AI Prediction Index

Aravind Srinivas

Co-founder and CEO, Perplexity AI

@AravSrinivas

17%
success rate · #104 of 144
0correct1half right2missed4on the clockFull leaderboard
“It might not be inconceivable that Micron, the supplier of HBMs, might be more valuable than Meta in the next 6 to 12 months. Whatever is the bottleneck will command the price.”

Co-founder and CEO, Perplexity AI

Micron briefly overtook Meta's market cap on June 25, 2026 after a blowout fiscal Q3 (revenue up 346% year over year), but the crossover did not hold: as of September 2026 Meta sits near $1.6–1.7T versus roughly $1.0–1.1T for Micron. The 6–12 month window runs to mid-June 2027.

“We have all the funding we've raised, and our revenue is only growing — we have no plans of IPOing before 2028.”

Co-founder and CEO of Perplexity AI

Said Mar 30, 2025Deadline Dec 31, 2027TechCrunch, March 30, 2025

Perplexity remains private; in June 2026 Srinivas and CBO Dmitry Shevelenko reaffirmed 2028 as the earliest IPO date.

“Agnostic of these two companies [Anthropic and OpenAI], we were planning for something in 2028, so that still remains the case.”

Co-founder and CEO of Perplexity AI

Said Jun 9, 2026Deadline Dec 31, 2028CNBC interview, June 9, 2026

Made after Anthropic's confidential IPO filing and reports of an OpenAI listing; Perplexity reported roughly $500M+ ARR in mid-2026 and has not filed.

“Power, not GPUs or chips, is the binding constraint on AI infrastructure — and it will remain the bottleneck for at least the next three-plus years, with roughly 40% of planned US data center builds blocked by public resistance.”

Co-founder and CEO of Perplexity AI

Said Jun 15, 2026Deadline Jun 15, 2029The Twenty Minute VC (20VC) with Harry Stebbings, June 15, 2026

Srinivas also flagged memory and CPUs as the next pinch points in the same interview, and said Rubin-generation data centers reach full capacity in 2027.

CapabilitiesPartially correct
“It's not going to work this year. It's still… you know, anyone who's saying agents will work in 2025 should be skeptical of it.”

Co-founder and CEO of Perplexity AI

He was largely right about general-purpose consumer agents: agentic browsers (including Perplexity's own Comet) shipped in 2025 but proved unreliable on complex, multi-step tasks and vulnerable to prompt-injection, and retrospectives described agents as overhyped or 'mostly a dud.' But he understated coding agents, which worked well enough commercially to reach billions in run-rate revenue by early 2026.

CapabilitiesIncorrect
“Next week, we will start giving access to a small set of initial users. We are trying our best to get it out in a few weeks. [On the release of Perplexity's agentic browser, Comet]”

Co-founder and CEO of Perplexity AI

Said Mar 21, 2025Deadline Apr 30, 2025X post by @AravSrinivas, March 21, 2025

Comet did not ship 'in a few weeks': Perplexity launched it on July 9, 2025 — nearly four months later — and even then only to $200/month Max subscribers plus a limited set of waitlist invitees, with broader invite-only rollout continuing for months afterward.

“On AI browsers taking over the work of recruiters and executive assistants: "I'm pretty sure [that within] six months to a year from now, it can do the entire thing" — "a recruiter's work worth one week is just one prompt: sourcing and reach outs," plus tracking responses, updating spreadsheets, resolving calendar conflicts, scheduling interviews and pushing a brief ahead of the meeting.”

Co-founder and CEO of Perplexity AI

By July 2026 no AI browser agent, Comet included, could execute an end-to-end recruiting or executive-assistant workflow from a single prompt; agentic browsers still require human supervision on long-horizon, multi-app tasks, and recruiter and administrative-assistant roles have not been eliminated.