Enterprise & Telecom Technology Industry News
On the radio side, Verizon and Ericsson completed the first pre-standard 6G field call in the 7 GHz band from a commercial network site in California, moving 6G from the lab into a real-world environment. Aglocell reported that its AI-powered SpectraMax software delivered a 4.1% capacity gain across 60-plus sites for a Tier-1 U.S. carrier, adding one terabyte of daily throughput and avoiding $2.6 million in hardware costs. ZTE and Comsol strengthened their partnership to bring 5G and network AI to South African home broadband, Telit Cinterion and Commstar Radio teamed up to offer multi-network SIM connectivity to utilities and local governments, and Calnex and VIAVI expanded their partnership so teams can recreate real-world network conditions before deployment. The common thread was proving performance early, whether in the field, in live trials or in the lab, before capital is committed at scale.
AI News
Sovereign AI was the dominant theme. MeetKai announced a six-country rollout of its NVIDIA-built sovereign AI stack, reaching more than 700 million people across Brazil, Ukraine, Pakistan, Kazakhstan, Uzbekistan and Bangladesh, including Ukraine’s first national-scale AI factory. Bell and Cisco signed an MOU on sovereign AI infrastructure so Canadian organizations can run sensitive workloads with strict security and local resilience. Singtel’s RE:AI partnered with the Singapore Institute of Technology and the NVIDIA AI Technology Centre to speed enterprise deployments and strengthen Singapore’s applied AI talent pipeline, while Sweden celebrated the launch of Sferical One, a sovereign NVIDIA DGX SuperPOD supercomputer that provides secure in-country capacity. The pattern is consistent: nations want the full AI stack, from infrastructure to models, to stay under local control, and operators are positioned as the natural builders and hosts.
Governments and hyperscalers also put money behind AI research and infrastructure. NVIDIA committed $1 billion over five years to U.S. science and super intelligence R&D, targeting quantum computing, healthcare, energy security and national labs in support of the Genesis Mission, and AWS pledged up to $50 million in cloud credits over three years to the expanded Genesis Mission through its Golden Age of Science Accelerator. Ericsson’s CTO report on the path to an intelligent fabric argued that collaboration is crucial, with programmable, intelligent and autonomous networks meeting AI at scale. Huawei unveiled an AI-native core network architecture designed to manage an estimated 900 billion AI agents by 2035, and Oracle and Yucca Growth Infrastructure refined the Project Jupiter microgrid proposal to cut permitted emissions by 35% in response to regulator and public feedback. Power, permitting and community acceptance are increasingly part of the AI infrastructure story, not an afterthought.
Enterprise AI moved toward governed autonomy. Oracle announced Fusion Claw, a governed agentic execution runtime that combines AI reasoning with deterministic computation and powers 25 new applications, and separately applied agentic AI to financial crime investigations with Nexus Case Flow and Nexus Reach, helping institutions automate workflows while keeping humans in control. IFS launched IFS Intelligence to turn live operational data into insights that users can request in plain language, and AT&T began working with OpenAI on responsible, human-led legal workflows. DE-CIX released an open-source Model Context Protocol server so customers can discover, order and manage connectivity services with AI agents using natural language.
Governance and trust remained front of mind. An IBM study of 1,500 CFOs found that 62% say their role has expanded into technology or AI strategy leadership, and 56% expect to help design financial and ethical guardrails for AI by 2030. Veeam made Intelligent ResOps for Microsoft 365 generally available, giving customers visibility into how AI agents interact with data while cutting redundant storage, and Lean SuperIntelligence closed its pre-seed round and launched self-improving, air-gapped cybersecurity models, LSI-Offense-1 and LSI-Defense-1.
Security & Assurance News
Securing AI agents was the focus of this month’s security announcements. NVIDIA launched the Open Agent Safety Platform, an open software and reference design for governance from agent testing to deployment, and pledged a $250 million compute donation over five years to UK agentic safety and security research. IBM research found that 80% of EMEA executives saw AI-related incidents reach their board or C-suite, yet fewer than half have updated their AI governance policies. AXON Networks introduced AXON Datum, billed as the first data governance and AI sovereignty layer for telecoms, and Akeyless achieved FedRAMP High certification for its identity platform, which governs AI-agent access to systems. Marvell, Microsoft and Oracle formed an Open Compute Project working group to standardize an integrated hardware security module, with an initial draft due in October 2026, and Credence Research forecast the threat intelligence market will reach USD 24.1 billion by 2032, a 19.52% CAGR.
