In the fast-changing world of artificial intelligence, not many startups have caught people’s attention and investors’ interest as quickly as Perplexity AI. This company, which is based in San Francisco and was founded in 2022, is a serious competitor to traditional search engines like Google. It has changed the way people use the internet to find information. Perplexity AI‘s value skyrocketed to $18 billion by July 2025. This meteoric rise shows how innovative its approach is, how strategic its vision is, and how much people want AI-powered solutions. This article goes into detail about the startup’s journey, its cutting-edge technology, and the plans that helped it become a unicorn in just three years.
Confusion The idea behind AI came from a simple but deep realization: traditional search engines, which were based on Google’s link-based model, weren’t giving users the direct, accurate answers they wanted. Aravind Srinivas, Denis Yarats, Johnny Ho, and Andy Konwinski, all veterans of AI and tech, started the company to connect scattered information with synthesized, conversational answers. Each founder had a different point of view. Srinivas, the CEO, had worked on AI at OpenAI and Google DeepMind; Yarats, the CTO, had worked on AI research at Meta; Ho, the Chief Strategy Officer, had worked on engineering at Quora; and Konwinski, the co-founder of Databricks, had business experience.
The team found a major problem in the search world: users were spending an average of 11 minutes per session looking through a lot of sources to find answers to hard questions, often having to deal with results that were full of ads and SEO-optimized content that put algorithms ahead of quality. Perplexity had a clear idea: if they could give direct, accurate answers with clear sources, they could make a new type of search engine called the “answer engine.” This vision would help them make decisions about their products and their overall strategy, which would set them apart from the big search engines.
Confusion AI’s main breakthrough is its conversational search engine, which combines advanced large language models (LLMs) with real-time web access to give you accurate answers with citations. Perplexity is different from regular search engines because it combines information from many sources and gives short answers in a conversational style. This method fits with the growing use of conversational AI, which has been driven by the success of models like ChatGPT. However, there is one big difference: it is open and honest. Perplexity’s answers are based on sources that can be checked, and they always use five sources to build trust and lower the chance of AI “hallucinations.”
The company’s technical architecture is an amazing piece of engineering. Perplexity lets users choose the best model for their needs by combining LLMs like GPT-4, Claude, and Mistral with proprietary models. This improves the quality and scope of research. Web scraping in real time makes sure that answers are up to date, and a two-step process of web search and LLM synthesis checks that the information is correct from many sources. This has led to some great numbers: 80% of users say the information is correct, and the platform has an 85% retention rate after the first use, with an average session length of 11 minutes. This is better than Wikipedia and Google for some uses.
The way people use Perplexity is also new. It feels more like a research assistant than a search tool because its conversational interface lets you ask follow-up questions and keep track of what you’ve already searched for. File upload capabilities and natural language processing for complex queries are just two of the features that have helped the company grow by 39% per month and 275% in monthly active users (MAU) from Q3 2023 to Q1 2024. Perplexity said that by April 2025, it had 30 million users around the world and handled over 780 million queries in May alone, with a growth rate of more than 20% each month.
Perplexity’s quick rise is based on a series of smart funding rounds and a clear way to make money. In 2022, the company got $2 million in a pre-seed round. In March 2023, it got $25.6 million in a Series A round, and in January 2024, it got $73.6 million in a Series B round. This made the company worth $520 million. By March 2024, Perplexity was worth $1.04 billion, making it a “unicorn.” By June, that number had tripled to $3 billion, and by December, it had grown to $9 billion. In 2025, Accel led a $500 million round that raised the company’s value to $14 billion. In July, an extra $100 million raised the value to $18 billion. Prominent investors like Jeff Bezos, NVIDIA, SoftBank, and Yann LeCun have helped this growth, showing that they believe Perplexity has the potential to shake up the search market.
The company’s business model is a mix of free and paid services. The free tier lets you make a small number of queries, but the Pro plan for $20 a month lets you make as many queries as you want, use advanced AI models, and upload files. The Enterprise plan costs $40 per seat per month and is for teams that need better security and analytics. API access makes it easier for developers to connect their apps. This tiered structure has led to amazing growth in revenue, from $7 million in annual recurring revenue (ARR) in 2023 to $63 million by the end of 2024. By July 2025, it is expected to reach $150 million. Perplexity also added contextual advertising and e-commerce integrations, and premium ad space cost $50 CPM.
Perplexity has worked with publishers instead of trying to control all of the ad revenue like Google has been accused of doing for a long time. Its revenue-sharing model makes sure that content creators get paid when their work is used, which has led to partnerships with more than 300 publishers, such as Fortune, Time, and The Washington Post. This model not only makes it easier to find content, but it also gives publishers analytics tools to keep track of how well their content is doing, making it a win-win for everyone. Aravind Srinivas, the CEO, has said that this way of doing things fits with Jeff Bezos’s idea of “your margins, my opportunity.” This makes Perplexity a fairer choice than regular search engines.
Perplexity is also different because it is dedicated to ethical AI. The company solves the “hallucination” problem that other AI models have by putting a high value on openness and being able to be checked. Srinivas has spoken out about this, saying, “People made fun of us for saying that in AI, hallucination is a feature, not a bug.” Perplexity’s focus on sourced answers has set a standard for reliable AI search, and other companies are following suit.
Perplexity released Comet, an AI-powered web browser, in July 2025. Its goal was to combine search, task automation, and summarization into one interface. Srinivas calls Comet a “cognitive operating system.” Its goal is to combine navigation, information, and activity, making the browser the best place for AI to fight. Comet’s capabilities have grown thanks to the acquisition of Invisible HQ, a stealth-mode startup founded by former Uber and CloudKitchens employees. It now has multi-agent orchestration and scalable automation.
Perplexity’s strategic partnerships, like giving Bharti Airtel’s mobile users in India a free Pro subscription for 12 months, show that the company is focused on expanding around the world. The company has also made deals to have its software pre-installed on Motorola phones and is in talks with Samsung to replace Google’s Gemini assistant. This shows that it wants to be the default AI experience on mobile devices.
Even though it has been successful, Perplexity is up against tough competition from tech giants like Google, OpenAI, and Microsoft, all of which are adding AI to their search services. Srinivas says that Google’s ad-driven model and huge resources are a big problem, but that its dependence on ad revenue makes it strategically stuck, which Perplexity can take advantage of. OpenAI’s SearchGPT and Microsoft’s Copilot are also trying to get a bigger piece of the market, but Perplexity stands out because it puts a lot of emphasis on being open and making the user experience better.
The company’s value has risen quickly, which has led some to wonder if it can keep going. Some people have even suggested that Apple might buy it, but Srinivas has said that Perplexity is committed to staying independent and might go public in the future.
The story of how Perplexity AI went from being a startup in 2022 to a powerhouse worth $18 billion by July 2025 is an amazing tale of innovation, smart planning, and breaking into new markets. Perplexity has changed the way we find and use information by creating the “answer engine” category, using cutting-edge AI, and encouraging fair partnerships. Perplexity is not only competing with Google; it’s also changing the way people search in the AI era as it grows through projects like the Comet browser and partnerships around the world. Perplexity AI is likely to stay a leader in the AI field for years to come because it has a visionary team, strong technology, and a focus on the user.
