Datum expands AI network cloud with DataBank PoPs in three U.S. hubs

Datum has deployed network points of presence inside DataBank facilities in Ashburn, Dallas and Salt Lake City to help AI companies, hyperscalers and enterprises connect distributed workloads. The move extends Datum’s open network cloud into three major interconnection markets as demand grows for AI infrastructure that blends physical connectivity with programmable networking.

Why it matters: - Datum is placing its AI-focused network layer in markets where cloud, inference and enterprise traffic already concentrates. - The deployment gives customers more options to connect across clouds and providers without relying on the same level of manual network coordination. - The move targets workloads that need secure routing, policy and observability across distributed infrastructure.

What happened: - DataBank said Datum has added network points of presence inside DataBank facilities in Ashburn, Dallas and Salt Lake City. - The new PoPs extend Datum’s open network cloud into three of North America’s key AI interconnection markets. - The companies framed the rollout as part of a broader buildout for AI infrastructure.

The details: - Datum serves neoclouds, inference providers, data platforms and enterprises that need to connect distributed workloads. - The platform includes an Envoy-based global edge, a microVM compute runtime and an SRv6-powered Galactic VPC. - Datum says the stack is designed to bring physical interconnection and private-network capabilities to builders through a developer- and agent-friendly experience. - Customers can apply consistent routing, policy and observability from a single control plane across clouds and providers. - Those capabilities have traditionally required large networking teams and direct relationships across the internet ecosystem. - Ashburn was selected as the largest interconnection hub in North America. - Dallas was selected as a growing central-U.S. exchange point. - Salt Lake City was selected for proximity to U.S.-West neocloud and public cloud regions. - Datum’s cloud is available in public and dedicated models and is backed by AGPLv3 open source. - Datum was founded in late 2024 by industry veterans. - Datum’s stated mission is to help the next 1,000 clouds thrive in the AI era. - More information is available in Datum’s announcement. - DataBank operates 70+ HPC-ready data centers across 25+ markets, with 20 interconnection hubs and on-ramps to a broad ecosystem of cloud providers. - DataBank has been recognized by Deloitte in 2023 and 2024 and by Inc. 5000 in 2024 as one of the fastest-growing private U.S. companies.

Between the lines: - The deal shows how AI infrastructure is shifting toward a mix of software-defined networking and physical colocation presence. - DataBank is positioning its facilities as a place where next-generation network control can sit closer to the edge. - The emphasis on application-layer interaction suggests AI agents and developers are changing how network services are consumed. - The long relationship between the companies’ leaders likely helped make the deployment possible.

What’s next: - DataBank said the partnership supports the continued buildout of AI ecosystems. - Datum is signaling broader U.S. expansion as it tries to unlock hyperscale-style connectivity and control for new clouds and builders. - The companies are likely to lean on these hubs as traffic, inference and enterprise AI workloads spread across more locations.

The bottom line: - Datum is using DataBank’s interconnection footprint to make its open network cloud more useful for AI builders that need fast, programmable connectivity across multiple environments.