UK Space Agency invests £13 million in AI-driven satellite broadcasting

This investment signals a shift toward intelligent edge processing in satellite delivery, moving away from raw data transmission toward AI-reconstructed content. By funding AWTG’s generative broadcasting and Airbus’s reprogrammable processors, the UK is prioritizing bandwidth efficiency that could lower the high overhead costs typically associated with satellite-based video streaming. For the broader ecosystem, these advancements in 5G and 6G non-terrestrial networks provide a blueprint for maintaining high-quality streams in underserved geographic markets. Watch for the results of the ScotRail satellite trial next year to determine if hybrid space-ground networks can finally provide stable bitrates for mobile passengers in remote regions.

The UK Space Agency's £13 million investment sits within a broader global push to apply AI to spectrum management and network efficiency in non-terrestrial and hybrid connectivity scenarios. Ericsson's AI-native link adaptation technology, which achieved up to 20 percent higher downlink throughput and 10 percent better spectral efficiency in field tests with Bell Canada in April 2025, represents the terrestrial counterpart to the kind of intelligent bandwidth optimization the UK Space Agency is funding for satellite links. The technology uses real-time AI models to dynamically adjust data transmission rates based on changing channel conditions and interference, executing directly on the baseband unit rather than as a bolt-on layer.

The commercial stakes of spectrum efficiency have been underscored by major satellite-terrestrial convergence deals. SpaceX agreed to acquire EchoStar's AWS-4 and H-block spectrum licenses for approximately $17 billion, consisting of up to $8.5 billion in cash and $8.5 billion in SpaceX stock, to deploy a next-generation Starlink direct-to-cell constellation. The deal, which also includes roughly $2 billion in funded interest payments on EchoStar debt through November 2027, signals the premium that satellite operators place on licensed spectrum for mobile connectivity. Ericsson's Head of Networks Strategy Mårten Narvinger pointed to that $17 billion figure at MWC 2026 to illustrate why even a 10 percent AI-driven spectrum efficiency gain translates into billions in value, noting that most customers would adopt the software on existing equipment within the year.

On the infrastructure side, the convergence of AI processing with cloud-native radio architectures is advancing rapidly. AT&T and Ericsson demonstrated AI-native Link Adaptation on a Cloud RAN stack powered by Intel Xeon 6 SoC in March 2026, marking the first call with portable Ericsson software on that hardware configuration. The test achieved up to 20 percent throughput gains compared to legacy rule-based link adaptation and validated a hardware-agnostic deployment path. For the UK Space Agency's funded projects, including AWTG's generative broadcasting work and Airbus's reprogrammable processor efforts, these terrestrial benchmarks provide a reference point for the efficiency gains that AI-native network architectures can deliver when applied to satellite and non-terrestrial network links.