NRI PULSE STAFF REPORT
Palo Alto, CA, August 19, 2025: Nearly three years after being unceremoniously ousted by Elon Musk during his $44 billion takeover of Twitter, former CEO Parag Agrawal has staged a bold comeback. The Indian-origin technologist has launched Parallel Web Systems Inc., a Palo Alto-based startup building infrastructure tools designed for the rapidly evolving world of artificial intelligence.
Backed by heavyweight investors including Khosla Ventures, Index Ventures and First Round Capital, the company has raised $30 million in funding. Agrawal’s new venture is already drawing attention with its flagship product, the Deep Research API, which allows AI agents to autonomously research, verify and organize information from across the web in real time.
According to Parallel, its tool not only streamlines workflows such as debugging, document discovery and long-form synthesis, but also outperforms both humans and advanced AI models like OpenAI’s GPT-5 on specialized benchmarks. The startup claims its accuracy on the “BrowseComp” test stands at 58 percent, compared with GPT-5’s 41 percent and humans’ 25 percent.
Agrawal, 40, was born in Ajmer, India, and moved to the United States for graduate studies after completing his engineering degree at IIT Bombay. He earned a PhD in computer science from Stanford University before joining Twitter in 2011 as an engineer. Rising quickly through the ranks, he became the company’s Chief Technology Officer and, in 2021, its CEO—the youngest leader of an S&P 500 company at the time. His tenure was cut short in October 2022, when Musk completed his Twitter takeover and dismissed him along with several top executives.
The company he now leads employs about 25 engineers and researchers, many with backgrounds at top tech firms including Google, Stripe, Airbnb, and Twitter. Agrawal says the goal is to build infrastructure for a future where AI agents, not humans, are the primary users of the web.
“It’s time to reimagine the internet for machine-native interaction,” he has said, outlining a vision of unified data infrastructure, transparent attribution, and declarative APIs to power the next wave of intelligent systems.
