Aravind Srinivas, CEO of Perplexity Labs, announced on June 7, 2025, via X that since the launch of Perplexity Labs, AI can now draft 98–99% of an IPO prospectus—up from the previously reported 95% by David Solomon, CEO of Goldman Sachs
🚀 Behind the Accuracy Leap
- Goldman Sachs Benchmark: At a Cisco event, Solomon revealed AI can generate 95% of an S-1 IPO prospectus in minutes, compared to two weeks by a six-person team
- Perplexity’s Advancement: Srinivas claims that Perplexity Labs’ deep web browsing, code execution, chart generation, and self-supervised workflows have improved drafting precision to 98–99%
- Public Confirmation: He shared this breakthrough on social media: “After Perplexity Labs, I would say probably 98-99%.”
💡 What Perplexity Labs Includes
- Automated tools: Integrated browsing, code execution, and chart creation help structure prospectus content
- Key use case: Srinivas posted on LinkedIn that Labs can perform end-to-end IPO analysis and investment thesis—citing Circle’s IPO as a real-world example
⚠️ Limitations & Industry Response
- Human oversight remains vital: Reddit users emphasize that while AI handles the bulk of drafting, legal review and regulatory compliance still need expert scrutiny
- Productivity vs trust: Debates surfaced around whether AI reduces man-hours or simply shifts effort into review phases, with comments like: “Obviously, you’d have to review the output… then you just need two sets of eyes to verify instead of six.” reddit.com
🌍 Broader Implications
- Financial industry transformation: Investment banks, compliance teams, and corporate counsel may save significant time and costs, allocating more attention to value-driven tasks instead of boilerplate drafting .
- Regulatory frameworks under pressure: As automation improves, regulators will need to define standards for accuracy, accountability, and disclosure in AI-generated financial filings.
- Competitive leap: With a 99% accuracy rate, Perplexity gains an advantage in efficiency and reliability over competitor AI tools.
📌 Key Takeaways
| Metric | Details |
|---|---|
| AI drafting accuracy | 98–99% with Perplexity Labs (up from 95%) |
| Benchmark comparison | AI completes 95% S-1 in minutes vs weeks by humans |
| Support beyond drafting | Includes web, code, visualization, and analysis tools |
| Human review necessity | Still essential for legal/regulatory validation |
| Industry impact | Big cost and time savings; expectations on regulation |
