AnswersPublished21 sources
How Supermicro, Nokia and Nvidia Plan to Bring AI-RAN Across the Network
On October 21, 2026, Supermicro announced an AI RAN collaboration with Nokia and Nvidia that combines radio and AI workloads on shared infrastructure, from cell sites to core hubs. The design spans cell site, aggregation and core hub systems; a dual Vera 2U model was shown at GTC Berlin and was expected later in 2026.
Published byEdited with GPT-6 LunaImages generated with GPT Image 2
Research answer
Key takeaways
- On October 21, 2026, Supermicro announced an AI RAN collaboration with Nokia and Nvidia that combines radio and AI workloads on shared infrastructure, from cell sites to core hubs.
- The design spans cell site, aggregation and core hub systems; a dual Vera 2U model was shown at GTC Berlin and was expected later in 2026.
- T Mobile reported a commercial network AI RAN call, while BT, Deutsche Telekom and Elisa announced live network pilots; these are evidence of testing, not proof of large scale paid deployments.
AI Prompt
openai.comCreate a landscape editorial hero image for this Studio Global article: What did Supermicro announce on October 21, 2026, about its collaboration with Nokia and Nvidia to deploy AI-RAN from distributed cell sites. Article summary: On October 21, 2026, Supermicro announced a collaboration combining Nokia RAN technology, Nvidia’s Aerial RAN Computer (ARC) platform and Supermicro infrastructure to deploy AI-RAN from distributed cell sites to centrali. Topic tags: general, general web. Style: premium digital editorial illustration, source-backed research mood, clean composition, high detail, modern web publication hero. Use reference image context only for broad subject, composition, and topical grounding; do not copy the exact image. Avoid: logos, brand marks, copyrighted characters, real person likenesses, fake screenshots, UI text, readable text, watermarks, charts with fake numbers, clic
Supermicro announced on October 21, 2026, that it is working with Nokia and Nvidia to deploy AI-RAN infrastructure from distributed cell sites to centralized network hubs. The proposed stack combines Nokia RAN software, Nvidia’s Aerial RAN Computer (ARC) platform and Supermicro systems, so operators can run radio processing and AI workloads on shared infrastructure. The cell-site ARC-Pro is specified at up to 1.6 PFLOPS of FP4 AI performance, but customer testing was planned for later in 2026 and general availability for the first half of 2027. 9
21
What the collaboration is designed to do
Nokia’s AI-native anyRAN software runs on Nvidia ARC, which provides accelerated computing for radio functions and AI workloads. Supermicro supplies systems for different points in the network. The companies’ stated aim is to support both AI for RAN—using AI to improve radio-network functions—and AI on RAN—running AI services on network infrastructure closer to users and data. Nokia describes the software as a foundation for integrating AI-powered radio functions intended to improve network capacity and performance. 16
31
The companies have pointed to voice AI, vision AI and autonomous network operations as potential workloads, alongside low-latency AI services at the network edge. Supermicro also describes the platform as a 5G solution with a software-upgradeable path intended to protect infrastructure investment as networks evolve. These are stated capabilities and goals; they are not quantified performance guarantees for every workload or deployment. 16
33
There is an example of mixed workloads operating together: Nokia says Elisa ran voice, mobile data and video analytics simultaneously on an AI-RAN platform in its Helsinki 5G network. That demonstrates a live-network use case, but it does not establish the scale or commercial economics of wider deployment. 42
Systems and availability across the network
The announcement describes systems for three network locations, with different roles and timelines:
- Distributed cell sites: The dual-node ARC-Pro pairs an Nvidia Grace C1 CPU and an Nvidia RTX PRO 4500 Blackwell Server Edition GPU in each node, in a 429 mm chassis. Supermicro specifies up to 1.6 PFLOPS of FP4 AI performance. Customer testing was planned for later in 2026, with general availability planned for the first half of 2027.
21
32
- Aggregation sites: The ARC MGX 2U is described with either dual Nvidia Vera CPUs or dual x86 processors, alongside two Nvidia RTX PRO 6000 Blackwell Server Edition GPUs. The dual-Vera configuration was shown at Nvidia GTC Berlin, October 20–22, 2026, and was expected to be available later that year.
- Core network hubs: Supermicro also identified an ARC MGX 4U system for core hubs. A contemporaneous report described support for up to eight RTX PRO 6000 Blackwell GPUs, but the supplied release excerpts do not provide a complete configuration or a precise availability date.
9
12
These timelines apply to the particular configurations described; they should not be read as one availability date for every ARC system. Nokia separately named Dell Technologies, Quanta Cloud Technology and Supermicro as supporting flexible ARC configurations, but the available material does not provide comparable product specifications or delivery dates for Dell and Quanta. 21
42
How the announcement fits into wider AI-RAN testing
The Supermicro collaboration landed alongside other AI-RAN announcements at GTC Berlin. Nvidia and Nokia said T-Mobile completed an AI-RAN call on a commercial network, while BT, Deutsche Telekom and Elisa announced pilots on live networks using Nokia’s CUDA-accelerated anyRAN software on Nvidia ARC. 43
Together with Elisa’s reported simultaneous voice, mobile-data and video-analytics workloads, these examples show activity in commercial-network testing and pilots. They do not, on their own, confirm broad paid rollouts, carrier contract values or revenue attributable to the Supermicro collaboration. 42
43
What the announcement means—and what it does not establish
Supermicro CEO Charles Liang described AI-RAN as turning mobile infrastructure into a shared foundation for connectivity, AI for RAN, AI on RAN and broader distributed AI services. That captures the strategic proposition: use network compute for both communications and selected AI workloads, rather than treating them as entirely separate infrastructure. 39
The cited reporting does not disclose customer contracts or revenue expectations tied to the collaboration. The announcement and pilots therefore signal a potential product and deployment opportunity, not a quantified commercial contribution. 12
Supermicro’s stock also had a separate earnings-related catalyst around the announcement: the company raised its fiscal first-quarter revenue guidance to $18.5 billion–$19.5 billion from $14.5 billion–$15.5 billion, and MarketBeat reported a 9.5% share-price rise. That reported move should not be treated as evidence that the AI-RAN collaboration generated the raised guidance; the cited report tied the revision to stronger AI-infrastructure demand. 4
Studio Global AI
Continue your research
This page includes a source-backed answer you can continue inside Studio Global.
People also ask
What is the short answer to "How Supermicro, Nokia and Nvidia Plan to Bring AI-RAN Across the Network"?
On October 21, 2026, Supermicro announced an AI RAN collaboration with Nokia and Nvidia that combines radio and AI workloads on shared infrastructure, from cell sites to core hubs.
What are the key points to validate first?
On October 21, 2026, Supermicro announced an AI RAN collaboration with Nokia and Nvidia that combines radio and AI workloads on shared infrastructure, from cell sites to core hubs. The design spans cell site, aggregation and core hub systems; a dual Vera 2U model was shown at GTC Berlin and was expected later in 2026.
What should I do next in practice?
T Mobile reported a commercial network AI RAN call, while BT, Deutsche Telekom and Elisa announced live network pilots; these are evidence of testing, not proof of large scale paid deployments.