Scalstrm and NETINT launch platform to meet EU cloud sovereignty requirements

The January 2027 deadline for the EU Data Act forces European operators to move beyond simple data residency toward true digital sovereignty. By utilizing NETINT video processing units, Scalstrm aims to offset the typically higher costs of sovereign cloud facilities through extreme hardware density and power efficiency. This shift suggests that compliance no longer requires a 'sovereign premium' that often stalls infrastructure budgets. As regulatory pressure mounts, the ability to maintain portable workloads across regional and national facilities will become a competitive necessity for broadcasters. Watch for whether other hardware-accelerated solutions emerge to challenge the cost-efficiency benchmarks set by this VPU-based architecture.

NETINT has been positioning its video processing units as a hardware alternative to GPU-based transcoding for cloud and edge deployments. The company's Quadra T1U and T1L accelerators target high-density encoding workloads where power efficiency and rack density matter most, a value proposition that aligns directly with the cost constraints of sovereign cloud facilities. In the broader competitive landscape, Nokia and AWS have been building integrated cloud platforms for telecom workloads, demonstrating how infrastructure vendors are layering AI automation on top of cloud-native architectures to reduce operational complexity for operators managing multi-domain environments. This trend toward unified, cloud-agnostic platforms mirrors the portability requirements that EU Data Act compliance will impose on video workloads. The EU Data Act, which takes full effect in January 2027, mandates that data generated by connected devices and services must be accessible and portable across providers, effectively forcing operators to avoid vendor lock-in in their cloud stacks. For video infrastructure, this means transcoding and packaging pipelines must be deployable across multiple sovereign cloud providers without re-engineering. Ericsson launched its AI in RAN commercial software subscription in June 2026, claiming up to 20% higher downlink throughput across more than 15 live deployments, illustrating how telecom vendors are already packaging software as subscription services that can run on existing silicon, a model that Scalstrm's approach extends to video processing. The shift from proprietary appliances to software-defined, hardware-accelerated pipelines is becoming the dominant architecture for operators seeking regulatory compliance without capital expenditure spikes. On the technical side, NETINT's VPU architecture differs from GPU-based approaches by offloading the full video codec pipeline to dedicated silicon rather than relying on general-purpose compute with software codecs. Ericsson and Nokia are diverging on how they use accelerators in their RAN platforms, with Nokia running all Layer 1 functions on Nvidia GPUs while Ericsson reserves GPU capacity for forward error correction alone, a distinction that parallels the debate in video processing between GPU-transcoding and dedicated VPU approaches. Scalstrm's claim of 110 HD channels per rack unit using NETINT hardware suggests a density advantage over GPU-based alternatives, which typically consume more power per channel and require larger cooling footprints. For operators building out sovereign cloud capacity under tight timelines, that density differential could determine whether compliance is achievable within existing facility budgets or requires new construction.