Building for the Web’s Second User: Parallel’s $100M Series A

Have you ever felt inevitability? That serene moment when you sense a shift on the horizon, long before the world gives it a name?

In the 1990s, the internet changed the way businesses operate and information flows. In the early 2000s, cloud computing transformed how software is built and delivered. By the late 2000s, social media had emerged as the dominant force in online interactions, turning social platforms like Facebook, Twitter, and LinkedIn into global powerhouses.

Each wave had opportunists who caught it early—but a rare few who sensed it coming and built before it broke.

Two years ago, it was already becoming clear that generative AI would drive the next great swell—one powerful enough to reshape how technology is built, distributed, and applied. What was less obvious—or what fewer people saw—was how that wave would break. Most believed large language models, once trained on the full sweep of the internet, would have no reason to return to it. Why would an AI need to search what it already knows?

Parag Agrawal and his team at Parallel took the opposite view: that AIs would depend on the web even more than humans do. Not the human web of individually-rendered pages and visual interfaces, but something more fundamental—what they call “the living corpus of human knowledge that grows and changes from moment to moment.”

As AI unleashes new productivity and creativity across industries, agents are becoming the web’s primary users. Yet the internet wasn’t built for them—rate limits restrict data, HTML wastes compute, and search engines optimize for clicks, not comprehension.

Parallel’s mission is to keep the web open, transparent, and competitive. They are building infrastructure for AI agents, applications, and systems to access and think on the web. Their web agents and tools are designed for how AIs actually think—ranking optimal tokens, minimizing hallucinations, and maximizing reasoning. Where today’s models try to memorize the world, Parallel helps them understand it.

When we first invested in Parag and his team in 2024, we wrote an unconventional memo. There wasn’t much about what they were building, the market, technology, traction, or business model. We actually weren’t sure we knew any of that. But we did have pages and pages of notes about Parag and his team. We look for imagination, execution, and decisiveness—and in AI, a religion about a lack of religion.

Parag stood out immediately. Beyond his 11 years at Twitter, where he built one of the world’s most influential platforms, he combines analytical precision with intuition—a rare balance essential for building infrastructure for a new kind of user: AI itself.

Since then, Parallel has made remarkable progress. In August, they launched the first web search API built for AI agents—now used by companies like Clay, Sourcegraph, Genpact, and several Fortune 100 enterprises. Their platform powers deep web research, enrichment, and automation, helping developers and systems access accurate, verifiable information.

We couldn’t be more excited to double down on Parallel with a $100 million Series A co-led with Kleiner Perkins. The investment will expand agent capabilities, develop faster and more intelligent web primitives, scale what’s already the highest-quality web index, and advance fundamental research on how AIs interact with information.

As the Parallel team puts it: “The web is humanity’s living memory of the present. Without it, AIs can only experience the past.”

Parallel is building the foundation for the next internet—one designed not just to be read by humans, but to be understood by machines.