Cisco’s Silicon One G300 as the dominant AI networking fabric, competing with Broadcom’s Tomahawk 6 series
On February 10, 2026, Cisco announced the Silicon One G300 102.4 Tbps Ethernet switch silicon, claiming it can power gigawatt-scale AI clusters for training, inference, and real-time agentic workloads, while maximizing GPU utilization with a 28% improvement in job completion time. The G300 was said to offer Intelligent Collective Networking, which combines an industry-leading fully shared packet buffer, path-based load balancing, and proactive network telemetry to offer better performance and profitability for large-scale data centers. It efficiently absorbs bursty AI traffic, responds faster to link failures, and prevents packet drops that can stall jobs, ensuring reliable data delivery even over long distances. With Intelligent Collective Networking, Cisco can deliver 33% increased network utilization, and a 28% reduction in job completion time versus simulated non-optimized path selection, making AI data centers more profitable with more tokens generated per GPU-hour. Also, the Cisco Silicon One G300 is highly programmable, enabling equipment to be upgraded for new network functionality even after it has been deployed. This enables Silicon One-based products to support emerging use cases and play multiple network roles, protecting long-term infrastructure investments. And with security fused into the hardware, customers can embrace holistic, at-speed security to keep clusters up and running.
The Cisco Silicon One G300 will power new Cisco N9000 and Cisco 8000 systems that push the frontier of AI networking in the data center. The systems feature innovative liquid cooling and support high-density optics to achieve new efficiency benchmarks and ensure customers get the most out of their GPU investments. In addition, the company enhanced Nexus One to make it easier for enterprises to operate their AI networks — on-premises or in the cloud — removing the complexity that can hold organizations back from scaling AI data centers.
“We are spearheading performance, manageability, and security in AI networking by innovating across the full stack – from silicon to systems and software,” said Jeetu Patel, President and Chief Product Officer, Cisco. “We’re building the foundation for the future of infrastructure, supporting every type of customer—from hyperscalers to enterprises—as they shift to AI-powered workloads.”
“As AI training and inference continues to scale, data movement is the key to efficient AI compute; the network becomes part of the compute itself. It’s not just about faster GPUs – the network must deliver scalable bandwidth and reliable, congestion-free data movement,” said Martin Lund, Executive Vice President of Cisco’s Common Hardware Group. “Cisco Silicon One G300, powering our new Cisco N9000 and Cisco 8000 systems, delivers high-performance, programmable, and deterministic networking – enabling every customer to fully utilize their compute and scale AI securely and reliably in production.”
The networking industry reaction to Cisco’s newest ASIC has been largely positive, with industry analysts and partners highlighting its role in reclaiming Cisco’s dominance in the AI infrastructure market. For example, Brendan Burke of Futurium thinks Cisco’s Silicon One G300 could be the backbone of Agentic AI Inference. His take: “Cisco’s latest announcements represent a calculated move to assert dominance in the AI networking fabric by attacking the specific bottlenecks of GPU cluster efficiency. As AI workloads shift toward agentic inference, where autonomous agents continuously interact across distributed environments, the network must handle unpredictable traffic patterns, unlike the structured flows of traditional training. Cisco is leveraging its vertical integration strategy to address the reliability and power constraints that plague these massive clusters. By emphasizing programmable silicon and rigorous optic qualification, Cisco aims to decouple network lifespan from rapid GPU innovation cycles, ensuring infrastructure can adapt to new traffic steering algorithms without hardware replacements. The G300 is a bid to make Ethernet the undisputed standard for AI back-end networks.”
- Industry-Leading Specs: Market analysts have noted that the G300’s 102.4 Tbps switching capacity sets a new benchmark for AI scale-out and scale-across networking.
- Efficiency Gains: Initial simulations showing a 28% reduction in job completion time (JCT) and a 33% increase in network utilization have been cited as major differentiators for large-scale AI clusters.
- Sustainability Focus: The shift toward liquid-cooled systems for the G300, which offers 70% greater energy efficiency per bit, is being viewed as a critical move for sustainable AI growth.
- Competitive Positioning: Experts from HyperFRAME Research suggest that the new silicon signals a “new confidence” from Cisco, positioning them as the “Apple of infrastructure” by tightly integrating hardware and software.
- AI Infrastructure Pivot: Financial analysts at Seeking Alpha have upgraded Cisco’s outlook, viewing the company no longer as just a legacy hardware firm but as a central player in the AI revolution.
- Partner Confidence: Major partners, such as Shanghai Lichan Technology, have expressed excitement about the Nexus 9100 Series powered by this silicon, specifically for its ability to simplify and scale AI deployments.
- Nvidia & Broadcom Competition: While the G300 is seen as a strong challenger to Nvidia’s Spectrum-X and Broadcom’s Tomahawk/Jericho lines, some observers note that Cisco still faces a steep climb to regain market share lost to these competitors in recent years.
- Complexity Concerns: Some industry veterans have pointed out that while the silicon is “hyperscale ready,” the success of these ASICs in the enterprise will depend on Cisco’s ability to maintain operational simplicity through tools like the Nexus Dashboard.
……………………………………………………………………………………………………………………………………………………………………………………………
|
Cisco Silicon One G300
|
Broadcom Tomahawk 6 (BCM78910 Series)
|
|---|---|
|
Bandwidth
102.4 Tbps  |
Bandwidth
102.4 Tbps  |
|
Manufacturing Process
TSMC 3nm  |
Manufacturing Process
3nm technology |
|
SerDes Lanes & Speed
512 lanes at 200 Gbps per link  |
SerDes Lanes & Speed
512 lanes at 200 Gbps per link, or 1024 lanes at 100G |
|
Port Configuration
Up to 64 x 1.6TbE ports or 512 x 200GbE ports |
Port Configuration
Up to 64 x 1.6TbE ports or 512 x 200GbE ports |
|
Target AI Cluster Size
Supports deployments of up to 128,000 GPUs |
Target AI Cluster Size
Supports over 100,000 XPUs (accelerators) Broadcom |
- Congestion Management: Cisco differentiates its G300 with an “Intelligent Collective Networking” approach featuring a fully shared packet buffer and a load-balancing agent that communicates across all G300s in the network to build a global map of congestion. Broadcom’s Tomahawk series also includes smart congestion control and global load balancing, though Cisco claims its implementation achieves higher network utilization (33% better).
- Programmability: Cisco emphasizes P4 programmability, allowing customers to update network functionality even after deployment.
- Ecosystem & Integration: Broadcom operates primarily in the merchant silicon market, with their chips used by various partners like HPE Juniper Networking. Cisco uses its own silicon to power its
Nexus 9000 and 8000 Series switches, tightly integrating hardware with software management platforms like Nexus One for a unified solution.
- Cooling Solutions: The Cisco G300 is designed to support high-density optics and is offered in new systems that include liquid-cooled options, providing 70% greater energy efficiency per bit compared to previous generations.
………………………………………………………………………………………………………………………………………………………………………………
References:
https://newsroom.cisco.com/c/r/newsroom/en/us/a/y2026/m02/cisco-announces-new-silicon-one-g300.html
https://blogs.cisco.com/sp/cisco-silicon-one-g300-the-next-wave-of-ai-innovation
Will Cisco’s Silicon One G300 Be the Backbone of Agentic Inference?
Analysis: Ethernet gains on InfiniBand in data center connectivity market; White Box/ODM vendors top choice for AI hyperscalers
Cisco CEO sees great potential in AI data center connectivity, silicon, optics, and optical systems
Networking chips and modules for AI data centers: Infiniband, Ultra Ethernet, Optical Connections
Nvidia enters Data Center Ethernet market with its Spectrum-X networking platform
Will AI clusters be interconnected via Infiniband or Ethernet: NVIDIA doesn’t care, but Broadcom sure does!
T-Mobile US announces new broadband wireless and fiber targets, 5G-A with agentic AI and live voice call translation
T-Mobile US (the “Un-carrier”) today announced new targets of 15 million 5G broadband customers by 2030, a 25% increase from its previous target of 12 million by the end of 2028, driven by increased spectral efficiency, better CPE technology, increased eligibility including to business customers with complementary usage profiles, and broadened product offerings to continue to meet evolving customer needs. T-Mobile is also leveraging its scale and nationwide 5G Advanced network to expand into new growth areas, including advertising, financial services, and long-term opportunities in edge and physical AI. The top rated U.S. wireless telco is also expecting between 3 and 4 million T-Fiber customers by 2030.
“T-Mobile is raising the bar on what customers, stockholders, and the industry can expect from the Un-carrier. T-Mobile has an unmatched combination of the Best Network, Best Value, and Best Customer Experiences — hallmarks of our unique Un-carrier differentiation — paired with our industry-leading portfolio of assets,” said T-Mobile CEO Srini Gopalan.
“This is why customers bring their connectivity relationship to T-Mobile. Looking ahead, we see an extraordinary runway to further expand this differentiation — through sustained momentum in network perception, digital and AI-driven transformation, and our future-forward innovation in areas like 6G and advanced AI. With this foundation, I’m confident that the future has never been brighter.”
Here are 2 of many impressive slides from T-Mo’s investor presentation referenced below:
The Un-carrier also plans to launch real-time and agentic AI services directly into its 5G-Advanced (5G-A) network by the end of 2026. This initiative, which began with a beta program in early 2026 for postpaid customers, allows for AI-driven features to function natively within the network, meaning users do not need to download specific apps or upgrade their hardware. This 5G-A offering will include live voice call translation in over 50 languages. By integrating AI directly into the 5G-A infrastructure (RAN, core network, and management layers), T-Mobile is enabling features that work on any eligible device, not just smartphones.
New 5G-A Agentic AI Highlights:
- The initial application is a “Live Translation” feature for voice calls, allowing for real-time translation in over 50 languages.
- “Agentic” AI and Automation: The network will use AI to enhance operational efficiency, including predictive optimization and dynamic resource allocation.
- The 5G-Advanced deployment also supports increased data speeds (up to 6.3 Gbps in tests), low-latency applications like XR and cloud gaming, and enhanced location services.
- The forthcoming capability will permit features to be active with only one participant needing to be on the 5G-A network.
- Infrastructure Partners: T-Mobile is collaborating with partners including NVIDIA, Ericsson, and Nokia to build an AI-RAN (Radio Access Network) framework. Telecompaper Telecompaper +3 This move is part of a broader strategy to transition from 5G to 5G-Advanced, with a focus on delivering “intent-driven” AI services and laying the groundwork for 6G (IMT 2030).
……………………………………………………………………………………………………………………………………….
References:
https://www.t-mobile.com/how-mobile-works/innovation/5g-advanced
https://www.t-mobile.com/news/business/t-mobile-capital-markets-day-update-feb-2026
https://www.t-mobile.com/benefits/live-translation
Analysis: Rakuten Mobile and Intel partnership to embed AI directly into vRAN
Today, Rakuten Mobile and Intel announced a partnership to embed Artificial Intelligence (AI) directly into the virtualized Radio Access Network (vRAN) stack. While vRAN currently represents a small percentage of the total RAN market (Dell’Oro Group recently forecasts vRAN to account for 5% to 10% of the total RAN market by 2026), this partnership could boost increase that percentage as it addresses key adoption hurdles—performance, power, and AI integration. Key areas of innovation include:
- Enhanced Wireless Spectral Efficiency: Optimizing spectrum utilization for superior network performance and capacity.
- Automated RAN Operations: Streamlining network management and reducing operational complexities through intelligent automation.
- Optimized Resource Allocation: Dynamically allocating network resources for maximum efficiency and subscriber experience.
- Increased Energy Efficiency: Significantly reducing power consumption in the RAN, contributing to sustainable network operations.
The partnership essentially aims to make vRAN superior in performance and TCO (Total Cost of Ownership) compared to traditional, proprietary, purpose built RAN hardware.
“We are incredibly excited to expand our collaboration with Intel to pioneer truly AI-native RAN architectures,” said Sharad Sriwastawa, co-CEO and CTO, Rakuten Mobile. “Together, we are validating transformative AI-driven innovations that will not only shape but define the future of mobile networks. This partnership showcases how intelligent RAN can be achieved through the seamless and efficient integration of AI workloads directly within existing vRAN software stacks, delivering unparalleled performance and efficiency.”
Rakuten Mobile and Intel are engaged in rigorous testing and validation of cutting-edge RAN AI use cases across Layer 1, Layer 2, and comprehensive RAN operation and network platform management. A core objective is the seamless integration of AI directly into the RAN stack, meticulously addressing integration challenges while upholding carrier-grade reliability and stringent latency requirements.
Utilizing Intel FlexRAN reference software, the Intel vRAN AI Development Kit, and a robust suite of AI tools and libraries, Rakuten Mobile is collaboratively training, optimizing, and deploying sophisticated AI models specifically tailored for demanding RAN workloads. This collaborative effort is designed to realize ultra-low, real-time AI latency on Intel Xeon 6 SoC, capitalizing on their built-in AI acceleration capabilities, including AVX512/VNNI and AMX.
“AI is transforming how networks are built and operated,” said Kevork Kechichian, Executive Vice President and General Manager of the Data Center Group, Intel Corporation. “Together with Rakuten, we are demonstrating how AI benefits can be achieved in vRAN. Intel Xeon processors power the majority of commercial vRAN deployments worldwide, and this transformation momentum continues to accelerate. Intel is providing AI-ready Xeon platforms that allow operators like Rakuten to design AI-ready infrastructure from the ground up, with built-in acceleration capabilities.”
Rakuten says they are “poised to unlock new levels of RAN performance, efficiency, and automation by embedding AI directly into the RAN software stack, this AI-native evolution represents the future of cloud-native, AI-powered RAN – inherently software-upgradable and built on open, general-purpose computing platforms. Additionally, the extended collaboration between Rakuten Mobile and Intel marks a significant step toward realizing the vision of autonomous, self-optimizing networks and powerfully reinforces both companies’ commitment to open, programmable, and intelligent RAN infrastructure worldwide.”
……………………………………………………………………………………………………………………………………………………………………..
- AI-Native Efficiency & Performance: The collaboration focuses on integrating AI to improve network performance and energy efficiency, which is a major pain point for operators. By embedding AI directly into the vRAN stack, they are enhancing wireless spectral efficiency, reducing power consumption, and automating RAN operations.
- Leveraging High-Performance Hardware: The initiative utilizes Intel® Xeon® 6 processors with built-in vRAN Boost. This eliminates the need for external, power-hungry accelerator cards, offering up to 2.4x more capacity and 70% better performance-per-watt.
- Validation of Large-Scale Commercial Viability: Rakuten Mobile operates the world’s first fully virtualized, cloud-native network. Its continued collaboration with Intel to make the vRAN AI-native provides a proven blueprint for other operators, reducing the perceived risk of adopting vRAN, particularly in brownfield (existing) networks.
- Acceleration of Open RAN Ecosystem: The collaboration supports the broader push towards Open RAN, which is expected to see a significant rise in market share, doubling between 2022 and 2026.
………………………………………………………………………………………………………………………………………………………………
- Market Share Shift: Omdia forecasts that vRAN’s share of the RAN baseband subsector will reach 20% by 2028. That’s a significant jump from its current low single-digit percentage.
- Explosive CAGR: The global vRAN market is projected to grow from approximately $16.6 billion in 2024 to nearly $80 billion by 2033, representing a 19.5% CAGR.
- Small Cell Dominance: By the end of 2026, it is estimated that 77% of all vRAN implementations will be on small cell architectures, a key area where Rakuten and Intel have demonstrated success.
References:
https://corp.mobile.rakuten.co.jp/english/news/press/2026/0210_01/
Virtual RAN gets a boost from Samsung demo using Intel’s Grand Rapids/Xeon Series 6 SoC
RAN silicon rethink – from purpose built products & ASICs to general purpose processors or GPUs for vRAN & AI RAN
vRAN market disappoints – just like OpenRAN and mobile 5G
LightCounting: Open RAN/vRAN market is pausing and regrouping
Dell’Oro: Private 5G ecosystem is evolving; vRAN gaining momentum; skepticism increasing
https://www.mordorintelligence.com/industry-reports/virtualized-ran-vran-market
https://www.grandviewresearch.com/industry-analysis/virtualized-radio-access-network-market-report
Virtualization’s role in 5G Advanced (3GPP Release 18) and a proposed new hardware architecture
Disclaimer: The author used Google Gemini to provide research contained in this article.
In a February 9, 2026 article, Ji-Yun Seol, Executive VP and Head of Product Strategy, Networks Business at Samsung, says: “The evolution from 5G to 5G-Advanced and 6G hinges on three interconnected pillars: virtualization for flexible networks, AI integration across all network layers, and automation towards autonomous networks.”
As the IEEE Techblog has extensively covered both AI RAN and the use of AI in 6G (IMT 2030), this post focuses on the role of virtualization in 5G Advanced.
In 3GPP Release 18 (5G-Advanced), virtualization is the foundational technology that enables several “software-defined” breakthroughs. 3GPP Release 18 components) have already been submitted to ITU-R WP 5D for inclusion in the next revision of ITU-R M.2150. Any remaining technical issues and the final decision for publication of ITU-R M.2150-3 are expected to be resolved during the WP 5D meeting concluding in Feb 2026.
3GPP Rel 18 features that depend most heavily on a virtualized, cloud-native architecture include:
- Host AI Models: Run complex machine learning algorithms for channel state information (CSI) feedback, beam management, and positioning.
- Automate Optimization: Enable “zero-touch” operations where the network dynamically adjusts power and resource allocation based on predictive traffic patterns.
While slicing existed in earlier releases, 5G-Advanced introduces more sophisticated, automated management. Virtualization is critical for:
Dynamic Resource Partitioning: Using Cloud-native Network Functions (CNFs) to create dedicated virtual networks on demand for specific use cases like Public Safety or industrial automation.
- SLA Assurance: Automatically scaling virtual resources to guarantee the ultra-low latency required for high-bandwidth applications like XR (Extended Reality).
To support lightweight headsets, 5G-Advanced relies on split-rendering.
- Edge Cloud Dependency: Virtualization allows heavy graphical processing to be moved from the headset to a virtualized Edge Cloud. This requires a highly agile, virtualized edge infrastructure to maintain the near-zero delay needed for immersive experiences.
- Infrastructure Visibility: New protocols provide the 3GPP layer with direct visibility into the underlying virtualized platform to detect vulnerabilities in the software-defined infrastructure.
Virtualization enables “self-organizing networks” (SON) where network entities can self-configure.
- Lifecycle Management: Standardized solutions in Rel-18 allow for the automated downloading, activation, and testing of software across virtualized network functions (VNFs).
| Feature | Primary Virtualization Dependency |
|---|---|
| AI/ML for RAN | Hosting and training models on COTS hardware |
| Edge-Based XR | Offloading computation to virtualized edge nodes |
| Automated Slicing | Rapid instantiation of CNFs for specific “slices” |
| Net Energy Saving | Software-driven power-down of virtual resources |
………………………………………………………………………………………………………………………………………………………………………………………………………………………………….
On the hardware side, traditional telecommunications infrastructure was defined by a tight coupling of network functions to proprietary, purpose-built hardware—resulting in siloed environments where routers, baseband units, and security appliances existed as distinct physical appliances. While providing reliable performance, this monolithic model introduced limitations in scalability, creating high demands for space, power, and capital expenditure for functional upgrades.
Virtualization transforms this paradigm by decoupling network functions from dedicated hardware, deploying them as software-defined workloads on commercial off-the-shelf (COTS) servers. This shift toward general-purpose compute platforms drives operational efficiency, enhances flexibility, and enables AI readiness. The industry adoption followed a staged evolution: starting with the virtualization of core networks—migrating packet gateways and subscriber databases to standard servers—followed by Virtualized RAN (vRAN), which disaggregates baseband processing from radio hardware to operate as cloud-native software.
In 5G-Advanced (Release 18), the hardware shifts from proprietary “black boxes” to a disaggregated architecture of General-Purpose Processors (GPPs) and Specialized Accelerators.
The physical infrastructure required to run these virtualized functions generally falls into three categories:
Virtual Network Functions (VNFs) and Cloud-native Functions (CNFs) run on Commercial Off-The-Shelf (COTS) servers designed for high-density environments.
- Processors: Typically
Intel Xeon Scalable or AMD EPYC processors with high CPU core counts (up to 48+ cores) to handle parallelized workloads.
- Memory: Large-scale deployments require 384GB to over 1TB of DDR4/DDR5 RAM to support multiple network “slices” simultaneously.
- Form Factor: Short-depth chassis (300mm to 600mm) to fit into standard telco racks or outdoor cabinets at the network edge.
- Inline vs. Lookaside:
- Lookaside: The CPU sends specific tasks (like Forward Error Correction) to the card and gets them back.
- Inline: The entire L1 data flow passes through the accelerator, reducing the load on the CPU and improving power efficiency.
- Chips: These cards use FPGAs (Field Programmable Gate Arrays), ASICs (Application-Specific Integrated Circuits), or GPUs.
- GPU Integration: Platforms like NVIDIA Aerial use GPUs to accelerate both 5G signal processing and AI workloads on the same hardware.
- DPUs (Data Processing Units): Used to offload networking and security tasks, ensuring that data moves between the radio and the virtualized core with sub-microsecond precision.
| Hardware Component | Function in 5G-Advanced |
|---|---|
| COTS Servers | Host virtualized core and RAN software (vCU, vDU) |
| L1 Accelerators | Handle compute-heavy signal processing (Beamforming, MIMO) |
| SmartNICs / DPUs | Manage high-speed data transfer and timing synchronization |
| GPUs | Power the AI/ML models for network optimization and XR rendering |
…………………………………………………………………………………………………………………………………………………………………..
References:
https://www.3gpp.org/specifications-technologies/releases/release-18
Samsung: Turning legacy infrastructure into AI-ready networks
Dell’Oro: Analysis of the Nokia-NVIDIA-partnership on AI RAN
RAN silicon rethink – from purpose built products & ASICs to general purpose processors or GPUs for vRAN & AI RAN
Comparing AI Native mode in 6G (IMT 2030) vs AI Overlay/Add-On status in 5G (IMT 2020)
Nvidia CEO Huang: AI is the largest infrastructure buildout in human history; AI Data Center CAPEX will generate new revenue streams for operators
Executive Summary:
In a February 6, 2026 CNBC interview with with Scott Wapner, Nvidia CEO Jensen Huang [1.] characterized the current AI build‑out as “the largest infrastructure buildout in human history,” driven by exceptionally high demand for compute from hyperscalers and AI companies. “Through the roof” is how he described AI infrastructure spending. It’s a “once-in-a-generation infrastructure buildout,” specifically highlighting that demand for Nvidia’s Blackwell chips and the upcoming Vera Rubin platform is “sky-high.” He emphasized that the shift from experimental AI to AI as a fundamental utility has reached a definitive inflection point for every major industry.
Jensen forecasts aa roughly 7–to- 8‑year AI investment cycle lies ahead, with capital intensity justified because deployed AI infrastructure is already generating rising cash flows for operators. He maintains that the widely cited ~$660 billion AI data center capex pipeline is sustainable, on the grounds that GPUs and surrounding systems are revenue‑generating assets, not speculative overbuild. In his view, as long as customers can monetize AI workloads profitably, they will “keep multiplying their investments,” which underpins continued multi‑year GPU demand, including for prior‑generation parts that remain fully leased.
Note 1. Being the undisputed leader of AI hardware (GPU chips and networking equipment via its Mellanox acquisition), Nvidia MUST ALWAYS MAKE POSITIVE REMARKS AND FORECASTS related to the AI build out boom. Reader discretion is advised regarding Huang’s extremely bullish, “all-in on AI” remarks.
Huang reiterated that AI will “fundamentally change how we compute everything,” shifting data centers from general‑purpose CPU‑centric architectures to accelerated computing built around GPUs and dense networking. He emphasizes Nvidia’s positioning as a full‑stack infrastructure and computing platform provider—chips, systems, networking, and software—rather than a standalone chip vendor. He accuratedly stated that Nvidia designs “all components of AI infrastructure” so that system‑level optimization (GPU, NIC, interconnect, software stack) can deliver performance gains that outpace what is possible with a single chip under a slowing Moore’s Law. The installed base is presented as productive: even six‑year‑old A100‑class GPUs are described as fully utilized through leasing, underscoring persistent elasticity of AI compute demand across generations.
AI Poster Childs – OpenAI and Anthropic:
Huang praised OpenAI and Anthropic, the two leading artificial intelligence labs, which both use Nvidia chips through cloud providers. Nvidia invested $10 billion in Anthropic last year, and Huang said earlier this week that the chipmaker will invest heavily in OpenAI’s next fundraising round.
“Anthropic is making great money. Open AI is making great money,” Huang said. “If they could have twice as much compute, the revenues would go up four times as much.”
He said that all the graphics processing units that Nvidia has sold in the past — even six-year old chips such as the A100 — are currently being rented, reflecting sustained demand for AI computing power.
“To the extent that people continue to pay for the AI and the AI companies are able to generate a profit from that, they’re going to keep on doubling, doubling, doubling, doubling,” Huang said.
Economics, utilization, and returns:
On economics, Huang’s central claim is that AI capex converts into recurring, growing revenue streams for cloud providers and AI platforms, which differentiates this cycle from prior overbuilds. He highlights very high utilization: GPUs from multiple generations remain in service, with cloud operators effectively turning them into yield‑bearing infrastructure.
This utilization and monetization profile underlies his view that the capex “arms race” is rational: when AI services are profitable, incremental racks of GPUs, network fabric, and storage can be modeled as NPV‑positive infrastructure projects rather than speculative capacity. He implies that concerns about a near‑term capex cliff are misplaced so long as end‑market AI adoption continues to inflect.
Competitive and geopolitical context:
Huang acknowledges intensifying global competition in AI chips and infrastructure, including from Chinese vendors such as Huawei, especially under U.S. export controls that have reduced Nvidia’s China revenue share to roughly half of pre‑control levels. He frames Nvidia’s strategy as maintaining an innovation lead so that developers worldwide depend on its leading‑edge AI platforms, which he sees as key to U.S. leadership in the AI race.
He also ties AI infrastructure to national‑scale priorities in energy and industrial policy, suggesting that AI data centers are becoming a foundational layer of economic productivity, analogous to past buildouts in electricity and the internet.
Implications for hyperscalers and chips:
Hyperscalers (and also Nvidia customers) Meta , Amazon, Google/Alphabet and Microsoft recently stated that they plan to dramatically increase spending on AI infrastructure in the years ahead. In total, these hyperscalers could spend $660 billion on capital expenditures in 2026 [2.] , with much of that spending going toward buying Nvidia’s chips. Huang’s message to them is that AI data centers are evolving into “AI factories” where each gigawatt of capacity represents tens of billions of dollars of investment spanning land, compute, and networking. He suggests that the hyperscaler industry—roughly a $2.5 trillion sector with about $500 billion in annual capex transitioning from CPU to GPU‑centric generative AI—still has substantial room to run.
Note 2. An understated point is that while these hyperscalers are spending hundered of billions of dollars on AI data centers and Nvidia chips/equipment they are simultaneously laying off tens of thousands of employees. For example, Amazon recently announced 16,000 job cuts this year after 14,000 layoffs last October.
From a chip‑level perspective, he argues that Nvidia’s competitive moat stems from tightly integrated hardware, networking, and software ecosystems rather than any single component, positioning the company as the systems architect of AI infrastructure rather than just a merchant GPU vendor.
References:
Big tech spending on AI data centers and infrastructure vs the fiber optic buildout during the dot-com boom (& bust)
Analysis: Cisco, HPE/Juniper, and Nvidia network equipment for AI data centers
Networking chips and modules for AI data centers: Infiniband, Ultra Ethernet, Optical Connections
Will billions of dollars big tech is spending on Gen AI data centers produce a decent ROI?
Superclusters of Nvidia GPU/AI chips combined with end-to-end network platforms to create next generation data centers
184K global tech layoffs in 2025 to date; ~27.3% related to AI replacing workers
SNS Telecom & IT: Mission-Critical Networks a $9.2 Billion Market
For nearly a century, the critical communications industry has relied on narrowband LMR networks for mission-critical voice and low-speed data services. Over time, these systems have evolved from relatively basic analog radios to digital communications technologies, such as APCO P25 and TETRA, to provide superior voice quality, end-to-end encryption, and other advanced features. However, due to their inherent bandwidth and design limitations, even the most sophisticated digital LMR networks are unable to support mobile broadband and data-driven critical IoT applications that have become vital for public safety, defense, utilities, transportation, oil and gas, mining, and other segments of the critical communications industry.
The 3GPP-defined LTE and 5G NR air interfaces have emerged as the leading radio access technology candidates to fill this void. Over the last decade, a plethora of fully dedicated, hybrid commercial-private, and secure MVNO-based 3GPP networks have been deployed to deliver critical communications broadband capabilities – in addition to the use of commercial mobile operator networks – for application scenarios as diverse as PTT group communications, multimedia messaging, high-definition video surveillance, BVLOS (Beyond Visual Line-of-Sight) operation of drones, situational awareness, untethered AR/VR/MR, collaborative mobile robots, AGVs (Automated Guided Vehicles), and automation in IIoT (Industrial IoT) environments. These networks range from nationwide PPDR (Public Protection & Disaster Relief) broadband platforms such as the United States’ FirstNet, South Korea’s Safe-Net, Saudi Arabia’s mission-critical broadband network, Great Britain’s ESN, France’s RRF, Sweden’s SWEN, and Finland’s VIRVE 2 public safety broadband service to defense sector 5G programs for the adoption of tactical cellular systems and permanent private 5G networks at military bases, regional cellular networks covering the service footprint of utility companies, FRMCS (Future Railway Mobile Communication System)-ready networks for train-to-ground communications, and NPNs (Non-Public Networks) for localized wireless connectivity in settings such as airports, maritime ports, oil and gas production facilities, power plants, substations, offshore wind farms, remote mining sites, factories, and warehouses.
Historically, most critical communications user organizations have viewed LTE and 5G NR as complementary technologies, used primarily to augment existing voice-centric LMR networks with broadband capabilities. This perception has changed with the commercial availability of 3GPP standards-compliant MCX (Mission-Critical PTT, Video & Data), QPP (QoS, Priority & Preemption), HPUE (High-Power User Equipment), IOPS (Isolated Operation for Public Safety), URLLC (Ultra-Reliable, Low-Latency Communications), TSC (Time-Sensitive Communications), and related service enablers. LTE and 5G networks have gained recognition as an all-inclusive critical communications platform and are nearing the point where they can fully replace legacy LMR systems with a future-proof transition path, supplemented by additional 5G features, such as 5G MBS/5MBS (5G Multicast-Broadcast Services) for MCX services in high-density environments, 5G NR sidelink for off-network communications, VMRs (Vehicle-Mounted Relays), MWAB (Mobile gNB With Wireless Access Backhauling), satellite NTN (Non-Terrestrial Network) integration, and support for lower 5G NR bandwidths in dedicated frequency bands for PPDR, utilities, and railways.
SNS Telecom & IT’s LTE & 5G for Critical Communications: 2025 – 2030 research publication projects that global investments in mission-critical 3GPP networks and associated applications reached $5.4 billion in 2025. Driven by public safety broadband, defense communications, smart grid modernization, FRMCS, and IIoT initiatives, the market is expected to grow at a CAGR of approximately 19% over the next three years, eventually accounting for more than $9.2 billion by the end of 2028. Looking ahead to 2030, the industry will be underpinned by operational deployments ranging from sub-1 GHz wide area networks for national-scale MCX services, utility communications, and GSM-R replacement to systems operating in mid-band spectrum such as Band n101 (1.9 GHz) and Band n79 (4.4-5 GHz), as well as mmWave (Millimeter Wave) frequencies for specialized applications.
Image Credit: SNS Telecom & IT
…………………………………………………………………………………………………………….
About SNS Telecom & IT
SNS Telecom & IT is a global market intelligence and consulting firm with a primary focus on the telecommunications and information technology industries. Developed by in-house subject matter experts, our market intelligence and research reports provide unique insights on both established and emerging technologies. Our areas of coverage include but are not limited to 6G, 5G, LTE, Open RAN, vRAN, small cells, mobile core, xHaul transport, network automation, mobile operator services, FWA, neutral host networks, private 4G/5G cellular networks, public safety broadband, critical communications, MCX, IIoT, V2X communications, and vertical applications.
References:
https://www.snstelecom.com/lte-for-critical-communications
SNS Telecom & IT: Private LTE & 5G Network Ecosystem – CAGR 22% from 2025-2030
SNS Telecom & IT: Private 5G and 4G LTE cellular networks for the global defense sector are a $1.5B opportunity
Fiber Optic Networks & Subsea Cable Systems as the foundation for AI and Cloud services
Emerging Cybersecurity Risks in Modern Manufacturing Factory Networks
RtBrick survey: Telco leaders warn AI and streaming traffic to “crack networks” by 2030
Alphabet’s 2026 capex forecast soars; Gemini 3 AI model is a huge success
Google parent company Alphabet forecast 2026 capital expenditures of $175 billion to $185 billion this year, a massive jump compared with average analyst expectations that it would spend about $115.26 billion this year, according to data compiled by LSEG. The midpoint capex forecast of $180 billion was well above the $119.5 billion projected by analysts tracked by Bloomberg. Alphabet’s fourth quarter capex of $27.9 billion was slightly less than the expected $28.2 billion for the period, per Bloomberg estimates.
“We’re seeing our AI investments and infrastructure drive revenue and growth across the board,” CEO Sundar Pichai said in the company’s press release. He said the higher 2026 spending would allow the company “to meet customer demand and capitalize on the growing opportunities.”
The release of Google’s Gemini 3 AI model — which outperformed competing models on benchmark tests and prompted rival OpenAI to declare a “code red” — as well as the announcement of a landmark deal with Apple, cemented Alphabet’s position as an AI winner. Google Gemini gained significant market share against ChatGPT. This changed the AI landscape from a near-monopoly to a more competitive duopoly. Although ChatGPT still leads in total traffic, Gemini’s growth has narrowed the gap. This growth is due Gemini’s integration into Google’s ecosystem, especially Chrome, Android phones/tablets and Google Cloud.
“The launch of Gemini 3 was a major milestone and we have great momentum,” Pichai noted. He added that the Gemini app now has more than 750 million monthly active users.
In a related comment, Emarketer analyst Nate Elliott wrote:
“Gemini continues to grow quickly, from 650 million monthly active users at the end of Q3 to 750 million at the end of the year. But it’s worth noting that Gemini’s user number grew only about one-third as fast in Q4 as in Q3. That might explain why Google continues to keep its flagship AI tool advertising-free, hoping the lack of ads makes it more attractive to users than ChatGPT. It also explains why the company is now more aggressively pushing search users from AI Overviews into AI Mode: it’s looking for additional avenues to increase usage of its full-fledged AI chatbots.”
Google offices in Mountain View, Calif. (Reuters/Manuel Orbegozo) · Reuters / Reuters
Alphabet’s 4th quarter revenue climbed 18% to $113.8 billion from the year-ago period, ahead of the $111.4 billion expected by analysts. The tech giant’s earnings per share rose to $2.82 from $2.15 in the previous year, also higher than the $2.65 projected. The big increase in sales was spurred by a 48% spike in Google Cloud revenue to $17.7 billion, more than the $16.2 billion expected by analysts.
………………………………………………………………………………………………………………………………………………………
Update:
“Both Alphabet and Amazon delivered strong underlying business performance, driven by better-than-expected growth in cloud. But that hasn’t been enough to distract markets from their ballooning capital investment plans,” said Aarin Chiekrie, equity analyst, Hargreaves Lansdown.
References:
https://s206.q4cdn.com/479360582/files/doc_financials/2025/q4/2025q4-alphabet-earnings-release.pdf
Google announces Gemini: it’s most powerful AI model, powered by TPU chips
Comparing AI Native mode in 6G (IMT 2030) vs AI Overlay/Add-On status in 5G (IMT 2020)
Analysis: Edge AI and Qualcomm’s AI Program for Innovators 2026 – APAC for startups to lead in AI innovation
China’s open source AI models to capture a larger share of 2026 global AI market
China gaining on U.S. in AI technology arms race- silicon, models and research
IDC Report: Telecom Operators Turn to AI to Boost EBITDA Margins
AI spending boom accelerates: Big tech to invest an aggregate of $400 billion in 2025; much more in 2026!
Amazon’s Jeff Bezos at Italian Tech Week: “AI is a kind of industrial bubble”
OpenAI and Broadcom in $10B deal to make custom AI chips
AI Frenzy Backgrounder; Review of AI Products and Services from Nvidia, Microsoft, Amazon, Google and Meta; Conclusions
Will billions of dollars big tech is spending on Gen AI data centers produce a decent ROI?
China’s telecom industry rapid growth in 2025 eludes Nokia and Ericsson as sales collapse
According to a Chinese government update, “Telecommunications business volume and revenue grew steadily, mobile internet access traffic maintained rapid growth, and the construction of network infrastructure such as 5G, gigabit optical networks, and the Internet of Things was further promoted.”
Figure 1. Cumulative growth rate of telecommunications service revenue and total telecommunications service volume
There were 4.83 million 5G base stations in service in China at the end of November 2025, an increase of 579,000 since late 2024 and 37.4% of the total number of mobile base stations in China. In one year, China claims to have added more 5G base stations than Europe has installed since the 5G technology was first put into service.
The total number of mobile phone users of the top four Chinese telcos (China Mobile, China Telecom, China Unicom, China Broadcasting Network) reached 1.828 billion, a net increase of 38.54 million from the end of last year. Among them, 5G mobile phone users reached 1.193 billion, a net increase of 179 million from the end of last year, accounting for 65.3% of all mobile phone users.
Meanwhile, the total number of fixed broadband internet access users of the three state owned telecom operators (China Mobile, China Telecom and China Unicom) reached 697 million, a net increase of 27.12 million from the end of last year. Among them, fixed broadband internet access users with access speeds of 100Mbps and above reached 664 million, accounting for 95.2% of the total users; fixed broadband internet access users with access speeds of 1000Mbps and above reached 239 million, a net increase of 32.52 million from the end of last year, accounting for 34.3% of the total users, an increase of 3.4 percentage points from the end of last year.
The construction of gigabit fiber optic broadband networks continues to advance. As of the end of November, the number of broadband internet access ports nationwide reached 1.25 billion, a net increase of 48.11 million compared to the end of last year. Among them, fiber optic access (FTTH/O) ports reached 1.21 billion, a net increase of 49.42 million compared to the end of last year, accounting for 96.8% of all broadband internet access ports. As of the end of November, the number of 10G PON ports with gigabit network service capabilities reached 31.34 million, a net increase of 3.133 million compared to the end of last year.
The penetration rate of gigabit and 5G users continued to increase across all regions. As of the end of November, the penetration rates of fixed broadband access users with speeds of 1000Mbps and above in the eastern, central, western, and northeastern regions were 34.6%, 33.8%, 35.8%, and 28.5%, respectively, representing increases of 3.4, 2.6, 4.1, and 4.9 percentage points compared to the end of last year; the penetration rates of 5G mobile phone users were 64.9%, 65.9%, 65.1%, and 65.9%, respectively, representing increases of 8.2, 8.7, 8.8, and 9.6 percentage points compared to the end of last year.
…………………………………………………………………………………………………………………………………………………………………
Separately, Light Reading reports that Ericsson and Nokia sales of networking equipment to China have collapsed.
Ericsson recently published earnings release for the final quarter of 2025 puts China revenues at just 3% of total sales last year. This would equate to revenues of 7.1 billion Swedish kronor (US$798 million). Based on a rounding range of 2.5% to 3.4%, it works out to be between SEK5.92 billion ($665 million) and SEK8.05 billion ($905 million) – down sharply compared with the SEK10.2 billion ($1.15 billion) Ericsson made in 2024, according to that year’s Ericsson annual report.
Nokia does not break out details of revenues from mainland China, instead lumping them together with the sales it generates in neighboring Hong Kong and Taiwan. But this “Greater China” business is in decline. Total annual revenues – which include Nokia’s sales of fixed, Internet Protocol and optical network products, as well as 5G – slumped from almost €2.2 billion ($2.6 billion) in 2019 to around €1.5 billion ($1.8 billion) in 2020, before creeping back up to nearly €1.6 billion ($1.9 billion) by 2022. Two years later, they had fallen to about €1.1 billion ($1.3 billion).
Bar Chart Credit: Light Reading
Nokia has recently indicated the complete disappearance of its China business. “Western suppliers, which is only us and Ericsson, have 3% market share now in China and it’s been coming down, and we are going to be excluded from China for national security reasons,” said Tommi Uitto, the former president of Nokia’s mobile networks business group, at a September press conference in Finland also attended by Justin Hotard, Nokia’s CEO. It implies China’s government is now treating the Nordic vendors in the same way Europe and the U.S. are banning Huawei and ZTE networking equipment.
Nokia revealed in its latest earnings update that Greater China revenues for 2025 had fallen by another 19%, to €913 million ($1.08 billion) – just 42% of what Nokia earned in the region seven years earlier. In the last few years, moreover, Nokia has cut more jobs in Greater China than in any other single region. While figures are not yet available for 2025, the Greater China headcount numbered 8,700 employees in 2024, down from 15,700 in 2019.
Ericsson has significantly reduced its China operations following greatly reduced 5G market share. In September 2021, the company consolidated three operator-specific customer units into a unified structure, impacting several hundred sales and delivery roles within its ~10,000-person local workforce. This followed the divestment of a Nanjing-based R&D center (approx. 650 employees), aligning with strategic pivots away from legacy 2G-4G technologies. The company’s total workforce in Northeast Asia plummeted from about 14,000 in mid-2021 to roughly 9,500 at the end of last year, according to Ericsson’s financial statements.
Exclusion from China would leave Ericsson and Nokia on the outside of the world’s most promising 6G market in 2030. That would intensify concern about a bifurcation of 6G into Western and Chinese variants of IMT 20230 RIT/SRIT standard and the 3GPP specified 6G core network.
References:
https://www.miit.gov.cn/gxsj/tjfx/txy/art/2025/art_7514154ec01c42ecbcb76057464652e4.html
https://www.lightreading.com/5g/ericsson-and-nokia-see-their-sales-in-china-fall-off-a-cliff
China’s open source AI models to capture a larger share of 2026 global AI market
Goldman Sachs: Big 3 China telecom operators are the biggest beneficiaries of China’s AI boom via DeepSeek models; China Mobile’s ‘AI+NETWORK’ strategy
China Telecom’s 2025 priorities: cloud based AI smartphones (?), 5G new calling (GSMA), and satellite-to-phone services
China ITU filing to put ~200K satellites in low earth orbit while FCC authorizes 7.5K additional Starlink LEO satellites
China gaining on U.S. in AI technology arms race- silicon, models and research
Analysis: Edge AI and Qualcomm’s AI Program for Innovators 2026 – APAC for startups to lead in AI innovation
Qualcomm is a strong believer in Edge AI as an enabler of faster, more secure, and energy-efficient processing directly on devices—rather than the cloud—unlocking real-time intelligence for industries like robotics and smart cities.
In support of that vision, the fabless SoC company announced the official launch of its Qualcomm AI Program for Innovators (QAIPI) 2026 – APAC, a regional startup incubation initiative that supports startups across Japan, Singapore, and South Korea in advancing the development and commercialization of innovative edge AI solutions.
Building on Qualcomm’s commitment to edge AI innovation, the second edition of QAIPI-APAC invites startups to develop intelligent solutions across a broad range of edge-AI applications using Qualcomm Dragonwing™ and Snapdragon® platforms, together with the new Arduino® UNO Q development board, strengthening their pathway toward global commercialization.
Startups gain comprehensive support and resources, including access to Qualcomm Dragonwing™ and Snapdragon® platforms, the Arduino® UNO Q development board, technical guidance and mentorship, a grant of up to US$10,000, and eligibility for up to US$5,000 in patent filing incentives, accelerating AI product development and deployment.
Applications are open now through April 30, 2026 and will be evaluated based on innovation, technical feasibility, potential societal impact, and commercial relevance. The program will be implemented in two phases. The application phase is open to eligible startups incorporated and registered in Japan, Singapore, or South Korea. Shortlisted startups will enter the mentorship phase, receiving one-on-one guidance, online training, technical support, and access to Qualcomm-powered hardware platforms and development kits for product development. They will also receive a shortlist grant of up to US$10,000 and may be eligible for a patent filing incentive of up to US$5,000. At the conclusion of the program, shortlisted startups may be invited to showcase their innovations at a signature Demo Day in late 2026, engaging with industry leaders, investors, and potential collaborators across the APAC innovation ecosystem.
Comment and Analysis:
Qualcomm is a strong believer in Edge AI—the practice of running AI models directly on devices (smartphones, cars, IoT, PCs) rather than in the cloud—because they view it as the next major technological paradigm shift, overcoming limitations inherent in cloud computing. Despite the challenges of power consumption and processing limitations, Qualcomm’s strategy hinges on specialized, heterogenous computing rather than relying solely on RISC-based CPU cores.
Key Issues for Qualcomm’s Edge AI solutions:
- Qualcomm® AI Engine: This combines specialized hardware, including the Hexagon NPU (Neural Processing Unit), Adreno GPU, and CPU. The NPU is specifically designed to handle high-performance, complex AI workloads (like Generative AI) far more efficiently than a generic CPU.
- Custom Oryon CPU: The latest Snapdragon X Elite platform features customized cores that provide high performance while outperforming traditional x86 solutions in power efficiency for everyday tasks.
- Specialization Saves Power: By using specialized AI engines (NPUs) rather than general-purpose CPU/GPU cores, Qualcomm can run inference tasks at a fraction of the power cost.
- Lower Overall Energy: Doing AI at the edge can save total energy by avoiding the need to send data to a power-hungry data center, which requires network infrastructure, and then sending it back.
- Intelligent Efficiency: The Snapdragon 8 Elite, for example, saw a 27% reduction in power consumption while increasing AI performance significantly.
- Instant Responsiveness (Low Latency): For autonomous vehicles or industrial robotics, a few milliseconds of latency to the cloud can be catastrophic. Edge AI provides real-time, instantaneous analysis.
- Privacy and Security: Data never leaves the device. This is crucial for privacy-conscious users (biometrics) and compliance (GDPR), which is a major advantage over cloud-based AI.
- Offline Capability: Edge devices, such as agricultural sensors or smart home devices in remote areas, continue to function without internet connectivity.
- Diversification: With the smartphone market maturing, Qualcomm sees the “Connected Intelligent Edge” as a huge growth opportunity, extending their reach into automotive, IoT, and PCs.
- “Ecosystem of You”: Qualcomm aims to connect billions of devices, making AI personal and context-aware, rather than generic.
- Qualcomm AI Hub: This makes it easier for developers to deploy optimized models on Snapdragon devices.
- Model Optimization: They specialize in making AI models smaller and more efficient (using quantization and specialized AI inference) to run on devices without requiring massive, cloud-sized computing power.
References:
Qualcomm CEO: AI will become pervasive, at the edge, and run on Snapdragon SoC devices
Huawei, Qualcomm, Samsung, and Ericsson Leading Patent Race in $15 Billion 5G Licensing Market
Private 5G networks move to include automation, autonomous systems, edge computing & AI operations
Nvidia’s networking solutions give it an edge over competitive AI chip makers
Nvidia AI-RAN survey results; AI inferencing as a reinvention of edge computing?
CES 2025: Intel announces edge compute processors with AI inferencing capabilities
Qualcomm CEO: expect “pre-commercial” 6G devices by 2028
SoftBank and Ericsson-Japan achieve 24% 5G throughput improvement using AI-optimized Massive MIMO
SoftBank Corp. and Ericsson Japan K.K. have announced a successful demonstration and deployment of an AI-powered, externally controlled optimization system for Massive MIMO, resulting in a 24% improvement in 5G downlink throughput, increasing speeds from 76.9 Mbps to 95.5 Mbps during periods of high traffic fluctuation.
- Dynamic Beam Patterns: The system automatically adjusts horizontal and vertical beam patterns every minute based on real-time user distribution.
- Packet Stalling Mitigation: By reacting to sudden traffic surges (e.g., during fireworks or concerts), the AI helps prevent “packet stalling,” where data transmission typically freezes due to congestion.
- Commercial Deployment: Following the successful trials at Expo 2025, SoftBank and Ericsson have begun deploying this AI-based system at other large-scale event venues, including major arenas and dome-type facilities in the Tokyo metropolitan area, to manage heavily fluctuating traffic patterns.
Overview of the System:
- An external control device (server) uses user distribution data collected from base stations at one-minute intervals to automatically determine event occurrence using AI
- Dynamically and automatically optimizes the horizontal and vertical coverage patterns of Massive MIMO base stations
Overview of demonstration at Expo 2025:
An AI model was constructed using performance results obtained when multiple coverage patterns were changed in advance as training data. Based on user distribution-related data such as Massive MIMO beam estimation information*2 acquired by an external control device from base stations at one-minute intervals, the system automatically determined event occurrence status and switched base station coverage patterns to optimal configurations.
……………………………………………………………………………………………………………………………….
ABOUT SOFTBANK:
Guided by the SoftBank Group’s corporate philosophy, “Information Revolution – Happiness for everyone,” SoftBank Corp. (TOKYO: 9434) operates telecommunications and IT businesses in Japan and globally. Building on its strong business foundation, SoftBank Corp. is expanding into non-telecom fields in line with its “Beyond Carrier” growth strategy while further growing its telecom business by harnessing the power of 5G/6G, IoT, Digital Twin and Non-Terrestrial Network (NTN) solutions, including High Altitude Platform Station (HAPS)-based stratospheric telecommunications. While constructing AI data centers and developing homegrown LLMs specialized for the Japanese language, SoftBank is integrating AI with radio access networks (AI-RAN), with the aim of becoming a provider of next-generation social infrastructure. To learn more, please visit https://www.softbank.jp/en/corp/
………………………………………………………………………………………………………………………………….
References:
SoftBank’s Transformer AI model boosts 5G AI-RAN uplink throughput by 30%, compared to a baseline model without AI
Softbank developing autonomous AI agents; an AI model that can predict and capture human cognition
Nokia, BT Group & Qualcomm achieve enhanced 5G SA downlink speeds using 5G Carrier Aggregation with 5 Component Carriers
T‑Mobile achieves record 5G Uplink speed with 5G NR Dual Connectivity
Huawei, Qualcomm, Samsung, and Ericsson Leading Patent Race in $15 Billion 5G Licensing Market
Ericsson announces capability for 5G Advanced location based services in Q1-2026
Highlights of Ericsson’s Mobility Report – November 2025
Ericsson integrates Agentic AI into its NetCloud platform for self healing and autonomous 5G private networks
