Perplexity plans IPO in 2028 as Anthropic, OpenAI prepare listings

Perplexity plans to go public in 2028, regardless of how the market receives Anthropic and OpenAI’s listings, CEO Aravind Srinivas told CNBC.

“We were planning something in 2028, regardless of these two companies, so that’s still the case,” Srinivas said in an interview published Tuesday. he said.

Srinivas had previously said that the company has no plans to go public before 2028. His latest comments suggest a more concrete timeline.

The CEO’s comments come after Claude developer Anthropic confidentially filed for an IPO last week. While there are no details on share pricing, Anthropic was last valued at around $1 trillion. OpenAI also confidentially filed for an initial public offering on Monday.

These listings, along with SpaceX this week, will be among the largest in history and will test investors’ appetite for these mega IPOs.

“I think there will definitely be ripple effects if they don’t do well, just like there’s no sugar coating on this. SpaceX’s IPO this week will definitely be a leading indicator of how Anthropic or OpenAI will come out,” Srinivas told CNBC.

“I think it’s important for the AI ​​industry that these IPOs do well, and I actually think they’ll do well because they’re doing well.”

The valuations of Anthropic and OpenAI, known as frontier laboratories because their models are among the world’s leading, are under the scrutiny of investors.

Srinivas said both companies deserve high valuations because “they are on the edge”. A slowdown in the pace of innovation could negatively impact valuations, he said, adding that there is currently no sign of that happening.

“If you don’t see any improvement in model capacity from either of these two companies for six months, that’s a problem for them,” Srinivas said.

Corporate spending on AI has become a major focus after OpenAI CEO Sam Altman said on a company livestream that companies are now discussing how much they spend on AI. Altman said AI costs are a “big problem.”

One emerging trend is “tokenmaxxing,” where employees increase the use of AI to signal productivity. “But people don’t just want tokenmax, they really want to use the best model for that task,” Srinivas said.

Perplexity’s product is based on models from several companies. AI will find the best model to use for a specific task, taking into account cost when desired.

“If there is an open source model that gets the job done 90 percent of the time, if it is 10 to 20 times cheaper than the edge model, I would probably use that,” Srinivas explained. “The future is still great for border intelligence, but there will be no mindless spending like we have seen over the last few months.”