AI for networks, security, sovereignty

Home RCR Wireless News AI for networks, security, sovereignty

A quick Friday news round-up, taking in telecoms and the wider AI infrastructure story. RCR Tech will also deal with the latter, so RCR Wireless shall deal with the former first. If you have been following Sean Kinney’s recent AI-RAN series, he has a new entry about brownfield open RAN transformation as a gateway to AI-RAN experimentation. Specifically, he discusses what Telus is doing with AI – for the RAN, in the first instance (with AI agents for certain operational tasks, to optimize flows and utilization), before AI in the RAN down the line. It has used an enforced hardware refresh to introduce virtualization, COTS compute, SMO, RIC, and programmable control – and, thus, ways to insert AI into its architecture. The Canadian firm’s approach is pragmatic, more credible than a wholesale rip-and-replace. Its key metric is spectrum efficiency, not AI for the sake of it. How do we deliver more data using the same spectrum asset? – it asks. The report works with Sean’s piece yesterday about T-Mobile’s AI-RAN twin-play in the US.

 

Otherwise, Deutsche Telekom has a de-riguer 4G/5G upgrade at a football stadium in Germany. Bayer 04 Leverkusen sounds better connected, for sure: 41 new antennas, connecting 30,000 fans on matchdays in prime low and mid-band spectrum, from 800 MHz to 3.6 GHz; five kilometers of new high-frequency cable; a shared set-up for other telcos; an ongoing small-cell densification project along routes to the stadium and in other areas. These upgrades are 10-a-penny these days, but the details are interesting. Meanwhile SpaceX is seeking more spectrum – as reported, and as discussed ad infinitum everywhere lately, including here. Also, Iliad Group in western Europe and United Group in southeastern Europe (in particular, which we haven’t covered much) have posted strong H1 results, highlighting the importance of regional challenger groups in the European telco landscape. Iliad continues to grow organically across France, Italy, and Poland, plus with sovereign cloud and AI infrastructure, while United is delivering double-digit growth across a portfolio of operators in Greece, Bulgaria, Croatia, and Slovenia.

 

Besides, on the straighter AI side, we have an OpenAI-led cybersecurity campaign, in the form of an open letter with a broad set of signatories (Anthropic, Arm, Broadcom, Cisco, Citi, DT, Equinix, General Motors, IBM, Lumen, Mastercard, Microsoft, NTT Data, Red Hat, SAP etc), to make cyber defense a collective effort as AI makes both attack and defense faster. There is a shrinking window in which defenders can use AI to harden systems before AI-enabled attacks become much more widespread and autonomous, they say. If AI can discover a vulnerability at machine speed, defenders need AI to develop a fix, test the fix, and deploy it at scale. The recent Hugging Face incident is clear enough. But OpenAI has a commercial interest, of course. The proposed solution is essentially to give defenders more capable models, such as its own. So it’s a commercial driver, as well, and the letter is advocacy, with broad recommendations about more government funding, information sharing, and corporate AI.

 

But the crisis is hardly in dispute. Corporate AI demand is outrunning supply forecasts. AWS had already planned more than a million Nvidia GPUs from 2026, and is now adding another two million for 2027/28 – because demand has exceeded expectations, notably among government agencies and big corporations, and with a keen eye on robotics besides. So what are we left with, this Friday lunch time? AI for AI networks, AI for AI security, AI for AI sovereignty, AI for AI robots – a self-perpetuating tech supply chain, and an infrastructure stack that must grow bigger, more distributed, more heterogeneous. AWS and Nvidia are also combining on CPUs, networking, memory, data processing, and robotics. The question for telcos, of course, is how that AI stack spills out of the data center and into the network, the factory, the warehouse, and the street. Telus and T-Mobile are asking what AI can do for spectrum utilization; but the telco questions are multiplying way beyond such parochial concerns.

James-Newsletter-rh9kgqsb2fng5qdzlyiq8g8ux9jouqcjh68bxs270g-10-rjao7r1w4lop2svtzn812vsuqvbme6iqfireyhnqps-rjyt8puypqg3izq2r8seeqbbi1pbceo4pqakrh4kxs-11

James Blackman
Executive Editor
RCR Wireless News

RCR Top Stories

Telus AI-RAN play: Telus is using its open RAN transformation to build a programmable foundation for AI, while testing automation, AI agents, and accelerated computing against real-world network performance.

SKT in AI rejig: SK Telecom is separating its AI data center infrastructure into SK Horizon and bringing in KRW 3.08 trillion from KKR and IMM to finance expansion across Korea.

Enterprise agents: Lightyear’s new Dispatch platform brings AI agents to enterprise telecom, automating carrier quoting and installation tracking while linking directly into expense management.

African AI infra: MTN is targeting 150MW of AI-ready data center capacity in South Africa and Nigeria, backed by a new partnership with UAE-based investor Tarek Al Ashram.

OpenAI Jalapeño: OpenAI has outlined performance of its Jalapeño chip, with the chip promising 1.7x lower latency and 1.9x higher tokens per watt than Nvidia GB200.

Logos SMCI NVIDIA 2021 2400x700 1 1 1
In partnership with

AI-Powered Telecom Infrastructure
Supermicro, in collaboration with NVIDIA, delivers AI-powered infrastructure tailored for telcos, enhancing operational efficiency, network management, and customer experiences. Explore now 

Beyond the Headlines

Gemini at Verizon: Verizon and Google Cloud are scaling their partnership into an enterprise-wide AI overhaul – from Gemini customer service to autonomous network agents with API access to live infrastructure.

AI-RAN paths diverge: Nokia and Ericsson have different AI-RAN paths, and T-Mobile US is testing both approaches – highlighting the split between incremental RAN intelligence and broader shared AI compute.

AI needs fabric: Compute orchestration and programmable connectivity are critical as AI inference workloads spread across AI factories, clouds, enterprises, telecom networks and the edge.

Regional AI fiber: Shentel has a hyperscaler deal for AI connectivity – a first for the regional carrier. The move is an early sign that AI demand is spilling out of the national networks into mid-tier regional fiber players.

Traffic management: Effective space traffic management is critical as debris and satellite constellations crowd low Earth orbit. But it will require tougher regulation, active debris removal, resilient satellite design, and robust cybersecurity.

What We're Reading

Starlink spectrum bid: SpaceX is seeking FCC approval to acquire three AWS-3 licenses from Tristar, covering about 4% of the US. The deal would expand spectrum for Starlink’s planned D2D satellite network.

DT’s match-day 5G: Deutsche Telekom has upgraded coverage at Bayer 04 Leverkusen’s BayArena with 41 new antennas, 14 ‘sectors’, and five kilometers of cabling. LTE and 5G capacity now supports more than 30,000 fans.

AWS nabs 2m GPUs: AWS is to deploy two million new Nvidia GPUs in 2027/28. The project also spans CPUs, networking, AI factories, open models, data processing and robotics, targeting enterprise and government workloads.

Iliad posts strong H1: Iliad claimed 3.4% growth in services revenue in H1, led by Italy and Poland; it also strengthened cash generation and reduced leverage. Its part-acquisition of SFR will make Iliad the EU’s third-largest telco. 

Double-digit gains: United Group saw revenue climb 11% to €1.45bn in H1 and adjusted EBITDAal up 12% to €487.5m. Its telecoms business grew across all its four markets in Southeast Europe.

Events

Virtual Program
Join industry experts for a day of strategic discussions, technical insights, and real-world deployment experiences as the ecosystem works to scale NTN integration and unlock new commercial opportunities worldwide. Register now
 
RCRTech’s Intelligent RAN Forum on September 22nd brings together industry experts to explore the technologies, strategies and real-world developments driving the evolution of intelligent networks. Register now
 
The virtual 6G Forum will tackle the most pressing questions relating to AI-native network architectures and service enablement as the industry moves towards the 6G era. Register now
 
Join industry stakeholders and innovative leaders in the communications service provider community this fall at CCA’s 2026 Annual Convention. Register now
 
Join 50 senior data center, energy and AI leaders at the Ritz-Carlton Dallas on October 21 for invitation-only roundtables on powering and scaling AI. Request your invitation 

What you need to know in 5 minutes

Join 37,000+ professionals receiving the AI Infrastructure Daily Newsletter

This field is for validation purposes and should be left unchanged.

This website uses cookies to improve your experience. We'll assume you're ok with this, but you can opt-out if you wish. Accept Read More