Ex-Twitter chief launches Parallel with $30M funding—says its Deep Research API surpasses GPT-5

Written by

Published 18 Aug 2025

Fact checked by

We maintain a strict editorial policy dedicated to factual accuracy, relevance, and impartiality. Our content is written and edited by top industry professionals with first-hand experience. The content undergoes thorough review by experienced editors to guarantee and adherence to the highest standards of reporting and publishing.

disclosure

parallel programmatic web surpass gpt5

Parag Agrawal refused to take a break after Elon Musk fired him as Twitter’s CEO in 2022. Instead, he spent his days in Palo Alto coffee shops building what would become Parallel Web Systems, an artificial intelligence (AI) startup that has already raised $30 million from top investors.

“Today, we’re announcing the only AI system to outperform both humans and leading AI models like GPT-5 on the most rigorous benchmarks for deep web research,” the company stated.

    The Palo Alto company creates what it calls a “Programmatic Web” designed specifically for AI systems rather than humans. Its platform allows AI agents to search, understand, and organize web content with greater accuracy than existing solutions.

    Parallel offers eight different research engines for various computational needs. The fastest delivers results in under one minute. The most advanced engine, called Ultra8x, spends up to 30 minutes conducting detailed research and analysis.

    Developers can access these tools through three different application programming interfaces (API).

    Parallel outperforms humans and all leading AI models.

    Source: Parallel

    The startup’s main product is the Deep Research API, which processes millions of queries daily for early customers. The system searches public web data in real time, verifies accuracy, and packages results with citations and confidence scores.

    Parallel’s technology supports multiple applications across different sectors. Coding assistants can pull code snippets directly from GitHub. Retailers can track competitor product catalogs. Market researchers can organize consumer reviews into spreadsheets automatically.

    Agrawal believes AI agents will soon outnumber humans on the internet. “You’ll probably deploy 50 agents on your behalf to be on the internet,” he predicted. “And that’s going to happen soon, like next year.”

    The company has attracted backing from prominent Silicon Valley investors, including Vinod Khosla through Khosla Ventures, Index Ventures, and First Round Capital. Parallel now operates with 25 employees from its Palo Alto headquarters.

    The former Twitter chief executive launched Parallel just three years after his dramatic dismissal. Musk fired Agrawal and other top executives immediately after completing his $44 billion Twitter acquisition, denying them severance payments worth nearly $50 million.

    “I’m not a person that can enjoy the beach in that moment,” Agrawal told Bloomberg about his decision to immediately start working on his next venture.

    The Stanford computer science PhD and former International Physics Olympiad gold medalist spent over a decade climbing Twitter’s ranks. He started as an advertising team engineer in 2011 before becoming CTO under Jack Dorsey, then CEO in late 2021.

    His brief tenure as Twitter’s leader was dominated by legal battles with Musk over the acquisition. Many Twitter employees were surprised when Dorsey pushed the board to name Agrawal as his replacement.

    Agrawal maintains he has no regrets about his Twitter experience. “I don’t think Twitter defines me,” he said. “If I do a good job, I’m hoping this company will define me.”

    Agrawal plans to expand beyond current developer tools. Phase two involves creating fair compensation systems for content creators whose work feeds AI research systems, addressing ongoing concerns about AI companies using online content without permission.