A startup led by Parag Agrawal raised $100 million [1] on April 28, 2026, to build a parallel web for AI agents.
This development signals a shift in how internet infrastructure may evolve. While the current web is designed for human navigation, a dedicated layer for AI agents could fundamentally change how data is accessed and processed by autonomous systems.
The funding comes at a time of rapid expansion in the artificial intelligence sector. The new venture intends to create an environment where AI agents can operate more efficiently than they do on the standard human-centric web, a goal that requires significant structural changes to how information is indexed and retrieved.
By establishing a parallel web, the company seeks to remove the friction agents face when interacting with traditional websites. This infrastructure would theoretically allow AI to communicate and exchange data in a native format, bypassing the need to interpret visual layouts designed for human eyes.
The $100 million [1] investment provides the necessary capital to recruit specialized engineering talent and develop the underlying protocols for this new network. The scale of the funding reflects the high stakes involved in the race to define the next generation of the internet.
Agrawal, who previously served as the chief executive officer of Twitter, is now applying his experience with large-scale platforms to the challenge of AI autonomy. The project represents a bet that the future of the internet will not be a single entity, but rather a dual system serving both biological and synthetic users.
“Parag Agrawal's startup has raised $100 million to develop a new web infrastructure.”
This investment highlights a growing industry belief that the current World Wide Web is an inefficient interface for artificial intelligence. By attempting to build a 'parallel web,' Agrawal is betting on a future where AI agents do not just scrape human websites, but interact within a purpose-built ecosystem. If successful, this could lead to a bifurcation of the internet, separating the visual web used by people from a data-centric web used by machines.
