The UK's Future RAN Competition (FRANC) closed out all 14 funded projects in early 2024, deploying £36M of DSIT money against ~£36M of matched private R&D, and the resulting Open RAN software, DU/CU stacks, and AI-driven security modules are now the most concrete data point in the 2026 5G industrial module supply chain [S3].
FRANC engaged 56+ organizations and seeded technical work in Distributed Unit software, far-edge Kubernetes platforms, and Open RAN hardware disaggregation, all of which feed directly into the 5G industrial module, RU, and DU bill of materials that sourcing teams now have to qualify.
FRANC funding scale and 5G Open RAN deliverables
DSIT committed ~£36M across 14 FRANC winners, with the projects running between December 2021 and March 2024, and the headline objective was to accelerate 5G Open RAN solutions meeting UK dense urban requirements by 2025 [S3]. The largest single award, ~£3.3M, went to the Proteus project (Bristol), which built a hardware-agnostic DU that runs on a universal platform instead of being locked to one radio vendor [S3]. That matters for industrial module sourcing because DU software portability is what lets buyers mix RAN base stations, chipsets, and servers from multiple suppliers without redoing integration work.
Cambridge/Edinburgh-based partners, Microsoft UK, Intel R&D UK Ltd., Capgemini, and the University of Edinburgh, ran the second large track (~£1.8M) and produced a far-edge platform on low-latency Linux, with vendor-tuned Kubernetes features and APIs that bring a new server into full operation in roughly thirty minutes [S3]. The platform also bundles a marketplace-style installer for cloud-native RAN software, AI-powered anomaly analytics, and ransomware detection/recovery tooling [S3]. For an industrial buyer, the deliverable is a directly traceable piece of the Open RAN stack that can be specified into a private 5G module reference design.
Parallel Wireless UK Limited, British Telecommunications PLC, the University of Bristol, Real Wireless Limited, and Benetel were the Proteus partners, demonstrating that the same UK consortium model now feeds the Industrial Ethernet Upstream and Downstream: Specs, Media, and Sourcing Map conversation around fronthaul, midhaul, and backhaul media choices inside private 5G plants [S3].
Disaggregated supply chain, open interfaces, security priority
FRANC's three explicit policy objectives, disaggregated supply chains, open interfaces by default, and security as a priority in network deployment, are now embedded in UK procurement language for 5G industrial modules [S3]. Disaggregation lets an industrial site pick RU, DU, CU, and core software from different vendors, which breaks the single-supplier lock-in that has historically dominated 5G module sourcing.
Open interfaces by default means that R1, F1, E1, and the O-RAN fronthaul M-plane split become hard procurement requirements, not optional features, so any module that ships without documented conformance faces a specification gate at RFQ. Security-as-priority maps onto zero-trust module design, signed firmware, and tamper-evident hardware roots of trust, which industrial OT teams now audit alongside traditional functional specs.
Industrial buyers carrying out 2026 RFQs should treat disaggregation, open interface conformance, and security as a three-axis selection filter, because that is the exact wording UK policy uses to score supplier bids [S3]. The same axis logic shows up in the Embedded part selection for industrial facilities: 2026 spec map coverage of long-lifecycle industrial silicon, where the same disaggregation pressure pushes buyers toward modular SoCs instead of monolithic radio chips.
Module-level comparison: legacy integrated RAN vs Open RAN
On a four-axis comparison the two 5G industrial module paths diverge sharply. Legacy integrated RAN ties the radio, baseband, and higher-layer software to one vendor, which compresses integration time but kills multi-source flexibility and tends to lock the buyer into a single security-update cadence. [S3]
Open RAN, the path FRANC explicitly funds, splits the radio unit, distributed unit, and centralized unit across independent vendors, with AI-driven RIC (RAN Intelligent Controller) software sitting on top, and it shows longer initial integration windows (FRANC data suggests new servers come online in about thirty minutes once APIs are in place, but the first-time bring-up is heavier) [S3]. The trade-off is multi-vendor sourcing, faster feature upgrades, and the ability to mix-and-match RUs against different DUs, which is precisely the flexibility that procurement teams in chemical, port, and energy sites now demand.
For hazardous-area or outdoor industrial sites, the Open RAN path also supports richer edge-AI workloads because the far-edge platform was specifically optimised for real-time performance and energy efficiency in the Cambridge/Edinburgh track [S3]. Buyers who only need a single small private 5G cell under 50 users with no edge analytics are often better served by integrated RAN, because the integration overhead of Open RAN does not pay back at that scale.
Who Open RAN industrial modules are for, and who should skip them
Open RAN industrial modules, the ones FRANC seeded, are aimed at sites with multi-vendor sourcing mandates, dense-urban or dense-industrial coverage requirements (the FRANC 2025 dense-urban target is the proxy for plant density), and an OT/IT team that can absorb Kubernetes-class operational overhead [S3]. Chemical plants, ports, large warehouses, and energy substation clusters fall into this bucket, and they map onto the buyer profiles covered in SCADA system suppliers 2026: B2B pricing, vendor types, and spec gates, where modular software-defined radio is increasingly specified alongside traditional PLC RTUs.
Smaller sites, single-vendor greenfield builds, or buyers without Linux/Kubernetes operations skills should usually stay on integrated RAN, because the FRANC reference architecture leans on cloud-native tooling, low-latency Linux tuning, and AI-powered analytics that require operational maturity [S3]. Sites that have not yet budgeted for security operations (the FRANC AI/ransomware recovery stack is non-trivial to maintain) should also defer Open RAN until those capabilities are in place.
Failure modes and sourcing constraints to watch
Three failure modes show up repeatedly in Open RAN industrial module rollouts. First, fronthaul timing and synchronization, where the M-plane split demands strict Class C or Class B timing budgets; if the industrial site uses non-deterministic Ethernet switches on fronthaul links, the DU/RU link will drop. Second, security-update latency: an Open RU from a smaller vendor may not have the same CVE patch cadence as a Tier-1 OEM, and the security-priority procurement language from FRANC does not waive that obligation [S3].
Third, module obsolescence: industrial sites run 10-15 year lifecycles while Open RAN silicon generations refresh on roughly 24-36 month cycles, so a 2026 module decision can become end-of-life inside the plant's own refresh window. Sourcing teams should require a minimum 7-year module availability commitment and a documented second-source plan for the DU software, which is exactly the model the Proteus universal-platform deliverable tried to prove out [S3].
Standards, sourcing signals, and what to track next
O-RAN ALLIANCE fronthaul specifications, 3GPP Release 16/17 industrial IoT features, and ETSI EN 303 645 for consumer-grade IoT cyber security are the governing references for Open RAN 5G industrial modules, though buyers should verify the exact revision in force at the time of RFQ rather than assume the latest. The FRANC closure reports themselves, including the Proteus closure report, are the most authoritative UK-government-backed evidence that the Open RAN industrial module stack is technically deliverable, and they remain a defensible citation in any sourcing justification [S3].
Trackable signals for the next 6-12 months include the publication of follow-up UK diversification funding rounds beyond the original £36M DSIT envelope, the volume of Open RAN industrial module SKUs entering distributor catalogs (versus custom-integration projects), and any new AI/RIC software releases that reuse the Cambridge/Edinburgh far-edge platform components [S3]. Buyers waiting on those signals can shortlist modules against the 30-minute server bring-up benchmark and the multi-vendor DU portability test that Proteus demonstrated in a live network environment.
For the relevant spec sheets and selection criteria, see power supply, dc power supply, and industrial ups.