Edgevana Secures $700 Million in Multiyear AI Computing Deals

By Nathan Brooks

Edgevana said it has secured more than $700 million in fully committed, multiyear agreements to deploy NVIDIA Blackwell computing systems across North America and Europe. The contracts lift the AI computing infrastructure company’s committed customer demand to approximately $7 billion. Its active business pipeline exceeds $20 billion.

The agreements involve take-or-pay commitments, under which customers contract for computing capacity over multiple years. Edgevana intends to deploy NVIDIA B300 GPUs and GB300 systems through its distributed data center infrastructure. The company is developing a model that combines existing power, data center facilities, computing equipment and high-speed network connectivity.

By coordinating those resources across multiple locations, Edgevana aims to accelerate delivery of production-ready AI computing capacity. The company says its approach can bring tens of megawatts of high-density AI infrastructure online within approximately 90 to 120 days. Edgevana says traditional data center development can take substantially longer because of land acquisition, utility connections, power infrastructure, equipment procurement and construction.

What does this change for AI customers?

Edgevana says its model can reduce deployment delays for customers that might otherwise have to build entirely new data centers. The company selects locations where power and other essential infrastructure already exist, then coordinates computing systems across its network. It calls the resulting distributed footprint a virtual campus, which customers access through a unified management environment.

Edgevana’s commercial structure links capital deployment to contracted demand. The company combines long-term energy supply arrangements, access to GPU computing equipment and committed customer contracts. By securing customer demand first, it seeks to reduce financial exposure and improve capital efficiency. Edgevana believes its pre-energized infrastructure can provide a competitive advantage as demand for high-performance AI computing increases.

Ryan Fay, Co-CEO Of Edgevana, said:

“We stayed in stealth through the initial training boom on purpose, and in many ways, we still operate that way because our model isn’t obvious to the broader market. Training rewards centralized mega–campuses, but inference is live production traffic. As performance demands increase, it plays directly into our core strengths. We cut our teeth delivering ultra-low latency for high-frequency trading and managing edge production for two of the world’s largest hyperscalers. Now that AI is pivoting to inference, the market has come to us.”

Edgevana said its experience includes delivering low-latency infrastructure for high-frequency trading and supporting edge computing operations for two major hyperscale technology companies. It is applying those capabilities to advanced AI accelerators.

Subhan Jahromi, Co-CEO Of Edgevana, said:

“The defining constraint on AI isn’t algorithms or software, it’s physical, heavy-duty infrastructure. While the market focused exclusively on mega-gigawatt campuses, we spent years quietly securing the power, land, and compute required to bring infrastructure closer to the end user. Getting high-performance GPUs to the edge reliably and at scale is incredibly difficult. As our name suggests, Edgevana makes edge deployment seamless.”

Edgevana believes AI workloads are shifting from centralized model training toward inference applications that support active users and business systems. The company says inference can require low latency, reliable network connectivity and geographically distributed computing resources. Those requirements create an opening for infrastructure providers that can place AI compute closer to end users.

According to Edgevana, its global infrastructure network includes:

• 17 operational data center locations.
• Another 10 sites under active development.
• Platforms serving more than 10,000 downstream enterprise customers.

Edgevana cited power availability as an important constraint on developing large AI data centers. It said companies seeking substantial computing capacity must secure reliable electricity while coordinating physical infrastructure and network requirements. Its platform is designed to address those constraints through coordinated deployment across multiple sites.

The $7 billion in committed demand represents customer contracts. The more-than-$20-billion pipeline represents potential additional business, not completed commitments. Edgevana was founded in 2019 and intends to continue expanding its network and accelerating AI infrastructure deployment.

The company positions its distributed model as an alternative to relying exclusively on large, centralized data center campuses. Combining available power, high-performance computing equipment and network infrastructure is intended to reduce deployment timelines and support customers requiring dedicated AI capacity. Edgevana’s next test is whether its pipeline converts into binding multiyear commitments. Its 90-to-120-day deployment window will matter as inference moves closer to live users.