Former Twitter CEO Parag Agrawal is charting a new course for the internet’s future with his latest venture, Parallel Web Systems, an AI infrastructure startup that has raised $100 million in Series A funding to power web access for artificial intelligence agents. The round values the company at $740 million and was co-led by Kleiner Perkins and Index Ventures, with participation from Khosla Ventures and other existing investors — signaling strong conviction from top Silicon Valley backers in the emerging field of AI-native web search.

Building the Internet for AI Agents

Parallel aims to become the connective tissue between AI systems and the open web. Unlike traditional search engines, which deliver ranked links for human users, Parallel’s technology enables AI agents to directly query and retrieve live web data through application programming interfaces (APIs) — effectively transforming the internet into machine-readable “fuel” for intelligent systems.

Agrawal said in an interview that the company’s enterprise customers already use Parallel to power AI tools that write software code, analyze sales data, and assess insurance risk, all tasks requiring a mix of live web data and proprietary information.

“Think about how many jobs could be done if you turned off web access,” Agrawal said. “You can’t deprive a merger-and-acquisition lawyer of the internet — so why would you deprive their AI agent?”

By offering real-time web access, Parallel positions itself as a critical infrastructure provider for enterprise AI deployment — especially as companies increasingly rely on large language models that require accurate, continuously updated context.

A Smarter Web for Smarter Machines

Parallel’s approach addresses a key limitation of AI systems: their tendency to “hallucinate”, or generate plausible but incorrect information. The startup’s system doesn’t just scrape pages — it parses, filters, and returns optimized data “tokens” formatted specifically for model ingestion.

This distinction, Agrawal said, allows AI systems to produce more accurate responses while reducing the computational cost of running queries. “Our platform doesn’t rank results for humans; it delivers structured knowledge for machines,” he said.

The new funding will accelerate product development, enterprise expansion, and partnerships with online content owners, as Parallel looks to tackle a growing challenge: the fragmentation of web data behind paywalls and restricted APIs.

The Economics of the New Web

As publishers and social platforms increasingly lock down content to curb automated scraping, Agrawal said Parallel is working on an “open market mechanism” that would compensate data owners for allowing controlled AI access to their content. While he did not elaborate, such a model could lay the groundwork for a new digital economy, where publishers monetize structured data rather than advertising impressions.

Industry observers see this as a pivotal step toward reconciling AI’s appetite for information with the internet’s long-standing business models. “Parallel is trying to create the web’s next layer — a trust-based ecosystem for machine-to-machine data exchange,” said Lisa Huang, a partner at Kleiner Perkins. “That’s a trillion-dollar opportunity if done right.”

Positioning for an AI-Native Internet

Founded in 2023 and officially launched in August 2025, Parallel reflects Agrawal’s deep technical roots in distributed systems and search infrastructure — areas that defined his tenure as Twitter’s chief technology officer and later CEO. The company previously raised $30 million in seed funding in early 2024, led by Khosla Ventures.

With the new capital, Agrawal said Parallel will “go all in on building the data access layer for AI,” expanding its team, refining its real-time APIs, and establishing partnerships with major enterprise clients and data providers.

As AI agents evolve from productivity tools into autonomous digital workers, the demand for live, permissioned data will only grow — and Parallel aims to be the bridge connecting the AI ecosystem to the real-time web.

“AI systems will soon outnumber human users on the internet,” Agrawal said. “Our mission is to make sure they can read, reason, and respect the web responsibly.”