Vikram Singh Mangat is an AI researcher, software engineer, and founder of Singularity Language. He is the architect behind LayerZero, an agentic computing framework built to help AI systems communicate, collaborate, and prove origin across multi-cloud environments.His research focuses on multi-agent AI, cross-cloud interoperability, provenance-backed licensing, and human-aligned governance. In this talk, Vikram explores why the next generation of AI systems will require a new business model—one built not around restricting access before use, but proving origin and enabling fair licensing after intelligent systems create value.

Beyond APIs: Proof of Origin as the New Business Model for Multi-Agent AIAbstract: The API economy was designed for a software world where access was the main event. Adeveloper called an endpoint, received a response, and paid for usage. That model workedwhen software interactions were relatively bounded, predictable, and human-initiated.Multi-agent AI changes this assumption. A single user request can now trigger chains of modelcalls, tool invocations, data retrievals, module instantiations, agent-to-agent handoffs, andcross-system workflows. In this environment, value is no longer created only at the point ofaccess. It is created through downstream use.This creates a structural challenge for the current AI business model. Infrastructure costs arerising, hardware efficiency improvements are no longer reducing costs at the same historicpace, and autonomous agents can generate large volumes of cross-system activity without thesame economic restraint a human operator would apply. Human approval can reduce runawayspending, but too much human oversight becomes a bottleneck to true self-supervisedautomation. At the same time, AI systems now operate across more available data points thanever before. Sourcing the right data at the right time can make applications more powerful, butsearching, filtering, ranking, verifying, and delivering relevant information across expanding dataenvironments also creates growing compute and energy costs.Our research indicates that proof of origin should become a foundational business model formulti-agent AI. Proof-of-origin systems can function like an algorithmic digital watermark: acryptographic record that helps identify where an AI module, dataset, behavior, workflow, oroutput originated and how it was used after implementation. Unlike traditional API accesscontrols, which mainly determine who can enter a system, proof-of-origin infrastructure can helpdetermine what happened after access was granted.This shift could accelerate innovation rather than restrict it. Instead of limiting powerful AIsystems only to users who can afford high upfront access costs, a proof-of-origin model couldencourage broader adoption by allowing modules, models, datasets, and agentic workflows tobe used more freely while preserving enforceable licensing records after implementation.Originators could allow wider experimentation, knowing that downstream commercial use,reuse, or monetization could still be traced, attributed, and licensed.Using Singularity Language as a practical research context, this talk explores howprovenance-aware interoperability can help the AI industry move from an access-first model to ause-aware model. The broader goal is not to promote a single platform, but to advocate for anindustry transition: if autonomous agents are going to collaborate across systems, the businessmodel must evolve from charging only for access to proving origin, measuring downstream use,and enforcing licensing after value is created.