considering legal action over alleged unauthorised scraping.

Industry observers have criticised Perplexity’s web-scraping tactics, with Cloudflare’s CEO comparing its behaviour to North Korean hackers. Reports indicate the company routinely ignores website protocols designed to limit scraping activities.

When asked about the Summit’s survey results, Perplexity’s Jesse Dwyer dismissed them as the product of a “judgmental valley conference.”

Investor momentum shifts toward Anthropic

Perhaps most notable at the Summit was the shifting investor sentiment toward Anthropic, which attendees named the top pick for future investment over OpenAI. Despite a consensus that OpenAI will likely lead next year’s global LMArena leaderboard for advanced AI models, Anthropic’s growth has stolen the spotlight.

Anthropic’s revenue vaulted from $87 million at the start of 2024 to more than $5 billion by August 2025. The firm completed a $13 billion Series F round in September, bringing its valuation to $183 billion.

Within the crowded enterprise AI market, Anthropic now holds a 32 per cent share—outpacing OpenAI’s 25 per cent, according to Menlo Ventures. Its developer toolkit, Claude Code, launched in May and already generates over $500 million in annualised revenue, with business clients growing from fewer than 1,000 two years ago to over 300,000 today.

Financial woes shadow OpenAI

OpenAI’s second-place finish in the “most likely to fail” poll surprised some, given its market dominance. The pessimism, however, may be rooted in financial concerns. While OpenAI expects to generate $13 billion in revenue this year, projected losses could reach $9 billion.

Internal forecasts suggest operating losses may balloon to $74 billion by 2028, with the company’s gargantuan infrastructure spend projected to top $1 trillion over the next decade.

Moreover, OpenAI’s conversion challenges remain acute: 95 per cent of ChatGPT’s 800 million users use the service for free. Unlike Anthropic, which aims to break even by 2028, OpenAI is not expected to do so until the decade’s end.

Other rising startups flagged in the Summit’s “failure poll” included Cursor, Figure, Harvey, Mercor, Mistral, and Thinking Machines. Yet optimism endures among AI bulls: Participants predicted Nvidia could reach a $6 trillion market capitalisation by the end of 2026.