While most people think the internet was built for humans, former Twitter CEO Parag Agrawal believes it’s time to create a version designed specifically for artificial intelligence. His new company, Parallel Web Systems, just raised $30 million to make this vision a reality.
Founded in 2023 after Agrawal’s departure from Twitter, the Palo Alto-based startup quietly assembled a 25-person team with talent from Google, Airbnb, and Waymo. The company focuses on building infrastructure that helps AI agents research information on the web more effectively than current methods allow. Agrawal abruptly fired following Elon Musk’s $44 billion takeover of Twitter in late 2022, which led to his entrepreneurial pivot into AI infrastructure. Their approach exemplifies how combining multiple AI models can enhance accuracy and adaptability in complex tasks.
Founded in 2023, the Palo Alto startup assembled top talent from Google, Airbnb, and Waymo to revolutionize AI web research infrastructure.
The funding round attracted impressive investors including Khosla Ventures, First Round Capital, and Index Ventures. The investment valued the company at $450 million, with plans for a follow-on $100 million round led by Kleiner Perkins and Index Ventures. The strong investor confidence demonstrates the market’s belief in AI infrastructure solutions addressing critical web research limitations.
Parallel Web Systems develops APIs that enable AI agents to conduct deep web research. Think of it like giving AI systems super-powered search abilities that go far beyond what regular search engines can do. The company built eight specialized research engines, with the fastest delivering results in under one minute and their most advanced system, Ultra8x, taking up to 30 minutes for detailed analysis.
The technology claims impressive results, outperforming GPT-5 in accuracy by 10 to 15 percent on rigorous benchmarks. The platform already processes millions of AI research tasks daily for early adopters, including some of the fastest-growing AI companies.
Agrawal’s team believes AI will eventually use the web far more than humans do. Current internet infrastructure creates challenges for AI with advertisements, paywalls, and restricted access to information. Their solution aims to create what they call a “programmatic web” that works seamlessly with AI reasoning and computation.
The company serves various use cases, from helping AI coding agents search technical documentation to enabling public companies to automate research workflows. Early results show the system surpasses human-level accuracy in certain research scenarios.
As AI becomes more prevalent in daily operations, Parallel Web Systems positions itself as a pioneer in building the infrastructure that AI agents will need to navigate and understand information effectively. Effective risk management and adaptability remain key as AI-driven systems evolve and integrate into complex environments.
