SignSplit Emerges from Stealth with $400M Seed to Build a 'Signed-Data' Economy for AI

The generative AI boom was built on scraping the open web, a model that is now colliding with copyright lawsuits and a looming shortage of high-quality training data. To build the next generation of models and autonomous robots, AI developers need access to verified human knowledge, skills, and likenesses—but lack a standardized infrastructure to acquire and compensate for them.

Attempting to solve this data bottleneck, SignSplit emerged from stealth today with a $400 million strategic seed commitment led by W Group, valuing the Delaware-based Public Benefit Corporation at $1 billion. Founded by [Glib Denisov] and [Alessandro Monterosso], the startup has built a platform for what it calls "signed data."

The platform allows individuals and institutions to upload their data, creative work, or digital likeness with explicit consent, clear provenance, and defined licensing terms. AI developers, researchers, and robotics firms can then access these specialized data pools to train their systems, while the original contributors receive financial compensation. A built-in verification layer also allows social media platforms to identify signed content and trace its origins.

Unlike traditional data brokers that scrape and sell aggregated datasets without creator consent, SignSplit is attempting to build a direct marketplace between human creators and AI consumers. If successful, this infrastructure could shift the AI industry from an extraction-based data model to a licensing-based economy, providing a scalable alternative to the ad-hoc data partnerships currently being negotiated by major AI labs.