Embedded LLM has officially unveiled TokenVisor, its new GPU monetisation and control platform, during the Advancing AI 2025 event in Santa Clara, California. Billed as the first purpose-built solution for the AMD GPU-powered neocloud ecosystem, TokenVisor is designed to streamline how neocloud providers and enterprises manage and monetize GPU clusters used for running Large Language Model (LLM) workloads.
For many early adopters, TokenVisor represents a significant leap forward as it is designed to make GPU infrastructure setup as frictionless as possible via features like custom pricing, automated resource allocation, rate limiting, and real-time usage tracking. These tools give both cloud providers and internal enterprise teams the ability to go live quickly, reduce overhead, and generate revenue sooner after installing their AMD GPUs.
According to Embedded LLM, the platform isn’t just about managing hardware; it’s about unlocking the economic potential of decentralised AI computing. The company’s CEO, Ooi Ghee Leng, described TokenVisor as “the hypervisor for the AI Token era”, emphasising that truly decentralised GPU computing needs software that’s just as powerful and adaptable as the hardware behind it.
Built closely with AMD, Team Red’s ecosystem also includes the broader neocloud community, thanks to native support for a wide range of popular LLMs and multimodal models that could appeal to startups and research teams who want a fast on-ramp to AI experimentation without massive backend complexity. Other than that, Embedded LLM’s is committed to providing responsive technical support and commitment to industry-grade standards to its clients.
TokenVisor is built with deep input from the AMD ecosystem and the broader neocloud community. Its launch at Advancing AI 2025, an event known for championing open-source and AI innovation, underscores the collaborative DNA behind the platform. Mahesh Balasubramanian, AMD’s Senior Director of Product Marketing for Data Center GPUs, praised the platform for giving providers the tools they need to better manage and monetize their LLM workloads.
With native support for a wide range of popular LLMs and multimodal models, TokenVisor also appeals to startups and research teams who want a fast on-ramp to AI experimentation without massive backend complexity. Embedded LLM’s responsive technical support and commitment to industry-grade standards further increase its appeal.
