Open 6G tools: what researchers are testing now

Open-source 6G tools link network simulation, AI training and real 5G test rigs

Open-source software is pulling 6G research out of pure simulation and closer to real radio hardware. On 6 November 2026, Nvidia published a round-up of its Aerial and Sionna platforms, drawing together several recent trials. The examples range from a robotic arm to a self-learning radio receiver to automatically switching off unused base-station capacity.

The headline numbers sound big, but they need a clear yardstick. What was tested were individual functions in laboratory and trial systems, mostly built on 5G technology. None of this proves a finished 6G specification or the performance of a future commercial network. For UK research institutions and companies, the release is still relevant: open source code can make experiments easier to reproduce and shorten the route into building a test rig of one’s own.

Open code for radio and simulation

Aerial CUDA-Accelerated RAN (github.com/NVIDIA/aerial-cuda-accelerated-ran) provides building blocks for a GPU-accelerated mobile base station: physical-layer processing, a radio resource scheduler, Python interfaces and reference models. It is released under the Apache 2.0 licence. Parts of it rely on Nvidia’s CUDA environment and matching hardware.

Sionna (github.com/NVlabs/sionna) is also published under Apache 2.0: a Python library for simulating individual radio links, whole systems and radio-wave propagation.

Nvidia puts usage at more than 1 million downloads and more than 8,000 research and higher-education institutions, with use said to have grown tenfold within a year. These are the vendor’s own figures, not an independent breakdown.

Four trials, four different goals

King’s College London tested goal-oriented transmission for a robotic arm: sending compact descriptions of objects instead of full camera images. Across 20 runs each, the approach cut task completion time by up to 52.6% in the best case and raised the success rate by as much as 45%. Source: arxiv.org/html/2609.15895 (14 September 2026). This is not a general speed gain for every robotics application.

SoftBank built a self-learning receiver into an Aerial-based 5G radio link. In a shielded lab test with a commercial smartphone, average uplink throughput beat the conventional receiver at all 15 interference levels tested, with the biggest reported gain at 60.5% (47.9 Mbit/s versus 29.9 Mbit/s at that measurement point). Source: softbank.jp (2 November 2026). Not a tariff or everyday speed, and a single-shift test inside a screened room.

The University of Leeds examined when radio capacity can be switched off without dropping devices: in a scenario with four base stations, power consumption of the radio access network fell by 45.6%, with no dropped connections during the trial. Source: eprints.whiterose.ac.uk (2026). This describes the chosen test set-up, not a nationwide network.

Spain’s CTTC research centre linked GPU-accelerated radio channel models to its own open network simulator: the run time of an early evaluation fell from around three hours to 10–20 minutes. Source: cttc.cat.

The UK is building its own route from lab to market

The UK already funds 6G as a research field in its own right. Since 2022 and 2023, the Department for Science, Innovation and Technology (DSIT, then DCMS) and UK Research and Innovation (UKRI) have launched a “Future Telecoms” national mission, with up to £100 million in initial funding and a further £70 million through the Technology Missions Fund for university-led Future Telecoms Research Hubs, alongside an earlier £110 million telecoms R&D package that included £28 million for 6G network research at the universities of York, Bristol and Surrey, working with Nokia, Ericsson and Samsung, plus an £80 million UK Telecoms Lab test facility in Solihull (sources: gov.uk, ukri.org, gov.uk). Open tools like Aerial and Sionna lower the barrier to building a radio test rig of one’s own, including for smaller UK companies and university spin-outs feeding into these hubs; “open source” does not mean vendor-independent (Aerial depends on Nvidia’s CUDA stack). The European Commission (digital-strategy.ec.europa.eu), whose 6G research the UK also takes part in through Horizon Europe association, expects the first 6G products and infrastructure towards the end of the decade, with robotics, connected vehicles, digital twins and sensing among the possible early uses; spectrum, standards, security and business models are all still open questions.

Good lab results are not yet a mobile network promise

These results answer narrowly framed research questions: specific objects and interference cases for the robotic arm, a screened room for SoftBank, a four-base-station scenario for Leeds. They do not replace testing in a live network with changing handsets, buildings, weather and many users at once.

For consumers, nothing changes yet about mobile tariffs or 5G reception. The benefit so far sits in research: earlier comparisons, traceable errors, shared test set-ups. Standardisation and investment decisions will decide the path into standards and commercial networks.

Sources and status

• Nvidia blog, 6 November 2026
• Aerial CUDA-Accelerated RAN repository
• Sionna repository
• King’s College London paper, arXiv, 14 September 2026
• SoftBank research topic, 2 November 2026
• University of Leeds paper (White Rose eprints), 2026
• CTTC (Spain)
• gov.uk: investment in telecoms innovation and R&D, DSIT, 11 April 2023
• UKRI: £70 million for new future telecoms technologies, 5 October 2023
• gov.uk: £110 million telecoms R&D package, DCMS, 13 December 2022
• European Commission: 6G policy page

Research and data as of: 6 November 2026.

Transparency note: AI-assisted tools were used in researching, structuring and writing this article. The sources used are listed in the article. Publication takes place only after manual editorial approval. Country adaptation based on an article from Net Trends, researched for the United Kingdom.