2 min Applications

Google challenger Perplexity nearing 500 million investment round

Google challenger Perplexity nearing 500 million investment round

In a new round of investment, Perplexity hopes to raise $500 million. This would take the AI search engine to a tripled value of $9 billion.

It is a significant increase from the previous investment round in April. Back then, Perplexity had raised more than $250 million with an estimated value of $3 billion. That was already a huge increase from January 2024, when it was supposedly worth 540 million.

Among the investors are SoftBank, Jeff Bezos and Nvidia. Interestingly, the new round of investment had come about not because Perplexity asked for it, but because the previous backers wanted to make additional investments.

No proprietary AI

Sizable investments are more common among AI companies. One advantage for Perplexity is that it largely relies on others’ AI models. While a party like OpenAI has to convert the many billions received into pricey training and R&D for new LLMs, Perplexity does not have to produce the foundation models itself. Instead, it ties other people’s models to its own algorithms.

Perplexity’s interface is very reminiscent of that of ChatGPT. Like OpenAI’s offering, the AI chatbot answers follow-up questions, but explicitly about search queries. A core promise of Perplexity, then, is that it is the “answer engine” for all kinds of questions, whether they be about politics, the weather forecast or a recreational topic. There are far fewer limitations to it than Microsoft Copilot, which also uses OpenAI models.

Read more: Is Perplexity a preview of online search’s AI-driven future?

Internal documents

Perplexity’s functionality has recently expanded, making it a lot more interesting. It can now do more than just search the Internet, but also internal documents. In addition, Perplexity is able to serve specific industries, such as the financial industry.

By the way, as with other GenAI apps, Perplexity’s approach regularly leads to hallucinatory answers. Wired also found out in June that the “scraper bot,” the robotic Internet visitor behind the search engine, manages to get hold of content it really shouldn’t have access to.