Ex Twitter CEO Parag Agrawal's New AI-Startup Parallel

Former Twitter CEO Parag Agrawal, who was ousted by Elon Musk following his 2022 takeover of the social media platform (now called X), has re-emerged with a new venture aimed at reshaping how artificial intelligence interacts with the web.

Agrawal has founded Parallel Web Systems Inc., a Palo Alto–based startup developing a cloud platform to help AI systems perform large-scale online research. Since launching Parallel in 2023, he has built a 25-member team and secured $30 million in funding from top investors, including Khosla Ventures, First Round Capital, and Index Ventures.

“We already power millions of research tasks every day, across ambitious startups and public enterprises,” Agrawal wrote on LinkedIn. “Some of the fastest-growing AI companies use Parallel to bring web intelligence directly into their platforms and agents. A public company automates traditionally human workflows with accuracy exceeding human levels using Parallel. Coding agents rely on our search to find documentation and debug issues.”

Parallel recently unveiled its Deep Research API, which Agrawal claims is the first system to outperform both humans and advanced AI models—including GPT-5—on two of the toughest industry benchmarks.

Inside Parallel’s Vision

Parallel is built on the idea that the internet, designed originally for humans, must now evolve for a new primary user: artificial intelligence.

In a blog post, the company outlined its philosophy:

The web as humanity’s memory: The open internet has enabled publishing, collaboration, and knowledge-sharing at scale, forming the foundation for modern AI.

AI as the web’s new user: Unlike humans, who browse selectively, AI systems can scan entire databases, process vast volumes of data, and generate insights instantly.

Limitations of today’s models: Business models based on clicks, ads, and paywalls don’t serve AI systems well and risk confining valuable knowledge.

A “Programmatic Web” for AIs: Parallel envisions an internet tailored for machines—optimized for reasoning, computation, and verified information.

Key Principles of Parallel’s “Programmatic Web”

Unified infrastructure that merges data, computing, and reasoning to deliver insights and actions, not just static pages.

Declarative interfaces where AIs specify what they need, while the system determines how to retrieve it.

Transparent attribution ensuring sources are credited and contributions measurable.

Open markets where value is rewarded economically, ensuring openness is sustained by incentives rather than goodwill alone.

How It Works

Parallel aims to create a machine-first internet. Instead of humans browsing or searching, AIs will directly request information. Parallel’s platform will then collect, process, and organize it, while crediting original sources and rewarding contributors.

In essence, Agrawal’s startup wants to build a fairer, more efficient web—one where artificial intelligence can reason, compute, and generate insights at scale without being constrained by outdated, human-centric models.