San Francisco, November 9 (Reuters/ PR Newswire ) – Edgevana is a data center, energy, and computing power orchestration platform that provides AI computing power ready for immediate deployment in production. The company announced today that it has reached a multi-year pay-as-you-go agreement worth over $700 million, with full commitment, to deploy NVIDIA Blackwell computing systems in North America and Europe. As power restrictions are slowing down the construction of traditional data centers, this commitment comes at a time when Edgevana has reached $7 billion in signed customer demand for its pre-powered platforms, with an active project pipeline exceeding $20 billion.
According to these agreements, Edgevana will orchestrate and deploy the NVIDIA B300 GPU and GB300 systems, covering North America and Europe. By integrating pre-electrified land, data center facilities, and high-speed networks into a single virtual park, Edgevana enables customers to launch high-density AI computing power of several dozen megawatts within 90 to 120 days. This allows enterprises to bypass the years-long delays associated with building new data centers from scratch, while still being able to manage their distributed computing infrastructure through a single interface.
Co-CEO Ryan Fay stated: "During the initial boom in training, we deliberately chose to remain low-key, and we still operate in this manner in many aspects today, as our model is not immediately apparent to the broader market. Training rewards centralized, ultra-large parks, but inference involves real-time production traffic. As performance requirements increase, this aligns perfectly with our core strengths. We initially laid the foundation for providing ultra-low latency services for high-frequency trading and managing edge production environments for the two largest hyperscale cloud service providers globally. Now, AI is shifting towards inference, and the market has come to us."
Edgevana has a global network covering 17 operational data center locations, with another 10 projects currently in progress. Overall, the platforms supported by Edgevana infrastructure serve over 10,000 downstream enterprise customers, most of whom are not even aware that Edgevana is the underlying support layer behind them.
Co-CEO Subhan Jahromi stated: "The decisive factor for AI is not the algorithm or software, but rather the physical heavy infrastructure. Although the market has always focused on ultra-megawatt facilities, we have been quietly securing power, land, and computing power over the years to bring the infrastructure closer to end-users. It is extremely challenging to deploy high-performance GPU at the edge in a reliable and large-scale manner. As our name implies, Edgevana makes edge deployments seamless."
Edgevana combines long-term energy supply, its unique GPU allocation method, and the committed customer demands to form a more capital-efficient deployment model. By locking in long-term contracts before deployment, Edgevana reduces financial risks, accelerates the launch of computing power, and provides exclusive AI capacity at a cost base lower than that of traditional suppliers.
About Edgevana:Edgevana was established in 2019 and is a global distributed AI computing power company dedicated to accelerating the deployment of the next generation of AI infrastructure. The company provides support for leading global hyperscale cloud service providers and cutting-edge AI laboratories, delivering high-performance AI clusters at unprecedented speed and scale. By integrating power, data centers, GPU infrastructure, and connectivity capabilities on a single platform, Edgevana significantly reduces the time to value realization, enabling AI laboratories and enterprises to deploy dedicated computing power in months rather than years. With strategic access to power, infrastructure, and advanced computing power, Edgevana offers the speed, scale, and economy required to drive the AI revolution. Edgevana states: “Together, Edgevana is powering the ideas that power the world.” For more details, see edgevana.com.
