news

New Intelligence Report

Recently, the market for artificial intelligence search engines has continued to "heat up."

On the one hand, Google, the search leader with a market share of over 90%, is integrating Gemini into its search engine.

Faced with fierce market competition, Perplexity does not seem to have lost a large number of users under the pressure from giants. Instead, it has ushered in a new round of surge in user usage.

Three months before ChatGPT launched, former OpenAI research scientist Aravind Srinivas founded Perplexity, an AI search engine that uses information extracted from the web (including news sites) in real time.

As a startup trying to compete with larger, earlier-started rivals, Perplexity is about to complete a $250 million investment round from investors including SoftBank Vision Fund 2, increasing its valuation to $3 billion from $1 billion in April.

Existing investors include Nvidia and Amazon founder Jeff Bezos, as well as AI industry heavyweights such as Andrej Karpathy and Yann LeCun.

In addition to investor optimism, Perplexity's monthly revenue and usage have also skyrocketed sevenfold since the beginning of this year.

According to company insiders, Perplexity's annualized revenue was $5 million at the beginning of this year (annualized revenue refers to the full-year revenue calculated based on the sales in the most recent month), but current revenue expectations have exceeded $35 million.

According to statistics, the Perplexity search engine answered about 250 million questions last month, while the total query volume for the whole year of 2023 was only 500 million. The KPI for the first half of last year was completed in just one month.

The surge highlights that Perplexity is the fastest-growing generative AI application after ChatGPT, despite some controversy surrounding Perplexity’s data collection technology.

Perplexity’s growth also points to a growing preference for search engines powered by artificial intelligence — whether from Google, OpenAI or Perplexity.

Integrating artificial intelligence into Internet search will not only revolutionize the way people search, but may also lead to changes in users’ search habits.

Aiming to become a leader in the search field, the "development system" Perplexity faces huge challenges.

Despite its rapid growth, Perplexity still lags far behind market leader Google.

As a search engine giant, Google has dominated the market for many years, monopolizing more than 90% of the global market share and processing about 8.5 billion queries every day.

In addition, Google has huge financial resources and massive amounts of data, and can continuously improve its AI search capabilities.

After overcoming its first-to-market issues, Google’s multimodal Gemini model has topped benchmarks.

However, in the view of Perplexity CEO Aravind Srinivas, the purpose of Perplexity is not to replace Google, but to do things that Google disdains.

And, as an "answer engine," as their homepage slogan says, Perplexity is "the starting point of knowledge."

Perplexity doesn't want to be a new Google, but rather to change the way we find answers on the Internet.

Perplexity’s core goal and positioning is to satisfy users’ curiosity and allow users to get the answers they want.

Users don't care if Perplexity has the most powerful model. They just care about getting good answers.

“At the end of the day, small players in search have two advantages: speed and focus,” said Dmitry Shevelenko, chief business officer at Perplexity.

“The Perplexity team only thinks about one thing: how to get users’ questions answered quickly. The fierce competition makes us focus on this even more.”

Having firmly established its positioning and goals, Perplexity was not intimidated by the fierce competition from technology giants.

OpenAI's business is very diverse and it did not start out as a search engine, so OpenAI does not focus on answering user search questions through high-quality news sources.

Perplexity focuses on the search track, which is why initial feedback from SearchGPT shows that it does not have an advantageous position compared to Perplexity.

So far, Perplexity's revenue mainly comes from consumer and business subscriptions. Perplexity recently announced on its official website that it will introduce advertising on its platform by the end of next month.

Shevelenko said Perplexity shares a “double-digit” percentage of revenue with news publishers for each sponsored article, and that it has signed deals with companies such as Time, Der Spiegel and Fortune.

As part of the program, publishers will also receive access to the Perplexity API, which can be used to create customized Answer Engine and Enterprise Pro accounts.

In addition, Perplexity will provide all employees of the above publishers with one year of access to the Enterprise Professional Edition product, which has enhanced data privacy and security features.

In the two weeks since the partnership was launched, 50 publishers have requested to join the program, and Perplexity hopes to include as wide a range of sites as possible.

However, before announcing its recent partnership with publishers, Perplexity was accused of plagiarism in June by media outlets including Forbes and Wired.

The companies criticized Perplexity for copying content without clearly citing the original link and scraping information on websites that explicitly block crawlers.

Perplexity subsequently made changes to its user interface to make citations more prominent and took steps to ensure that responsive interface content was not a hodgepodge of information scraped from other sites.

In order for Perplexity to become a competitive search engine, it is necessary to maintain its operation with a good business model.

From a long-term perspective, revenue sharing is a more effective way than a one-time payment, based on Perplexity's positioning and company specifics. However, OpenAI adopted a one-time payment solution.

In addition, Perplexity is different from Google and OpenAI in one aspect: Perplexity does not build its own large AI model.

Instead, it licenses AI systems from companies like OpenAI.

Like many potential Google competitors, Perplexity’s search engine was initially powered by a licensed version of Microsoft’s Bing web index but has since moved away from using Bing as a core system.

Although it uses technical support from various engines, Perplexity always has its own proprietary search index and ranking system.

Earlier this year, a Perplexity employee said that compared with traditional search engines such as Google, Perplexity has more professional and reliable information sources and is more suitable for the search needs of the news and academic circles.

Using unclear spam information for training can only produce a large amount of spam information, which has always been a problem that has plagued most companies. Therefore, more data sources need to be used when training models.

However, some people have suggested that introducing advertisements may scare off users, causing them to doubt whether the search environment and search results are trustworthy, and that advertisements will make web pages look less professional and reliable.