Parag Agrawal’s startup raises $100M to build a parallel web for AI agents

Former Twitter Inc. Chief Executive Parag Agrawal has raised $100 million to fund his new startup Parallel Web Systems Inc., which is focused on enabling artificial intelligence agents to search the web more efficiently.

Sequoia Capital led today’s Series B round and was joined by existing investors Kleiner Perkins, Index Ventures and Khosla Ventures, bringing the startup’s valuation to $2 billion.

Agrawal, who serves as Parallel’s founder and CEO, told the Wall Street Journal that his new company will use the funds to establish a sales and marketing team and accelerate its research and development plans. The startup is building a platform that enables autonomous AI agents that perform work on behalf of humans to search the internet for information with much greater accuracy than they do now. It previously raised $100 million in funding in November.

Agrawal believes that agents will ultimately “use the web a lot more than humans,” and require a different kind of infrastructure to access it properly. More specifically, he said they need the tools to perform “deep research” so they can complete tasks related to insurance claims processing, sorting through government contracts and so on. While humans would use a web browser to do this kind of thing, agents can potentially do it much faster if they have the right tools.

Parallel’s solution is a programmatic web infrastructure based on a suite of specialized application programming interfaces for searching the internet, performing tasks online, extracting information from websites and monitoring the web. The APIs utilize a proprietary web index that’s optimized for “machine retrieval.” Agrawal founded the company in early 2024, less than two years after he was forced out of his job at Twitter by Elon Musk, who acquired the social media startup in 2022 before transforming it into X Corp.

One of the first companies to use Parallel’s tools is the legal AI startup Harvey AI Inc., which has developed autonomous agents designed to perform research-heavy tasks on behalf of lawyers. Harvey co-founder and President Gabe Pereyra told the Journal that it’s not enough for agents to just be given access to Google Search. He believes that they require more granular control over which websites they should be accessing, and said that’s what Parallel provides.

Sequoia partner Andrew Reed said Parallel is providing the core infrastructure needed to support “long-running” AI agents that can operate continuously in the background and maintain context for longer periods of time. “One of the things that is a core shared function amongst all of these long-horizon agents is the ability to use the web,” he pointed out.

Since launching in early 2024, Parallel has amassed a user base of more than 100,000 developers, including many from AI-native startups and large enterprises, Agrawal said. But Parallel is not the only game in town, with rivals including Tavily Inc. and Exa Labs Inc. also building similar infrastructure to help AI agents navigate the web.

Image: Parallel Web Systems