AI search engine startup Perplexity AI has seen its monthly revenue increase sevenfold over the past two quarters, and the number of monthly hits to the system has grown to 250 million, up from 500 million in all of 2023. Search engine Perplexity has become one of the fastest-growing generative AI applications since the launch of OpenAI’s ChatGPT in November 2022, despite allegations from copyright holders of copyright infringement in data collection.
Image source: freepik.com
San Francisco-based Perplexity AI, founded by former Google intern Aravind Srinivas just three months before the launch of ChatGPT, based its eponymous search engine on artificial intelligence that generates real-time responses to user queries based on information from the Internet. . Perplexity did not develop its own model, but went the route of licensing a combination of AI systems from several manufacturers, such as OpenAI and Google.
Aravind Srinivas / Image source: Noah Berger (Associated Press)
Initially, the Perplexity search engine used a licensed version of the Microsoft Bing index, but then abandoned it in favor of its own index and its own ranking system. “We use signals from all types of search engines, but we have our own crawler and ranking system,” says Perplexity business director Dmitry Shevelenko.
The company has raised $250 million in investment this year, tripling its valuation from $1 billion in April to $3 billion today. Perplexity’s investors include Nvidia, Amazon founder Jeff Bezos, OpenAI co-founder Andrej Karpathy and Meta✴ chief AI developer Yann LeCun.
Perplexity’s growth comes as it integrates AI capabilities into Google’s search engine and launches a prototype of OpenAI’s SearchGPT search AI tool. “At the end of the day, being a smaller player in this space has two advantages: speed and focus,” says Shevelenko. “Our users and team have one thing on their mind when it comes to Perplexity: a place where you get answers to your questions.” Competition sharpens our attention even more.”
At the start of the year, based on last month’s sales extrapolation, Perplexity’s annual revenue was estimated at $5 million, but the revenue forecast has now increased sevenfold to $35 million. The company has now begun to transition from a subscription business model to one that generates revenue from paid advertising. That puts Perplexity in closer competition with Google, which dominates the $300 billion advertising industry.
The platform targets journalism and academia as sources of fairly reliable information and data. Perplexity sees this source material as an advantage over traditional search engines, which index a much wider range of sites. Some analysts believe that the advent could have a negative impact on user growth due to an “untrusted environment” and skepticism about the impartiality of search results.
«For Perplexity to become a useful product on the open Internet, there must be good business models for publishing new and updated facts about the world, Shevelenko is sure. “If you want to align incentives [with journalism] over the long term, revenue sharing is a more efficient way to do it than the one-time lump sum payments that OpenAI went with.”
The company will share a “double-digit” percentage of its revenue with each news publisher partner, Shevelenko said. He said that such agreements have already been concluded with Time, Der Spiegel and Fortune, and in total more than 50 publishers have expressed a desire to join the Perplexity revenue sharing program within two weeks of its launch.
Dmitry Shevelenko / Image source Ringo Chiu (EPA-EFE)
It is worth noting that Forbes and Wired previously accused Perplexity of violating intellectual rights, “reproducing stories without clear attribution, and copying sites that clearly blocked its search robots.” The company took the criticism into account – it made changes to the user interface to make quotes more visible and changed the format of aggregated responses.
Many experts believe that the AI search market is heating up. In their view, general web search as a whole is becoming redundant due to new ways of matching users with information, products and services. The risk is whether these new methods will be reliable enough for mass use. “AI is stubbornly prone to confabulation,” says analyst Joseph Teasdale. “At the scale of billions of requests per day, major disruptions are inevitable.”