Vapor IO and Supermicro have announced a unique 5G- and AI-as-a-service platform, called Zero Gap AI. Powered by the NVIDIA MGX platform with the NVIDIA GH200 Grace Hopper Superchip and operated by Vapor IO, the new platform delivers private 5G and GPU-based micro-clouds to specific locations. These locations are adjacent to where the capabilities are needed, such as next to retail storefronts, factory floors and city intersections.
Zero Gap AI is built around Supermicro servers that implement the NVIDIA MGX platform with the GH200 Grace Hopper Superchip, which delivers a high-bandwidth connection between the Nvidia Grace CPU and the Nvidia Hopper architecture-based GPU for high-performance accelerated computing and AI servers.
(Source: Shutterstock)
With the new platform, hyperlocal AI services can be delivered over private wired and wireless networks without running servers on premises, eliminating the need for onsite AI and 5G hardware, while leveraging private fiber from nearby access points. Key benefits include fast response times, a reduced reliance on centralized computing and the ability to handle sensitive data locally, according to the companies.
For example, a manufacturer can use Zero Gap AI real-time inferencing to support factory IoT to reduce risk and lower capital expenditures. In addition to manufacturing, the new platform can be used in a range of applications such as smart cities, healthcare, retail, warehouse and logistics, hospitality, transportation and public venues.
The wide range of use cases, from real-time inferencing to supercomputing at the edge, is thanks to Supermicro’s Nvidia MGX servers with the GH200 Superchip, which can support 5G and AI running on the same machine. This enables the optimization of the Zero Gap AI offerings for different use cases. These high-performance servers are being deployed in specific geographies across Vapor IO’s Kinetic Grid platform, which is currently available in 36 U.S. cities.
Due to the platform’s highly distributed architecture and automation capabilities, AI micro-clouds can be created where they are needed, such as next to a nearby factory or local hospital, on demand and instantaneously, the companies said. This on-demand scalability also provides scalable GPU resources and inferencing services to meet changing demands without significant upfront costs.
In addition, since the platform is delivered locally, the data can stay local, so it does not have to be sent to a centralized cloud. “This can save time and money while also providing fine-grained control over data sovereignty, locality and compliance,” the companies said.
Other features include near-zero latency for real-time data processing and decision-making in time-sensitive AI applications and high reliability thanks to zero downtime and zero congestion of Vapor IO’s Zero Gap network.
Zero Gap AI will be available first in Atlanta and Chicago, leveraging Comcast’s network infrastructure. Zero Gap AI customers can work at their choice of level, said the companies. They can operate at a low level with bare-metal instances, or they can use prepackaged third-party models and solutions for specific use cases.
Advertisement
