FCC plans 6G spectrum auctions before IMT 2030 frequencies have been identified and without a 6G frequency arrangement standard

Executive Summary:

Federal Communications Commission (FCC) Chairman Brendan Carr said Wednesday that a series of planned wireless-spectrum auctions could generate more than $100 billion in proceeds over the next several years.  In July, the FCC voted to conduct a 2027 auction offering 160 MHz of mid-band spectrum in the Upper C-Band.  Mr. Carr said the agency is also preparing three additional auctions to follow the C-Band sale. The FCC has formally notified stakeholders of those plans and is targeting completion of all four auctions by the end of 2028.

Carr said demand for licensed spectrum extends beyond the three national mobile network operators—AT&T, Verizon, and T-Mobile, which collectively provide nearly all U.S. mobile service. He said the planned auctions could broaden participation in the wireless ecosystem and promote additional competition.  “We’re seeing real response ​in the market in ways that we didn’t see just a couple of years ago,” Carr said.

–>We sincerely doubt that after Dish Wireless’ spectaular 5G O-RAN failure!

The FCC’s auction strategy follows several high-profile spectrum transactions. In May, the agency approved EchoStar’s $40 billion sale of wireless-spectrum assets to SpaceX, AT&T, and Verizon. It also approved Verizon’s 2024 $1 billion transaction to acquire selected spectrum assets from U.S. Cellular.

SpaceX acquired spectrum to support Starlink’s direct-to-device service and other satellite-enabled connectivity offerings. Last week, the FCC said it was advancing efforts to make additional spectrum available for space-based broadband services. The agency had previously approved SpaceX’s plan to deploy thousands of additional satellites intended to support next-generation mobile services and broadband speeds of up to 1 Gbit/s.

The Trump administration said Friday that it is laying the policy and regulatory groundwork for multiple 6G-focused spectrum auctions in 2028. The FCC is accelerating its spectrum-auction agenda in response to rapid growth in wireless demand and emerging applications, including artificial intelligence workloads, autonomous vehicles, connected sensors, and other advanced connectivity use cases.

“It’s more ​competition. It drives prices down for consumers. It raises money for the ​Treasury,” Carr ⁠said, noting that a prior auction raised funds to replace Chinese gear in U.S. networks that raised national security concerns.

The planned auctions come as mobile-data demand continues to increase, albeit at a slower rate. CTIA, the wireless-industry association, reported that U.S. consumers used 159.3 trillion MB of mobile data in 2025, a 20% increase from 2024.  CTIA also said AI-related traffic is growing at roughly three times the rate of conventional wireless traffic and could account for nearly one-third of all broadband traffic by 2034.

Nokia’s 6G Spectrum Vision:

From a regulatory perspective, IMT’s identification of new spectrum in the extended mid-band will be vital for the industry, as it can enable global or regional harmonization, provide regulatory certainty for technological investments in the ecosystem, and create economies of scale for faster development and broader adoption.

Nokia on 6G Drivers:

Analysis- What the FCC is Totally Missing:

What the FCC’s 6G auction planning appears to overlook is the international spectrum-harmonization timeline.  The FCC’s accelerated 6G auction agenda risks getting ahead of the global standards and spectrum-harmonization process. Frequency bands for 6G, formally addressed by the ITU as IMT-R-2030, will not be identified at the international level until the ITU-R World Radiocommunication Conference 2027 (WRC-27) which will take place in Shanghai, China, from October 18 to November 12, 2027.

Starting at ITU-R WP 5D meeting #57, which takes place in Jan/Feb 2028, 5D will start to develop IMT-2030 frequency arrangements for the  designated bands identified at WRC-27. Those frequenc arrangements are vitally important because they establish the technical basis for globally or regionally harmonized use of spectrum, including band plans and deployment approaches that support equipment interoperability, scale, and international roaming.  They should greatly simplify IMT 2030 roaming between carriers that use the same 6G frequency bands.

As a result, auctioning spectrum specifically characterized as “6G” before that process is complete could force U.S. policy ahead of the eventual international framework. The FCC can prepare candidate bands, study sharing and coexistence conditions, and develop auction authority and rules in advance, but it cannot yet know which bands will ultimately receive broad international support for IMT-2030.

Importance of a 6G Frequency Arrangements Standard:

The explicit purpose of this ITU-R recommendation (international standard)  is to guide administrations in selecting transmitting and receiving frequency arrangements for terrestrial IMT and to promote efficient spectrum use.

Core functions of IMT frequency arrangements:

Function What the frequency arrangement specifies Why it matters
Duplexing mode Whether spectrum is used as paired FDD spectrum, unpaired TDD spectrum, or—where appropriate—either option Determines the basic UL/DL operating model and whether two separated blocks or one unpaired block are required
Uplink/downlink allocation The specific frequency ranges assigned to mobile-station transmission and base-station transmission in FDD arrangements Establishes the direction of transmissions and supports compatible handset, base-station, and filter designs
Duplex direction Conventional FDD—UE uplink in the lower band and base-station downlink in the upper band—or reverse duplex where coexistence requires it Affects uplink link budget, interference exposure, adjacent-band compatibility, and device design
Duplex separation The fixed frequency offset between corresponding uplink and downlink channels in an FDD plan Enables paired-channel operation and drives duplexer/filter feasibility and ecosystem compatibility
Centre gap The guard separation between the upper edge of the lower FDD block and lower edge of the upper FDD block Helps define the paired-band geometry and affects duplexer bandwidth and isolation performance
Band segmentation The usable sub-bands, block boundaries, and alternative arrangements within the larger IMT-identified allocation Allows a regulator to choose an arrangement fitting regional allocations, incumbent services, and licensing structure
Channel or carrier placement The relationship between particular uplink and downlink carrier positions, including consistent pairing rules Permits terminals and networks to map channels predictably across deployments
Bandwidth scalability Practical bandwidth options and the extent to which a national administration may implement a full arrangement or only a contiguous portion of it Supports staged awards and deployments without breaking the fundamental pairing relationships
TDD operating framework The unpaired frequency range available for TDD deployment, while leaving the specific DL/UL time split to the air-interface and operator configuration Supports asymmetric traffic and wide-channel operation, particularly in mid-band and mmWave spectrum
Coexistence support Arrangement choices designed to mitigate interference with adjacent services or neighboring IMT systems Makes the band plan usable in real national and cross-border spectrum environments
Harmonization reference A common international menu of arrangements rather than one mandatory global plan Supports economies of scale, multivendor equipment availability, international roaming, and cross-border coordination

In summary, IMT frequency arrangements do considerably more than identify whether a band uses FDD or TDD. They define the uplink and downlink frequency blocks, duplex direction, duplex separation, center gap, carrier-pairing relationships, usable sub-band segmentation, and alternative arrangements needed to translate an IMT identification into an interoperable and commercially deployable spectrum plan.

A Very Bad Omen from ITU-R WP5D – 2.5 Year Gap Between IMT 2020 Co-Recommendations (M.2150-0 and M.1036-7):

One should not equate a WRC-27 IMT frequency identification with a completed 6G spectrum standard. The IMT-2020 precedent is highly instructive: after WRC-19 identified new 5G bands, ITU-R WP 5D did not complete the corresponding ITU-R M.1036 frequency arrangements recommendation on the same timetable as the IMT-2020 RIT/SRIT specification. despite being a co-requisite standard. The ITU-R M.2150-0 radio-interface recommendation (IMT 2020 RIT/SRIT) was approved and published on February 1, 2021, while the internationally agreed IMT 2020 frequency band (M.1036-7) recommendation wasn’t approved until December 13, 2023.  [All versions of ITU-R M.1036 (IMT 2020 frequency arrangements) are available for download here.]

For over two and 10 months, 5G (IMT-2020 RIT/SRIT ) was actively deployed globally without a finalized, globally harmonized UN/ITU-R standard for its internationally specified frequency arrangements (specifically the mmWave bands designated by WRC-19). During that interval, the formal 5G radio-interface standard existed, but the internationally agreed spectrum-arrangements framework remained incomplete for the major new WRC-19 IMT 2020 frequency bands.  The nearly three-year disconnect demonstrated that an IMT RIT/SRIT recommendation, by itself, was not a complete international 5G standardization outcome. Without settled arrangements for duplexing, spectrum segmentation, uplink/downlink operation, carrier placement, and coexistence, WRC-identified spectrum does not automatically become harmonized, deployable IMT spectrum.  Let’s hope WP 5D learned a lesson from that fiasco.

–>Since ITU-R Recommendations are technically voluntary and non-binding, national regulators and regional bodies simply bypassed the ITU-R impasse to build their 5G networks.   Nonetheless, several 5G deployment problems resulted:

1.  Stagnation and “Death” of 5G mmWave
The biggest casualty of the M.1036 delay was the deployment of millimeter-wave (mmWave) 5G (such as the 24.25–27.5 GHz and 37–43.5 GHz bands). 
    • The Problem: WRC-19 had identified these bands for 5G, but because a geopolitical impasse (largely driven by the Russian Federation over existing satellite/military service protections) blocked the M.1036 frequency arrangement consensus, there was no UN-sanctioned blueprint for channel channeling plans or guard bands.
    • The Impact: Outside of early adopters like the US (via the FCC), most global operators refused to touch mmWave spectrum. Capital expenditure shifted entirely to mid-band (C-band) frequencies, stalling the rollout of ultra-low latency, high-capacity 5G applications for years. 

2. Fragmentation and Bypassing the ITU via 3GPP
Because wireless network operators could not wait for the ITU to resolve its internal political deadlock, they relied on alternative specifications.
    • The Problem: The industry treated 3GPP Release 16 and subsequent spectrum definitions as the de facto authority.
    • The Impact: Rather than utilizing an internationally validated ITU framework, global network vendors and national regulators (like the FCC in the US or CEPT in Europe) executed their own domestic spectrum rules. This reduced the ITU’s role during that window from an active coordinator to a passive archivist validating rules after the networks were already built.

3. Delays in Global Hardware Economies of Scale
The primary purpose of M.1036 is to build a unified global market so device manufacturers can put the same antennas into phones worldwide, dropping production costs.
    • The Problem: Without a finalized international agreement on exact band arrangements, device OEMs (Original Equipment Manufacturers) faced uncertainty about which exact block matrices and duplexing directions would become standard globally.
    • The Impact: Early 5G smartphones required highly fragmented, region-specific RF front-end architectures. This slowed down the decline of 5G handset prices, particularly for phones capable of international roaming on high-frequency bands.

4. Severe Cross-Border Coordination Friction
  • The Problem: In regions like Europe, Africa, and parts of Asia, countries sit in close geographical proximity. Without M.1036 specifying standard guard bands and TDD (Time Division Duplexing) synchronization models for the newly opened frequencies, there was no international baseline for interference mitigation.
  • The Impact: Neighboring nations had to negotiate messy, bilateral spectrum-sharing agreements to prevent base stations in one country from bleeding over and blinding mobile networks or satellite receivers in another

Bottom Line:

For IMT-2030, the same distinction will be decisive. WRC-27 may identify additional bands for IMT-2030, but identification alone will not produce globally usable 6G spectrum. WP 5D must subsequently develop agreed IMT-2030 frequency arrangements for those bands. Until that work is concluded, it is premature to presume that frequencies auctioned in 2028 will have internationally harmonized 6G band plans or broad device-ecosystem support. Make no mistake that detailed IMT 2030 (6G) frequency arrangements will be needed for interoperable 6G deployment.

The ITU-R IMT 2030 Frequency Arrangements recommendation, expected to be approved in late 2030 or early 2031, will provide the internationally recognized implementation alternatives for arranging that spectrum. A national regulator may still adopt a different domestic plan, but departure from that standard will likely reduce device scale, raise RF complexity, and weaken prospects for roaming and cross-border 6G compatibility.

An FCC auction in 2028 may be feasible as a U.S. domestic spectrum-policy action, but calling it a “6G auction” would be technically way premature until the relevant IMT-2030 frequency arrangements have been completed by WP 5D  and internationally supported.

………………………………………………………………………………………………………………………………………………………………………………………………..

References:

https://www.reuters.com/business/media-telecom/us-official-says-upcoming-spectrum-auctions-could-generate-more-than-100-billion-2026-09-17/

https://www.itu.int/rec/R-REC-M.1036/en

https://www.nokia.com/6g/spectrum-for-6G-explained/

GSMA Vision 2040 study identifies spectrum needs during the peak 6G era of 2035–2040

ITU-R M.[IMT-2030.EVAL] & ITU-R M.[IMT-2030.SUBMISSION] reports: Evaluation & Submission Guidelines for 6G RIT/SRITs (6G)

Roles of 3GPP and ITU-R WP 5D in the IMT 2030/6G standards process

ITU-R M.2150: Detailed specifications of the radio interfaces of IMT-2020

IMT 2020.SPECS approved by ITU-R but may not meet 5G performance requirements; no 5G frequencies (revision of M.1036); 5G non-radio aspects not included

Only domestic network equipment may be used for 5G in Russia; Revision of ITU-R M.1036 urgently needed

Do ITU Radio Regulations Matter? China allocates 6 GHz spectrum for 5G and 6G services prior to WRC 23; CTIA objects!

Huawei’s Ascend Silicon Roadmap Extends From AI Chips to Cluster-Scale Infrastructure

Note:  Perplexity.ai was use to research this article.

……………………………………………………………………………………..

Executive Summary:

Huawei is accelerating development of an indigenous AI-computing platform intended to reduce China’s reliance on U.S.-supplied accelerators, networking technologies, and associated software ecosystems. The effort aligns with the company’s stated strategy of becoming a full-stack supplier of AI-era computing and connectivity infrastructure—rather than primarily a developer of frontier foundation models.

Huawei said it plans to introduce two new Ascend AI accelerators in 2027: the Ascend 960DT in the first quarter and the Ascend 960PR in the third quarter. The company also disclosed a longer product roadmap extending through 2029 and said it has shipped more than 1,000 AI-computing systems to over 370 customers.

A necessary distinction is that Huawei designs its Ascend processors through HiSilicon, its semiconductor-design subsidiary, but does not own or operate leading-edge chip fabrication facilities. Production therefore depends on external foundry, memory, advanced-packaging, and equipment supply chains—areas that remain materially constrained by U.S. export controls and by the capabilities of China’s domestic semiconductor ecosystem.  For example, Huawei can’t use TSMC to make its silicon so must rely on Chinese chipmakers.

As per an earlier IEEE Techblog post this week, Guo Ping, chairman of Huawei’s supervisory board,  told new employees that Huawei’s objective in computing and communications is to “become Nvidia.” In technical terms, Huawei is not attempting merely to replicate a GPU. It is seeking to assemble an alternative AI-infrastructure stack: Ascend accelerators; Kunpeng general-purpose processors; servers and rack-scale systems; high-speed interconnect; storage; cloud services; and the CANN software environment that must serve, in China, some of the platform role CUDA performs globally for Nvidia.

Visitors tour Huawei’s Ascend AI exhibition booth during a conference last year. Andy Wong/AP

From accelerator to system architecture:

Huawei’s principal response to a per-device performance gap is system-scale integration. Rather than relying exclusively on the performance of a single accelerator generation, the company is assembling larger clusters of Ascend devices and attempting to improve the efficiency of communication, memory access, workload partitioning, and collective operations across those systems.

This is strategically relevant because training and serving large AI models are increasingly limited not only by floating-point throughput, but also by memory capacity and bandwidth, interconnect latency, bisection bandwidth, power delivery, cooling, and software efficiency. A large accelerator cluster can only approach the behavior of a unified computing resource if its networking and systems software prevent communication overhead from overwhelming the benefits of adding more processors.

Huawei has promoted UnifiedBus as an architectural approach for improving communication among processors, memory, and devices across server and rack boundaries. The underlying objective is familiar to data-center architects: reduce data-movement overhead and make a distributed accelerator cluster behave more like a coherent, programmable system. Whether Huawei can deliver this at scale will depend on achievable latency, bandwidth, congestion control, fault tolerance, topology-aware scheduling, and software maturity—not simply on the number of accelerators installed.

That systems emphasis is consistent with Huawei’s established position in telecom infrastructure. Unlike a pure-play AI-chip supplier, Huawei can combine AI computing with data-center networking, optical transport, IP infrastructure, cloud platforms, mobile networks, and edge-computing systems. The potential differentiator is therefore architectural integration, particularly for AI-RAN, autonomous network operations, edge inference, digital twins, and operator cloud deployments where AI workloads must interact with network telemetry and real-time control functions.

Interconnect and optics matter:

The company is also pursuing near-packaged optics, an approach intended to shorten electrical paths between switching or compute silicon and optical transceivers. In principle, placing optical components close to high-speed silicon can reduce electrical-channel loss and potentially improve energy efficiency as link speeds and port densities rise.

Huawei’s terminology should not be conflated automatically with co-packaged optics. Co-packaged optics generally integrates optical engines and switch or compute ASICs in a common package or closely coupled assembly; near-packaged approaches may retain more physical separation while still reducing copper reach. The key engineering question is not the label, but the extent to which the architecture can deliver lower power per transmitted bit, higher density, manufacturability, serviceability, and operational reliability at scale.

For large AI clusters, the interconnect fabric is increasingly a first-order design constraint. As clusters expand, network performance determines how effectively distributed training workloads can scale. The challenge is particularly acute where a vendor seeks to compensate for lower accelerator performance by deploying more devices: more devices can increase aggregate compute capacity, but they also increase synchronization traffic, power consumption, cabling complexity, failure exposure, and the burden on cluster-management software.

Constraints remain substantial:

Huawei’s roadmap should be assessed as a bid to establish a credible domestic alternative AI platform, not as evidence that it has achieved parity with Nvidia’s highest-end systems. Its Ascend roadmap faces several interdependent constraints:

  • Advanced fabrication remains dependent on external foundry capacity and equipment availability, even though HiSilicon can design sophisticated processors.

  • High-bandwidth memory availability, yield, packaging capability, and supply-chain scale can materially affect system output and performance.

  • Cluster-level competitiveness depends on interconnect bandwidth, latency, memory architecture, power efficiency, and the ability to operate reliably at very large scale.

  • CANN must attract developers, framework integrations, tools, libraries, and application vendors in an ecosystem where CUDA remains deeply embedded.

  • Customer adoption will depend on total cost of ownership, application portability, model performance, support quality, and availability of hardware at predictable volumes.

The strategic rationale is nonetheless clear. Export controls have increased the value of a domestically supplied AI-computing stack, even if individual components or systems lag the frontier in some metrics. As Huawei rotating chairman Eric Xu put it, the company’s concern is not only whether it can access the most advanced technology, but whether it can avoid strategic dependence on supply decisions made elsewhere.

For telecom operators, the central issue is whether Huawei can turn this silicon-and-systems program into deployable AI infrastructure for network operations. A credible offering would require more than Ascend processors. It would require validated reference architectures for AI-RAN and telco cloud, high-performance east-west networking, operational automation, observability, model lifecycle management, security controls, and a software ecosystem that can support carrier-grade availability.

Huawei’s advantage is that it already participates across many of those domains. Its challenge is proving that the integrated stack can deliver competitive performance, efficiency, ecosystem breadth, and supply assurance under sustained technology restrictions. That is the practical test of its ambition to become China’s Nvidia-equivalent in AI infrastructure.

……………………………………………………………………………………………………………………………………………………………..

References:

https://www.wsj.com/tech/ai/huaweis-plan-to-become-chinas-nvidia-8af8d8a1 [paywall]

https://www.wsj.com/tech/chinas-huawei-develops-new-ai-chip-seeking-to-match-nvidia-8166f606 [paywall]

Huawei Chairman: Strategic Focus on AI Computing & Connectivity Infrastructure

Huawei unveils AI Centric Network roadmap, U6 GHz products, 5G Advanced strategy and SuperPoD cluster computing platforms

Key take-aways: “16th Smart City and Intelligent Economy Expo” for Huawei, Alibaba & China’s three state backed carriers

Omdia’s 2025 Mobile Core Network Leaders: Huawei #1 in market share; Nokia #1 for portfolio competitiveness

Analysis: Huawei”s upgraded Xinghe Intelligent Network Solution for South Africa

Huawei’s AI-Centric Network Vision: Six Imperatives for the Next Decade; Critical Questions for IEEE Techblog Community

Huawei FY2025: 2.2% YoY revenue increase; strategic pivot to AI and intelligent automotive solutions

Huawei, Qualcomm, Samsung, and Ericsson Leading Patent Race in $15 Billion 5G Licensing Market

Huawei Cloud Review and Global Sales Partner Policies for 2026

Huawei’s Electric Vehicle Charging Technology & Top 10 Charging Trends

Huawei to Double Output of Ascend AI chips in 2026; OpenAI orders HBM chips from SK Hynix & Samsung for Stargate UAE project

 

4.8 GHz to 4.9 GHz frequency band uses & Verizon Wireless experimental license for testing ISAC with Ericsson

Introduction:

The 4.8 GHz to 4.9 GHz frequency band is a critical slice of mid-band spectrum located within the larger IEEE C-band (4.0 to 8.0 GHz) and the ITU Ultra High Frequency (UHF)/Super High Frequency (SHF) boundary.  It is currently used primarily for public safety operations (especially the 4.9 GHz portion allocated to FirstNet for drone control and crisis response) and select 5G deployments (part of 3GPP 5G NR Sub-6 GHz Band n79 in parts of Asia).  It is extensively deployed for commercial 5G networks across Asia (such as China and Japan) and Europe, while in the United States, it is being studied for repurposing and testing advanced 5G/6G integrated sensing and communication (ISAC).  Finally,  it is actively being studied for future 5G-Advanced and 6G mobile broadband and sensing use cases.

Executive Summary:

1. Global 5G Spectrum Allocation Context:
The chart below shows how the 4.8–4.9 GHz range fits into international wireless allocations:

Region / Country 5G Spectrum Range Mid-Band Status
China 4.8 GHz – 4.9 GHz (and 3.3–3.6 GHz) Actively deployed for 5G NR Band n79
Japan 3.6–4.1 GHz & 4.5–4.9 GHz Allocated to major carriers (e.g., NTT Docomo)
United States 4.4 GHz – 4.94 GHz Historically Federal/Military; currently under review for commercial 5G/6G

2. RF Characteristics of 4.8–4.9 GHz:
  • Favorable Propagation: It provides a strong balance between coverage area and high data throughput (bandwidth). [1]
  • Capacity: This frequency handles heavy data loads—like high-definition real-time video streaming—much better than lower bands (like 700 MHz). [1]
  • Indoor Penetration: It experiences higher atmospheric and structural attenuation compared to 2.4 GHz or 3.5 GHz, meaning it requires denser small-cell deployments for deep indoor coverage

Verizon’s Experimental License for the 4.8GHz-4.9GHz Band:

Verizon Wireless (under the company name Cellco Partnership) has applied to the FCC for an experimental license to use the 4.8GHz-4.9GHz band in and around its lab in Los Angeles, California. Verizon Wireless’s request is temporary, expiring one year from grant.  The application is pending approval, but the license is expected to be granted soon, according to Steve Crowley.

Verizon said, “this experimental authorization is necessary to test Ericsson software and hardware in a noncommercial environment. Verizon Wireless requests this authorization to demonstrate 5G advanced technology use cases such as object sensing as a precursor to 6G Integrated Sensing and Communication (ISAC). This work will facilitate the adoption of use cases across commercial, public safety, and defense sectors.”

Grant of the experimental authorization will allow Verizon Wireless to set up hardware radiating in 4.8-4.9 GHz as part of a private network not connected to commercial operators. The radios are certified for use under 47 C.F.R. Part 27 and Part 96. The testing will be conducted using varying bandwidths to allow for range, resolution, and confidence assessments. Testing will occur outdoors and all transmissions will be controlled at the locations provided in the application. The fixed base stations will employ directional antennas and will have a maximum antenna elevation as described in the application. The base station antenna has a half-power beam width of approximately 24º vertically and 65º horizontally.

Verizon Wireless is unable to determine the incumbent users in the 4.8-4.9 GHz band. Verizon Wireless will coordinate with the incumbents once the FCC and NTIA have provided the agencies impacted, and prior to commencing testing, to avoid any potential disruptions to their operations. If incumbent licensees experience interference, Verizon Wireless will cease interfering operations.

Verizon is one of the first companies to apply for a test license in 4.4GHz specifically for 6G-related trials, according to Crowley. He noted it is also first time the carrier has requested a test license in any of the frequencies under study, which include 1.6GHz, 2.7GHz, 4.4GHz and 7GHz.

Other Entities Pursuing the 4GHz Band:

Samsung Research America requested an experimental license this month to use 4720MHz-4820MHz in Plano, Texas, but its application did not specify use cases to be tested.

The 4.8GHz-4.9GHz spectrum falls within the so-called 4.4GHz band (4400MHz-4940MHz) that the NTIA has identified for study to repurpose for commercial licensed use as part of the U.S. plan to make more frequencies available for 5G and 6G.

The CTIA refers to the band as 4GHz, rather than 4.4GHz. The group published a report this week to make an economic and technical case for the “prime midband spectrum” and urged policymakers to move quickly on studies this fall. They described 4GHz as the “a cornerstone” of the US spectrum pipeline, along with upper C-band, 2.7GHz and 7GHz, because it offers potential 400MHz of bandwidth and already has some equipment and device support. Further, it is a “compelling alternative” to China’s push for licensed use of the 6GHz band in spectrum negotiations at ITU World Radiocommunication Conference 2027 (ITU WRC-27).

Will It Be Used for 6G?

  • International Consideration: The ITU-R  and organizations like Nokia are studying the 4.4–4.8 GHz and neighboring ranges as potential mid-band spectrum for future IMT-2030 (6G) services. There are several ongoing compatibility studies in ITU-R WP 5d between terrestrial, maritime and satellite use of this band.  They are in preparation for ITU WRC-27 which will specify IMT 2030 (6G) frequencies.
  • Industry Support: Industry groups like the CTIA advocate for the broader 4 GHz mid-band range as a prime, lower-risk spectrum for wide-area mobile capacity to bridge 5G-Advanced and early 6G rollouts.
  • Global Hurdles: Formal decisions regarding international mobile allocation for this band will be evaluated globally at the ITU-R World Radiocommunication Conference (WRC-27), though support varies by region due to existing incumbent and defense users.
  • Verizon added nine new members to its 6G Innovation Forum this week and signaled intent to test more use cases in the areas of ISAC, digital twins, robotics, AI and wearables.  The new members are Amazon Web Services (AWS), Cisco, Intel, Keysight Technologies, MediaTek, Nvidia, Palo Alto Networks, Rohde & Schwarz and Viavi Solutions. They join founding members Ericsson, Samsung, Nokia, Meta and Qualcomm Technologies.

References:

https://x.com/StevenJCrowley/status/2097450891444142433

https://apps.fcc.gov/els/GetAtt.html?id=412698&x=

https://apps.fcc.gov/oetcf/els/reports/442_Print.cfm?mode=current&application_seq=154088&license_seq=156054

https://www.lightreading.com/6g/verizon-seeks-fcc-approval-to-test-6g-use-cases-in-4ghz

https://urgentcomm.com/network-tech/an-introduction-to-4-9-ghz

ETSI Integrated Sensing and Communications ISG targets 6G

Key take-aways: “16th Smart City and Intelligent Economy Expo” for Huawei, Alibaba & China’s three state backed carriers

Analysis: Cohere’s $28M U.S. DoD FutureG ISAC contract; OTFS vs OFDM; 6G-NR/IMT 2030 RIT standards outlook

3GPP approves timelines for Release 21 which will specify 6G RAN, Core and 5G Advanced

 

 

Dell’Oro: Data Center capex grew 92% in 2Q-2026 (caveats galore)

According to a new report by Dell’Oro Group, the global data center capital expenditures strongly accelerated in 2Q 2026. Continued AI infrastructure investment supported growth across compute, storage, networking, and physical infrastructure, while rising memory and storage prices significantly increased server average selling prices.

“Data center capex growth broadened in the second quarter as investment accelerated across both established Cloud Service Providers and emerging AI infrastructure customers,” said Baron Fung, Vice President of Research at Dell’Oro Group.

“Spending remained concentrated in NVIDIA Blackwell Ultra and hyperscaler custom accelerators, while agentic AI created incremental demand for general-purpose compute, storage, and complementary networking. Neocloud providers and AI model builders are also becoming increasingly important contributors to infrastructure investment. These companies are rapidly expanding their own capacity while deepening partnerships with cloud service providers.”

“Looking ahead, ongoing accelerator deployments and emerging agentic AI and AI-related storage workloads should sustain strong capex growth through the remainder of 2026 and beyond, although supply constraints could limit the pace at which planned infrastructure is deployed,” explained Fung.

Additional highlights from the 2Q 2026 Data Center IT Capex Quarterly Report:

  • Neocloud and AI Model Builder capex grew the fastest among the customer segments, reflecting the early stages of their infrastructure buildouts.
  • Higher memory and storage prices provided an additional lift to capex by driving server average selling prices higher.
  • Dell led server OEM revenue, followed by SuperMicro and Lenovo, while white-box server revenue reached a record high.

………………………………………………………………………………………………………………………………………………………….

On August 18th Dell’Oro Group forecasted that worldwide data center capital capex is to maintain growth momentum and surpass $3 trillion by 2030. High-end accelerators powering accelerated servers optimized for AI are expected to represent the largest share of data center capex and remain the primary driver of capex growth over the forecast period.

“Our 2030 data center capex outlook has nearly doubled since the January 2026 forecast, reflecting higher hyperscale capex guidance, increased projections for global data center power capacity, and higher commodity costs,” said Baron Fung, Vice President of Research at Dell’Oro Group. “High-end accelerators powering AI-optimized servers are expected to account for the largest share of data center capex and remain the primary driver of growth over the forecast period.

“However, the pace of growth will depend on the sustainability of investment, power availability, and supply chain conditions. Accelerated and heterogeneous computing, along with innovations in server efficiency, could help mitigate the rising cost and infrastructure demands of AI. The Top 4 US hyperscalers alone could represent about half of global capex, while enterprise investment remains constrained by uncertain AI returns,” according to Fung.

Additional highlights from the Data Center IT Capex 5-Year July 2026 Forecast Report:

  • High-end accelerators are expected to account for the largest share of data center capex and remain the primary driver of spending growth through 2030.
  • General-purpose server demand is expected to benefit from growing inference, agentic AI, and storage workloads, broadening infrastructure growth beyond accelerated computing.
  • The newly added AI-specialized cloud segment, comprising AI model builders and neocloud service providers, is projected to grow at nearly a 60 percent CAGR, outpacing the growth of other customer segments.

Additional highlights from the Data Center IT Capex 5-Year July 2026 Forecast Report:

  • High-end accelerators are expected to account for the largest share of data center capex and remain the primary driver of spending growth through 2030.
  • General-purpose server demand is expected to benefit from growing inference, agentic AI, and storage workloads, broadening infrastructure growth beyond accelerated computing.
  • The newly added AI-specialized cloud segment, comprising AI model builders and neocloud service providers, is projected to grow at nearly a 60 percent CAGR, outpacing the growth of other customer segments.

IEEE Techblog Analysis:

  • Dell’Oro had already raised its 2026 global data-center capex outlook to more than $1 trillion in its June 2026 report. It also said 2H26 growth was expected to accelerate because of NVIDIA Rubin deployments and hyperscaler custom-accelerator refreshes. That makes the new 92% 2Q figure much more significant: this isn’t simply a strong quarter; it is occurring within a $1-trillion-plus annual investment cycle.
  • Epoch AI’s tracking shows combined hyperscaler quarterly capex has been increasing at an average 72% annual rate since 2Q23 and projects approximately $770 billion for 2026 if the trend continues.
  • Another estimate from Moody’s put 2026 hyperscaler capex at $785 billion, including Microsoft, Amazon, Meta, Alphabet, Oracle and CoreWeave.
  • Dell’Oro’s separate 2Q semiconductor/component report says data-center component revenue increased 182% YoY, while DRAM and storage-drive average selling prices per bit more than doubled.  Therefore, some of the 92% increase in data-center capex is clearly inflation in the cost of the equipment, not necessarily an equivalent increase in physical infrastructure.
  • Therefore, the 92% increase in capex should not be interpreted as a 92% increase in deployed computing capacity, because sharply higher DRAM, NAND/storage and other component prices are inflating server system costs.

What’s Missing from this Report:

The press release for this report notes that worldwide data center capital expenditures grew 92% in 2Q02026, driven by surging AI demand and memory costs. However, it omits precise spending figures for individual hyperscalers (Alphabet, Amazon, Meta, Microsoft and Oracle) as well as OEM market share details.  Moreover, YoY growth can conceal the current trajectory. Sequential growth would show whether the AI infrastructure spending acceleration actually intensified during 2Q of 2026.

Conclusions:

The 92% year-over-year increase in 2Q26 data-center capex needs to be viewed against a much larger AI infrastructure investment cycle. Dell’Oro had already raised its 2026 global data-center capex forecast to more than $1 trillion, with 2H26 spending expected to accelerate further as NVIDIA’s Rubin systems and hyperscaler custom accelerators ramp. At the same time, Dell’Oro reported that data-center semiconductor and component revenue surged 182% in 2Q26, with DRAM and storage-drive prices per bit more than doubling year over year. Consequently, a significant portion of the reported capex growth reflects higher equipment prices rather than a comparable increase in physical computing capacity. Meanwhile, hyperscaler capex is increasingly being supplemented by debt-financed Neocloud and AI-model-builder infrastructure, broadening the investment cycle beyond the traditional cloud giants.

About the Report:

Dell’Oro Group’s Data Center IT Capex Quarterly Report details the data center infrastructure capital expenditures of the largest hyperscale cloud service providers, AI Model Builders, Neocloud, Rest of Cloud, Telco, and Enterprise customer segments. It provides the allocation of data center infrastructure capex for general-purpose and accelerated servers, storage systems, and other auxiliary data center equipment. The report also discusses market trends, drivers of the leading cloud service providers’ capex growth during the quarter, and the outlook for the next year. To purchase this report, please contact us at [email protected].

…………………………………………………………………………………………………………………………………………………………………………………..

References:

Data Center Capex Grew 92 Percent in 2Q 2026, Driven by Surging AI Demand and Memory Costs, According to Dell’Oro Group

AI Buildout Maintains Momentum as Data Center Capex Surpasses $3 Trillion by 2030, According to Dell’Oro Group

https://www.gate.com/news/detail/data-center-capital-expenditure-to-exceed-1-trillion-in-2026-social-graph-23974521

https://www.gate.com/news/detail/global-ai-data-center-spending-to-hit-316-trillion-by-2050-pwc-projects-23945476

Dell’Oro: 2H2026 Data Center Capex to Accelerate due to massive AI Deployments

PwC: Global AI data center spending to hit $31.6tn by 2050; Role of full stack orchestration layer explained

Nvidia CEO Huang: AI is the largest infrastructure buildout in human history; AI Data Center CAPEX will generate new revenue streams for operators

China vs U.S.: Race to Generate Power for AI Data Centers as Electricity Demand Soars

AI risks and backlash increase; Recap of the circular loop of fake AI profits and hyperscaler markups of private AI companies

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

Expose: AI is more than a bubble; it’s a data center debt bomb

Will billions of dollars big tech is spending on Gen AI data centers produce a decent ROI?

Huge Risks for the proposed $500B AI Investments from Giant Wall Street firms

 

 

 

 

 

Huawei Chairman: Strategic Focus on AI Computing & Connectivity Infrastructure

Note:  Perplexity.ai was use to research this article.

According to Chairman Guo Ping, Huawei aims to position itself as a full-stack provider of AI-era computing and connectivity infrastructure—effectively a Chinese counterpart to Nvidia in the infrastructure layer rather than in foundation-model development. In remarks posted to Huawei’s internal Xinsheng bulletin board, Guo told a group of new employees that the company’s strategic priority remains the advancement of its core connectivity and computing capabilities, rather than diversification into unrelated businesses. “There are no plans for business expansion,” he said, framing artificial intelligence (AI) as Huawei’s most significant strategic opportunity.  Via Google Translate:

“Huawei’s core strategy is “focus”—specifically, deepening our expertise in the fields of connectivity and computing. We have no plans for business expansion. We aim to achieve even better results by providing digital infrastructure solutions to global customers, offering diverse solutions for the development of artificial intelligence, and contributing to the rise of China’s electronics industry. This is our vision for the future: to deepen our focus in our core areas and create greater value.”

Guo characterized AI as potentially humanity’s “last technology revolution,” arguing that Huawei must establish a durable competitive position in the sector. His comments suggest that Huawei sees its principal AI opportunity not in competing directly with hyperscalers and model developers on proprietary large language models (LLMs), but in supplying the underlying platforms on which a broad range of models can be trained, inferred, and deployed.

That strategy centers on Huawei’s Ascend AI processors and associated computing systems, together with Kunpeng general-purpose server platforms. The company’s objective is to provide an alternative AI infrastructure stack spanning processors, servers, networking, cloud platforms, and AI software tooling—capable, in Huawei’s view, of supporting diverse foundation models and enterprise AI workloads.

Image from Huawei

However, Guo indicated that Huawei does not regard AI model development as entirely optional. Its device and intelligent-automotive businesses will require domain-specific large models because open-source alternatives remain insufficient for certain proprietary, embedded, and safety-sensitive use cases. In addition, the increasing convergence of cloud infrastructure and AI services means Huawei Cloud must continue developing its own LLM capabilities to preserve platform competitiveness.

Strategic Implications:

The distinction is important. Huawei appears to be pursuing a layered AI strategy:

  • Infrastructure first: Ascend AI accelerators, Kunpeng processors, AI servers, high-performance interconnects, storage, and data-center infrastructure form the company’s primary competitive focus.

  • Models where strategically necessary: Huawei’s AI developed models (such as its enterprise Pangu LLM suite) are intended to support cloud services and vertical applications where open models do not meet performance, integration, security, or product-control requirements.

  • Connectivity as a differentiator: Unlike Nvidia, Huawei can combine AI computing platforms with extensive capabilities in wireless RAN, core networks, IP transport, optical networking, enterprise networking, and cloud infrastructure.

  • Domestic ecosystem resilience: The strategy also advances China’s effort to reduce dependency on U.S.-origin AI processors, software ecosystems, and advanced semiconductor supply chains.

For telecom operators, the relevant issue is whether Huawei can translate this integrated portfolio into credible AI-RAN, autonomous-network, edge-AI, and cloud-network offerings. The company’s position in radio access, transport, optical, and data-center infrastructure could give it an architectural advantage in deployments where AI workloads are tightly coupled with network telemetry, operational data, and real-time control loops.

Controversies, Constraints and Competition:

Guo’s remarks are more confirmatory than revelatory, but they provide an uncommon view of the company’s internal strategic framing. Huawei has substantially curtailed engagement with many foreign media and public-policy audiences, particularly as U.S. sanctions, export controls, and security restrictions have constrained its international operations.

The internal question-and-answer format also allows executives to articulate strategic priorities without directly addressing the numerous controversies facing the company, including the racketeering trial involving Huawei’s business with Iran that began in New York earlier this month. 

Guo said Huawei would continue supplying digital infrastructure and AI solutions to customers in China and international markets. He ruled out expansion into satellite or launch-vehicle manufacturing and said the company does not intend to enter automobile manufacturing—consistent with Huawei’s previous position that it will provide intelligent-vehicle technology rather than build branded cars.

Huawei’s hardware production is heavily bottlenecked by U.S. sanctions (yield issues on advanced chips, SMIC node restrictions, and high-bandwidth memory shortages).  On the prospect of additional sanctions or technology restrictions, Guo argued that China’s domestic market—home to roughly 1.4 billion people—provides Huawei with a sufficient base for long-term survival. But he also emphasized that the company’s aspirations extend beyond domestic demand: “Our concern is whether we are competitive compared to domestic and international competitors.”

AI infrastructure relies as much or even more on software than hardware. Huawei’s CANN software layer must  compete with  Nvidia’s CUDA, which makes AI developer adoption a tough uphill battle.

Analysis & Conclusions:

That will be the central test of Huawei’s AI strategy. Its ability to build a viable alternative to Nvidia-led AI infrastructure faces immediate supply-chain constraints under U.S. sanctions.  There’s also the software challenge of driving developer adoption toward Huawei’s CANN architecture over Nvidia’s CUDA. Long-term viability will depend not only on Ascend AI silicon performance, but also. but also on software maturity, developer adoption, supply-chain scale, interconnect and memory performance, power efficiency, and the breadth of its ecosystem outside China.

Guo’s comments underline Huawei’s attempt to convert its established strengths in telecommunications, enterprise infrastructure, cloud, and silicon into an integrated AI platform strategy. The approach offers the company a potentially differentiated position in AI-enabled networks and cloud infrastructure, but its success will depend on whether the company’s Ascend AI silicon and the broader Huawei software ecosystem can achieve sufficient performance, scale, developer support, and supply-chain resilience to compete beyond China.

Achieving Huawei’s AI infrastructure aspirations faces severe headwinds, including ongoing U.S. chip sanctions that limit advanced node and memory production, as well as the steep challenge of driving developer migration from Nvidia’s CUDA to Huawei’s CANN ecosystem. Furthermore, while Huawei insists models are secondary, its domain-specific Pangu LLM suite will be critical to proving its cloud and enterprise viability.

“Large corporations have both strengths and weaknesses; to adapt to the future, a company must constantly evolve and reinvent itself,” Guo said.  That will be Huawei’s formidable challenge in the age of AI.

……………………………………………………………………………………………………………………………………………

References:

https://xinsheng.huawei.com/next/detail/#/index?uuid=1316483709694681088

https://www.reuters.com/legal/litigation/huawei-heads-trial-us-over-its-business-dealings-iran-2026-09-08/

https://www.lightreading.com/ai-machine-learning/we-want-to-be-the-nvidia-of-ict-says-huawei-boss

Huawei unveils AI Centric Network roadmap, U6 GHz products, 5G Advanced strategy and SuperPoD cluster computing platforms

Key take-aways: “16th Smart City and Intelligent Economy Expo” for Huawei, Alibaba & China’s three state backed carriers

Omdia’s 2025 Mobile Core Network Leaders: Huawei #1 in market share; Nokia #1 for portfolio competitiveness

Analysis: Huawei”s upgraded Xinghe Intelligent Network Solution for South Africa

Huawei’s AI-Centric Network Vision: Six Imperatives for the Next Decade; Critical Questions for IEEE Techblog Community

Huawei FY2025: 2.2% YoY revenue increase; strategic pivot to AI and intelligent automotive solutions

Huawei, Qualcomm, Samsung, and Ericsson Leading Patent Race in $15 Billion 5G Licensing Market

Huawei Cloud Review and Global Sales Partner Policies for 2026

Huawei’s Electric Vehicle Charging Technology & Top 10 Charging Trends

Huawei to Double Output of Ascend AI chips in 2026; OpenAI orders HBM chips from SK Hynix & Samsung for Stargate UAE project

 

 

 

 

 

Bain: AI to greatly increase network operator expenses; network re-engineering needed!

Introduction by Bain:

“Over the next three to five years, the operating cost structure used by telecom operators is likely to undergo one of the most significant shifts in decades. As AI agents become embedded across customer care, network operations, software engineering, and enterprise functions, tokens will account for a growing share of operating expenditures.”

Key Points:

  • Unless telcos proactively manage their costs, scaling up AI will simply add expenses to an already-heavy legacy base.
  • An agentic operating model is emerging: a 70-to-30 ratio of legacy to AI costs, with AI automating or augmenting work across processes.
  • Cost traps lurk: cheaper AI models, bigger bills; bolting AI onto legacy processes; and demos mistaken for transformation.
  • Telco leaders can take five key actions today to avoid the traps.

Executive Summary:

Telecom operators are extending AI agents across network operations as they pursue higher levels of autonomy, but the shift could introduce a significant new operating-cost burden unless legacy processes, tooling and organizational structures are retired alongside the automation, according to Bain & Company.

Bain’s warning comes as operators accelerate plans for autonomous networks. TM Forum reported in June that 81% of 80 surveyed operators are targeting Level 4 autonomous networks or higher by 2030, and 20% expect to reach that threshold by 2027.

Under TM Forum’s Autonomous Networks framework, Level 4 moves beyond rule-based or preconfigured automation toward closed-loop, intent-driven decision-making within defined network domains. Current Level 4 work includes deployment of closed-loop operations in production networks, agent-based operating architectures, and metrics intended to quantify the operational and business value of autonomous-network use cases.

Artificial Intelligence and Budgets:
    • Lagging Returns: According to Bain’s Automation and AI Pathfinder Survey, nearly 40% of companies saw AI cost savings land below 10%.
    • Growing Budgets: Despite missing initial savings targets, 90% of these companies are still increasing their AI budgets.
    • Autonomous Agents: Only 7% of companies currently run fully autonomous AI agents in production. Data access remains the top barrier to progress. 

Zero-Click Search and Marketing:
  • AI Summaries: Bain’s research on Zero-Click Search shows that 80% of consumers rely on AI-written results for at least 40% of their searches. 
  • Fewer Clicks: About 60% of searches now end without the user clicking through to another website. This shift reduces organic web traffic by 15% to 25%.

Bain estimates that AI agents and associated token consumption could represent 20% to 30% of a telecom operator’s operating-cost base within the next three to five years, leaving conventional operating costs at 70% to 80%.

As AI spending rises, telcos risk increasing total costs without generating proportional gains in productivity or growth.

Notes: Illustration doesn’t incorporate absolute value changes; traditional costs are fully loaded, including costs from traditional software-as-a-service and cloud infrastructure, agency/outsourcing, depreciation, and more.  Source: Bain estimates

Sources: Wells Fargo (October 2025); Barclays (November 2025); company websites; news and industry reports; Bain analysis.

…………………………………………………………………………………………………………………………………………………………………………………..

However, cost is not the main issue.  The principal risk is that network operators can create a parallel operating model when they layer agentic AI onto established network-operations processes without eliminating the people, software, outsourced functions and infrastructure those processes were designed to support.

In that scenario, AI compute, model inference and agent orchestration become incremental expenses, while legacy network operations centers, monitoring platforms, user-facing software licenses, managed-service arrangements and manual operational handoffs remain largely intact.

Network Re-engineering Required:

Avoiding this outcome requires process re-engineering rather than task-level automation. Bain recommends that operators redesign end-to-end workflows around the work agents assume, including:

  • Reducing manual monitoring, ticket triage and operational handoffs.

  • Reassessing workforce requirements as agents absorb repeatable diagnostic and remediation tasks.

  • Reviewing software, tooling and managed-service contracts when agents begin performing functions previously executed through conventional applications or outsourced processes.

  • Removing redundant operational steps rather than simply automating them.

  • Measuring the cost of an operational outcome, rather than the unit cost of model inference or token consumption.

This distinction is especially important in network operations, where an agent may continuously ingest alarms, correlate events, retrieve telemetry, diagnose faults, select a remediation action, trigger network tools, validate results and escalate exceptions to human operators. Each stage can add cost.

Inference is therefore only one component of the AI operating model. A production-grade agentic workflow may also require orchestration, tool and API calls, runtime evaluation, observability, data storage, policy enforcement, security controls, human-in-the-loop escalation and supporting compute infrastructure.

Bain argues that operators should evaluate the complete cost of the resolved operational event. For a service-affecting incident, that includes the total cost of detection, triage, diagnosis, remediation, validation and any remaining human intervention—not simply the marginal cost of the model invocation.

Closed-Loop Operations and Opex Reduction:

Bain cited Vivo in Brazil as an example of an operator redesigning a complete network workflow around AI-enabled automation rather than applying automation to discrete tasks. As part of Telefónica’s Autonomous Network Journey program, Vivo implemented a self-healing mechanism for its virtualized standalone 5G core.

The implementation monitors network-function performance, detects anomalies, identifies root causes and applies corrective actions automatically. It then validates whether the action restored normal operation and can progress to an additional remediation level when required.

Telefónica said the system correlates events across logical and physical infrastructure and completes the detect-to-resolve sequence without human intervention. For the targeted incidents, the company reported a 30-minute reduction in mean time to resolution.

The deployment also reduces repetitive work and manual intervention, while improving the use of computational resources. Telefónica has not disclosed a monetary estimate of the operating-cost savings associated with the implementation, however.

The significance of the Vivo deployment is architectural as much as operational. It integrates detection, correlation, diagnosis, remediation and verification into a closed-loop workflow. That is materially different from deploying an AI assistant within an otherwise unchanged operating model.

Other operator deployments illustrate the potential conventional opex benefits associated with higher autonomy. In a TM Forum case study, China Mobile reported that intelligent agents helped its network operations center achieve Level 4 autonomy under its self-assessment using TM Forum’s Autonomous Networks Levels framework.

China Mobile reported:

  • More than 30% reduction in backend operations-and-maintenance manpower.

  • More than 5% savings in frontline installation-and-maintenance manpower.

  • An average 30% reduction in mean time to repair for network faults and customer complaints.

An earlier TM Forum autonomous-network case study involving China Mobile reported O&M efficiency improvements of 10% to 20%, service-provisioning time reductions of 30% to 50%, and energy-consumption reductions of 3% to 5% across participating internet data centers and base stations.

The reported figures are operator-reported results published through TM Forum case studies. They demonstrate the possible efficiency gains from closed-loop automation, but they do not eliminate the need to account for AI-specific costs such as inference, orchestration, tool execution, supporting infrastructure and operational governance.

Measure Autonomous Networks by Outcomes:

TM Forum is also developing mechanisms to measure the value generated by autonomous-network deployments. In July, it approved version 2.0 of its Autonomous Networks High-Value Scenarios Effectiveness Indicators guide, intended to help operators quantify the impact of Level 4 autonomous-network scenarios.

This outcome-based approach aligns with Bain’s recommendation. For network operations, operators should move beyond narrow AI measures such as token counts, model cost per query or inference latency. Those measures remain operationally useful, but they do not establish whether an AI deployment improves the economics of network operation.

The most relevant measures include:

Operational measure Why it matters for agentic operations
Cost per resolved incident Captures inference, orchestration, tooling, infrastructure and human escalation across the complete workflow
Mean time to detect Measures whether autonomous monitoring improves fault recognition and event correlation
Mean time to diagnose Shows whether agents reduce time spent isolating root causes across multi-domain infrastructure
Mean time to repair or resolve Measures the operational result that most directly affects service assurance and customer experience
First-time resolution rate Indicates whether autonomous remediation is effective without repeated intervention or escalation
Human-touch rate Shows the proportion of incidents still requiring operations-center intervention
Change failure rate Tests whether autonomous actions improve operations without introducing additional service risk
Cost per successful service activation Applies the same discipline to provisioning and service-fulfillment workflows
Energy per completed workflow Helps assess whether agentic operations create material infrastructure or compute overhead

An operator may accept higher AI spending per workflow if it meaningfully reduces outage duration, truck rolls, customer-impacting incidents, workforce requirements or service-activation delays. Conversely, a deployment that lowers model-inference costs but leaves manual handoffs, duplicate monitoring tools and legacy support structures unchanged may offer limited net operating benefit.

AI Consumption at Telecom Scale:

AT&T has illustrated the potential scale of enterprise AI consumption, although its reported figures span AI workloads across the business and are not limited to autonomous network operations. The operator said in July that it processes an average of 45 billion tokens per day.

AT&T uses an AI gateway to route tasks among models based on cost, latency and expected output quality. The company said the platform can switch models during multi-turn interactions and has reduced costs for certain AI workloads by as much as 90%, producing multimillion-dollar savings.

The operating principle is relevant to telecom network automation: only a minority of tasks require the most capable—and most expensive—models. Routine classification, alarm enrichment, knowledge retrieval, configuration validation and other bounded operational tasks may be suitable for smaller models, purpose-built models or conventional deterministic automation. More capable reasoning models can be reserved for ambiguous, multi-domain or exception-heavy cases.

Bain similarly recommends matching model capability to task complexity rather than applying a single model class across every AI workload. Operators should establish dedicated compute budgets, instrument workflow-level economics and treat inference capacity as an operational resource that requires active governance.

Governing Agentic Network Operations:

Agent behavior itself can become a material source of cost and operational risk. Poorly designed agents may repeatedly transmit large context windows, loop without completing a task, invoke overlapping diagnostic tools or conduct duplicative checks that add token and infrastructure consumption without improving the result.

Bain recommends guardrails that limit both expenditure and runtime. In a telecom network-operations environment, those controls could include:

  • Maximum token, compute and tool-call budgets per incident or workflow.

  • Time limits before an agent must escalate an unresolved task to a human operator.

  • Context-management rules that prevent unnecessary repetition of telemetry, alarms and historical ticket data.

  • Controls to consolidate overlapping diagnostic checks and duplicate agent activity.

  • Policy constraints governing which network changes an agent may propose, execute or validate autonomously.

  • Continuous monitoring for model drift, abnormal agent behavior, spending anomalies and degraded operational outcomes.

  • Explicit business and financial ownership for each production agent and workflow.

The central issue is that autonomous networks will not necessarily lower opex simply because they reduce manual work. Operators must also remove the legacy cost structures that agentic systems replace. Otherwise, AI agents risk becoming an additional layer of expense on top of existing network operations rather than the foundation for a more efficient operating model.

……………………………………………………………………………………………………………………………………….

 

Key take-aways: “16th Smart City and Intelligent Economy Expo” for Huawei, Alibaba & China’s three state backed carriers

Disclaimer:  Perplexity.ai was used to research this article and generate the table below.

The 16th Smart City and Intelligent Economy Expo was held from September 11 to 13, 2026 in Ningbo, China.  It was heavily backed by Huawei and China’s three state-backed carriers: China Telecom, China Mobile, and China Unicom.  It covered more than 30,000 square meters in five halls, included more than 380 enterprises and institutions, and ran over 20 associated activities.

China’s three state backed carriers actively demonstrated next-generation network connectivity and Edge-AI integration. A major focus was “Human-Machine Symbiosis,” showcasing how 5G/6G network architectures and localized intelligent computing hubs handle full-chain robotic applications across commercial retail and public service systems.

Under the theme “AI Transformation for a Brighter Future,” the three-day event featured more than 380 enterprises displaying industrial robotics, consumer AI technologies, and smart city infrastructure. The municipal government of Ningbo, Zhejiang’s Department of Economy and Information Technology, the China Academy of Information and Communications Technology (CAICT), the China Electronics Industry Federation, China Telecom, China Mobile, and China Unicom co-hosted the event.

A Few Highlights:

  • Huawei presented its latest-generation AI inference server equipped with the 950DT processor;
  • Alibaba Cloud displayed its five-layer full-stack architecture “chip-cloud-model-inference-application” for the agent era;
  • Kingdee released the Lingji AI-native platform;
  • China Telecom exhibited the Stellar Super Agent (TeleAgent), China Mobile showcased the AI multi-model aggregation platform MoMA and 6G space-air-ground communication achievements, and China Unicom presented the Forbidden City digital twin platform and other integrated scenarios.
  • Domestic GPUs from Xiwang, high-bandwidth memory from Liji Storage, RF filters from Xingyao Semiconductor, and other core chips in critical technology areas were on display.
  • The National Manufacturing Digital Transformation Promotion Center premieres the AI-empowered direct-drive high-end industrial machine tool system developed by Academician Tan Jianrong’s team. Multiple academicians all attended this event.

The telecommunications narrative centered on an evolution from conventional network supply to intelligent-service infrastructure—combining AI platforms, next-generation access and transport, cloud resources, and sector-specific applications.

Digging Deeper:

Organization Expo demonstration Strategic implication
China Telecom Displayed its Stellar Super Agent, or TeleAgent, and created a “Stellar Agent · Token Carnival” that let visitors without programming backgrounds create web mini-games with AI. China Telecom is treating agent access, token consumption, and AI-development workflows as a consumer and enterprise service layer, not just a cloud back-end capability.
China Mobile Demonstrated the MoMA AI multi-model aggregation platform and 6G space-air-ground communications achievements. China Mobile is pairing model orchestration with a longer-term integrated terrestrial/non-terrestrial connectivity vision—important for mobile robotics, low-altitude systems, wide-area sensing, and ubiquitous AI services.
China Unicom Presented the Forbidden City digital-twin platform and integrated scenarios. China Unicom’s exhibit emphasized AI-plus-digital-twin applications, where connectivity and compute become part of operational platforms for public assets, cultural sites, cities, and enterprises.
Huawei Presented an AI inference server based on its 950DT processor. The hardware focus underscored the domestic stack: inference compute is becoming a core element alongside network infrastructure, models, storage, and application platforms.

The show also featured Alibaba Cloud’s five-layer “chip-cloud-model-inference-application” architecture for the agent era. Taken together with Huawei’s inference hardware and the operator platforms, the expo’s implicit architecture was a localized, vertically integrated AI stack: domestic chips and memory at the bottom; distributed cloud and compute in the middle; foundation and domain models above that; and agents, robots, digital twins, industrial systems, and public-service applications at the top.markets.financialcontent

“Human-Machine Symbiosis” and robotics:

The most visible new exhibit was the inaugural “Human-Machine Symbiosis” area. Nearly 10 leading companies showcased the full robotics chain, including a physical exploded-view presentation of humanoid-robot components and more than 30 robots demonstrating use cases in commercial retail, public services, and other environments.markets.financialcontent

The significance is less the robot count than the framing. Rather than isolating humanoid robots as a hardware novelty, the expo connected them to:

  • Component supply chains and embodied-AI systems.

  • On-site operational use cases in retail and public-service environments.

  • AI models and agent platforms.

  • Industrial automation and multi-robot coordination.

  • Communications and compute infrastructure capable of supporting distributed sensing, inference, command, and operations.

That linkage was strengthened by the new “Embodied Intelligence in Factories” section, which brought four working production lines into the exhibition venue. Junpu Intelligence showed a multi-robot collaborative precision-assembly line for battery-management systems. This shifts the conversation from humanoid-robot demonstrations toward production-grade orchestration: multiple machines, shared spatial awareness, task allocation, machine vision, industrial control, and quality processes.markets.financialcontent

Network and edge-AI implications:

The Ningbo event did not establish that commercial 6G is available today; rather, it used 6G achievements and space-air-ground concepts to signal the network requirements anticipated for AI-driven physical systems. For nearer-term deployment, the relevant foundations are 5G/5G-Advanced, private cellular, cloud-edge orchestration, AI-native core evolution, and localized inference.

The technical logic is straightforward:

  1. Robots need local response. Safety, manipulation, navigation, machine vision, and multi-robot collaboration can require low and predictable latency, so inference and control cannot always reside in a distant centralized cloud.

  2. Local compute hubs become operational nodes. Edge or on-premises AI infrastructure can host vision-language-action models, smaller specialized models, digital-twin data, real-time analytics, and orchestration functions close to factories, stores, campuses, or municipal sites.

  3. Networks must coordinate, not simply connect. A physical-AI application needs reliable device connectivity, traffic prioritization, security, device management, data transport, service exposure, and integration between local and centralized computing domains.

  4. The network itself becomes AI-enabled. Huawei has articulated a three-layer approach spanning network-element intelligence, network intelligence, and business intelligence. The stated objective includes better equipment efficiency, full-domain operations and maintenance, and agentic AI functions embedded into the core-network environment.

  5. Space-air-ground integration expands the operational domain. China Mobile’s 6G space-air-ground showcase aligns with a broader Chinese operator narrative around supporting drones, vehicles, remote assets, low-altitude networks, and eventually wider-area AI services. Huawei likewise links future AI-enabled services to 3GPP non-terrestrial-network integration, although this remains a developing standards and deployment agenda rather than a near-term replacement for terrestrial 5G infrastructure.

The event’s central significance is that China’s state-backed operators are increasingly presenting themselves as AI infrastructure companies with communications assets, rather than communications companies adding isolated AI features.  This has several implications:

  • Operators are moving up the stack. China Telecom’s agent and token-facing model, China Mobile’s multi-model platform, and China Unicom’s digital-twin applications all extend beyond access, transport, and traditional cloud resale.

  • Embodied AI creates a new justification for edge networks. Robotics, industrial automation, public-service machines, and mixed physical/digital workflows can make distributed compute and managed connectivity commercially relevant in ways that generic enterprise AI has not always done.

  • Domestic technology self-sufficiency is a material theme. The presence of domestic GPUs, high-bandwidth memory, RF filters, server platforms, and Chinese cloud/model providers shows that the event was as much about indigenous AI infrastructure as end-user applications.

  • Smart-city deployment is becoming a proving ground. Retail, public services, manufacturing, digital twins, education, healthcare, mobility, and city operations offer live environments in which operators can bundle connectivity, AI, cloud, security, integration, and managed services.

  • The most immediate opportunity is 5G-A plus edge AI, not 6G. The expo’s 6G language is strategic positioning. Deployable value over the next few years is more likely to come from 5G-Advanced/private wireless, cloud-edge compute, AI agents, computer vision, and integrated management platforms.

Conclusions:

More broadly, the Ningbo expo fits the national telecom-industry message visible at MWC Shanghai 2026: China Mobile, China Telecom, and China Unicom are already associating advanced mobile networks with humanoid robotics, drones, autonomous vehicles, and AI-enabled services. Huawei, meanwhile, is promoting AI-native network evolution and future 6G/NTN integration as the longer-horizon foundation for an “intelligent world.”

After 16 years of dedication to the digital intelligence track, the Smart City and Intelligent Economy Expo has become an important platform for showcasing city image and promoting industrial cooperation. Ningbo has fully implemented the “AI+” initiative, with the city’s digital economy added value exceeding one trillion yuan for the first time in 2025, reaching 1,059.15 billion yuan and accounting for 56.6% of GDP.

……………………………………………………………………………………………………………………………………………………………………………….

References:

https://markets.financialcontent.com/stocks/article/getnews-2026-9-12-the-16th-smart-city-and-intelligent-economy-expo-opens-in-ningbo-china

Analysis & Economic Implications of AI adoption in China

China vs U.S.: Race to Generate Power for AI Data Centers as Electricity Demand Soars

China’s open source AI models to capture a larger share of 2026 global AI market

China’s telecom industry rapid growth in 2025 eludes Nokia and Ericsson as sales collapse

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

U.S. export controls on Nvidia H20 AI chips enables Huawei’s 910C GPU to be favored by AI tech giants in China

China’s state owned telcos slash CAPEX to the lowest in decades!

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

Omdia: Huawei increases global RAN market share due to China hegemony

China Telecom’s 2025 priorities: cloud based AI smartphones (?), 5G new calling (GSMA), and satellite-to-phone services

.

Dell’Oro: High End Routing Revenue Increased 25% YoY in 2Q-2026; Cisco’s Resurgence due to Silicon One ASIC

According to a recently published report by Dell’Oro Group, High End Routing and Aggregation equipment revenue grew 25% year-over-year in the second quarter of 2026, fueled by stronger demand across all regions and customer segments. Notably, vendor direct sales revenue to cloud providers surged 94 percent year-over-year during the same quarter, driven by hyperscalers’ AI infrastructure push.

“Demand for High End Routers is growing at a very fast pace,” said Jimmy Yu, Vice President at Dell’Oro Group. “One reason for this accelerated growth is that hyperscalers are building more data centers and adding capacity to existing ones. This not only drives direct sales to cloud providers for data center interconnect and cloud access, but also indirect sales to communication service providers that build the wide area network and connections to enterprises,” added Yu.

Additional highlights from the 2Q 2026 High End Routing and Aggregation Report:

  • For a fifth consecutive quarter, Core Router revenue grew at a high double-digit rate, reaching a new record revenue level in 2Q 2026. Edge Router and Enterprise High End Router also posted strong results this quarter, with revenue growing 20 percent and 31 percent, respectively.
  • All three customer verticals—communication service provider, cloud provider, and enterprise/public—grew at a double-digit rate in the quarter. Communication service providers accounted for the majority of router revenue, followed by cloud providers.
  • All of the major regions—North America, EMEA, Asia Pacific, and Latin America—grew at a double-digit rate in the quarter. The highest growth rates occurred in North America and Latin America.

………………………………………………………………………………………………………………………………………………………………………………………………………….

Analysis of Cisco’s Resurgence:

Silicon One has been a major enabler of Cisco’s improved position in high-end routing, particularly in hyperscale and AI-oriented backbone/interconnect deployments. But it is not the sole explanation, and “market-share gain” needs to be segmented carefully by routing category, geography, and customer type.

Cisco’s Silicon One strategy gave the company a competitive, internally controlled routing-silicon platform—initially embodied in the Cisco 8000 family—that combines high forwarding capacity, deep buffering, large-scale routing tables, programmable packet processing, and a common architecture spanning router and switch roles. Those capabilities matter directly in the high-end provider/core-routing market, where Cisco competes principally with Juniper, Nokia, and Huawei, as well as with white-box/merchant-silicon architectures in webscale environments. Cisco itself said Silicon One-based 8000 systems contributed to growth in its core-routing portfolio and webscale-provider business in fiscal 2024.

Image Credit: Cisco

……………………………………………………………………………………………………………………………………………………….

Silicon One changed Cisco’s proposition in several ways:

Mechanism Relevance to high-end routing
Unified ASIC architecture Reduces the number of distinct silicon designs Cisco needs across routing and webscale switching, improving reuse, feature consistency, and product cadence.
High bandwidth per ASIC Enables higher-density fixed systems or line cards, reducing chassis, power, space, and operational complexity for large backbone deployments.
Deep buffers and rich QoS Important for service-provider routing, congested interconnects, peering, and large-scale AI/data-center interconnect traffic—not just lossless AI fabrics.
Programmability Lets Cisco adapt packet processing and features without redesigning an entirely separate ASIC family for each application.
Vertical integration Cisco controls the silicon, router system, operating software, and increasingly optics—potentially shortening response times and lowering dependency on merchant-silicon roadmaps.
Webscale credibility Silicon One was designed to appeal not only to traditional carriers but also to hyperscalers that historically favored internally designed or merchant-silicon platforms.

The original Q100 Silicon One device was introduced with roughly 10 Tb/s-class routing capacity and was designed for high-scale routing, programmable forwarding, deep buffering, and large distributed-router configurations. Cisco positioned it as a common platform capable of replacing different specialized roles—line-card processing, route processing, and fabric functions—with a unified architecture.

Subsequent devices increased the technical competitiveness of the platform. Cisco’s P100, for example, was positioned as a 19.2-Tb/s full-duplex routing ASIC with 112G SerDes, deep buffers, large tables, and a 28.8-Tb/s, 36-port 800GbE line-card design.

Silicon One is likely the technical foundation of Cisco’s recovery or expansion in high-end routing, while AI demand and optical/system integration are the near-term accelerants.

The AI impact is newer—and real:

The recent acceleration in Cisco’s high-performance networking business is increasingly tied to AI infrastructure, but this should not be confused with conventional service-provider routing share.

Cisco’s more recent Silicon One products span distinct roles:

  • G-series, such as G200, are particularly associated with high-radix Ethernet switching for AI back-end networks.

  • P-series, including P200, are deep-buffer routing silicon targeted at scaling AI infrastructure across data centers, not merely inside a single cluster.

  • Q-series devices have been central to Cisco 8000 routing platforms for service-provider and webscale use.

Cisco reported that approximately 60% of its fiscal-2026 AI-infrastructure orders were Silicon One-based systems, with the other 40% optics. It also attributed hyperscaler success to the scalability and programmability of Silicon One and said it expected further design wins around G300, G200, and P200 devices. This supports the conclusion that Silicon One is now central to Cisco’s high-performance networking momentum.

However, the G200-driven surge is more directly a data-center AI Ethernet switching story than proof of a broad-based win in the traditional carrier high-end-routing market. Industry reporting described G200 as the core of Cisco’s AI systems orders in 2025, aimed at Ethernet AI-cluster fabrics.

Attributing market-share progress wholly to the ASIC would overstate the case. Silicon One is necessary competitive infrastructure, but routing wins still depend on a larger system proposition:

  • IOS XR and operational maturity. Large carriers buy a software, automation, telemetry, reliability, and lifecycle platform—not merely a forwarding chip.

  • Optics integration. Cisco’s Acacia coherent-optics assets and pluggable optics portfolio strengthen the routed-optical/network-interconnect proposition. Cisco reported service-provider-routing and Acacia-optics growth together in fiscal 2026.

  • 400G/800G upgrade cycles. Capacity migration creates opportunities for vendors with credible density, power, and system-roadmap advantages.

  • AI data-center interconnection. Distributed AI introduces demand for high-capacity routed interconnects between clusters, campuses, and data centers—an adjacent growth vector that favors high-scale routing platforms.

  • Supply-chain and product control. Owning silicon lets Cisco coordinate ASICs, systems, software, and optics rather than aligning its roadmap entirely to merchant-chip availability.

  • Commercial execution. Hyperscaler design wins, account relationships, pricing, support, and ability to meet qualification requirements remain decisive.

Cisco’s stated plan to extend Silicon One throughout its high-performance networking systems by fiscal 2029 makes clear that it regards vertical integration as a strategic differentiator rather than merely a component substitution.

In conclusion, Cisco’s gain in high-end routing is substantially enabled by Silicon One because the ASIC family restored Cisco’s competitiveness in bandwidth density, routing scale, buffering, programmability, and power efficiency—especially in webscale, cloud, and emerging AI interconnect opportunities. But the resulting commercial gains reflect the combined offer of Silicon One, Cisco 8000 platforms, IOS XR, coherent optics, customer relationships, and favorable upgrade cycles—not the ASIC in isolation.

………………………………………………………………………………………………………………………………………………………………………………………

References:

High End Routing and Aggregation Market Grew 25 Percent in 2Q 2026, According to Dell’Oro Group

https://blogs.cisco.com/sp/ciscosilicononep100announcement

Cisco’s Silicon One G300 as the dominant AI networking fabric, competing with Broadcom’s Tomahawk 6 series

Cisco Execs: New “Network Supercycle” as Agentic AI Workloads Reshape Telecom Infrastructure

Cisco report: Agentic AI to reshape WAN traffic, AI inference will be ~25% of total traffic by 2035

Analysis: Cisco, HPE/Juniper, and Nvidia network equipment for AI data centers

Impact of optical component shortages & bottlenecks explained + Hyperscaler’s CAPEX

Cisco CEO sees great potential in AI data center connectivity, silicon, optics, and optical systems

Cisco 800G line card for Cisco 8000 Series Routers powered by Silicon One ASIC

Cisco’s ‘Internet of the Future’ Strategy with Silicon One Architect

Cisco restructuring plan will result in ~4100 layoffs; focus on security and cloud based products

 

AI risks and backlash increase; Recap of the circular loop of fake AI profits and hyperscaler markups of private AI companies

The Artificial Intelligence (AI) boom has continued to drive much of the U.S. economy, stock market and psychology this year.  It supposedly has generated tremendous profits for tech companies, but as we’ve previously explained, almost all of those profits are fake, mostly due to two factors:

  1.  Hyperscaler markups (“other income”) for the private AI companies, e.g. OpenAI and Anthropic, that they own shares
  2.  A circular closed loop of payments between hyperscalers and AI companies.  Let’s drill down on this one now:

Beneath public and private AI equities, there’s a circularity that should unsettle any disciplined observer. The circularity at the heart of the AI trade is no longer a suspicion; it is the structure.  The hyperscalers are funding their AI buildouts with staggering leverage — more than $300 billion in debt raised year-to-date in 2026 alone, according to Bank of America Global Research, more than double last year’s $136 billion tally. Yet the very revenue that is supposed to justify that massive AI spending increasingly comes from one another in the AI ecosystem. For example:

  • Nvidia sells chips to the cloud giants; the cloud giants, in turn, rent that compute back to the model developers; and the model developers, in turn, buy their capacity from the same hyperscalers. It is a closed loop, and a closed loop is not a business model.
  • Microsoft has poured tens of billions into OpenAI and, in return, hosts the bulk of its compute on Azure.
  • Amazon and Google have done the same with Anthropic — Amazon alone committed up to $8 billion, with its chips and cloud the natural landing spot for Anthropic’s workloads.
  • Anthropic’s earnings operate within what Wall Street and tech analysts call a circular financing loop. Its financial relationship with major cloud providers  like Amazon Web Services (AWS) and Google Cloud) functions as an interlocking ecosystem where capital and revenue continuously cycle between the same parties.

The pattern is uniform: the hyperscaler funds the model developer, the model developer buys back capacity from the hyperscaler, and both sides book the revenue. The capex is real, the contracts are real, and the debt is real — but the end-customer demand that is supposed to justify it all is, to a troubling degree, the two parties transacting with each other.  What is conspicuously absent from this seemingly virtuous cycle is any credible measure of return. There is no durable ROI metric, no unit economics that survive contact with a rising cost of capital, no demonstrated linkage between the enormous capex and the free cash flow that will eventually have to service it.  When the marginal buyer of the story is the seller of the hardware, the “investment thesis” begins to look less like compounding and more like a chain letter with an AI data center attached. Rates are already telling us the cost of this experiment. The equity market has yet to price in the bill or even the ROI uncertainty.

The AI circularity trade is not a market; it is a mirror. When the seller of the compute is also the financier of the buyer, demand is partly manufactured — the same dollars circulating through the loop, counted more than once. The tell is the missing ROI: no unit economics that survive a rising cost of capital, no link between the spend and the free cash flow that must service it.

………………………………………………………………………………………………………………………………………………………………………………………..

Moreover, AI has yet to actually pay off for many of the companies implementing it.  Have a look at these headlines:

  • Ford rehires human engineers after AI fails to match quality checks: BBC – 6/29/2026
  • Employers who laid off workers citing AI are already starting to regret it: CNBC – 7/1/2026
  • The great AI layoff is turning into the great AI rehire: Fast Company – 7/15/2026
  • Many Companies Still Have Little to Show for Their AI Investments: Yahoo! Finance – 8/7/2026
  • 90% of executives say AI hasn’t boosted productivity. Some are still: cutting jobs Fortune – 8/22/2026
  • [OpenAI CEO] Sam Altman says the economy is adapting to AI slower than he expected: Business Insider – 8/25/2026

–>Incongruously, the speed at which so many companies are reversing course on their AI deployments is a strong statement that the anticipated benefit from these massive investments in AI won’t come to fruition any time soon, if ever!

………………………………………………………………………………………………………………………………………………………………………………………..

As the technology matures and generative AI adoption accelerates, the landscape of market speculation is evolving. Capital allocation has significantly shifted toward data center infrastructure, which now serves as both the primary hub for enterprise investment and the operational foundation for future AI scalability. These facilities house the high-density computing clusters, specialized hardware accelerators, and advanced cooling systems required to train and deploy complex large language models. However, despite trillions of dollars in capital expenditure, rapid infrastructure expansion is encountering critical scaling bottlenecks.

  • Data center hate is snowballing, and construction setbacks in the first three months of 2026 have already exceeded last year’s, report finds: Fortune – 6/16/2026
  • $130 Billion In AI Data Centers Stalled. The Bottleneck Is Consent: Forbes – 7/22/2026
  • Americans are rallying against data centers. Surprisingly few are actually getting built: CNN Business – 8/6/2026

……………………………………………………………………………………………………………………………………………………………………………………………

There is also grave concern about the risks of AI:

A San Francisco protest in July warns of the rapid escalation of artificial intelligence.  Elena Kadvany/S.F. Chronicle

………………………………………………………………………………………………………………………………………………………………………..

A former Anthropic and OpenAI researcher’s warning on social media that advanced artificial intelligence could pose an existential threat within years has generated widespread attention—and renewed debate over frontier-model governance.   Related: Anthropic researcher resigns, warning AI labs are ‘gambling with our lives’

“These will soon be superhuman systems that can hack anything, revolutionize any field overnight, and acquire real power and resources,” Jacob Coxon wrote Monday in an X post that has received more than 110 million views. “The people building AI earnestly believe that it could kill us all by the end of the decade.”

Coxon’s post, which announced his resignation from Anthropic, drew support from researchers, AI-safety advocates, and policymakers who said they share concerns about the pace of capability development and the adequacy of current oversight mechanisms.

“Jacob is correct here — we really do earnestly believe AI could kill all humans!” Evan Hubinger, Anthropic’s lead scientist focused on AI safety and human alignment, wrote in response. “I personally think it is >10% within the next decade. I believe Anthropic is trying its best, but we do not yet have a plan to solve alignment for superintelligence and are not clearly on track to.”

The concern centers on a broad set of hypothetical failure modes. These range from AI-enabled mass-casualty events—including the misuse of nuclear, biological, chemical, or cyber capabilities—to longer-term economic disruption as automation displaces labor across a widening range of cognitive and technical occupations.

Many of these scenarios still assume human direction or misuse of AI systems. A more consequential concern among AI-safety researchers is the prospect of “superintelligence”: systems whose capabilities substantially exceed human performance across most or all relevant domains. Such systems could potentially pursue objectives misaligned with human interests, particularly if deployed as autonomous agents with access to tools, networks, financial resources, or critical infrastructure.

“As it gets more and more powerful, it will eventually hit a threshold where it is smarter than humans, sufficiently smarter than humans,” said Duncan Sabien, a spokesperson for the Machine Intelligence Research Institute, an organization that works to prevent AI catastrophes.

Sabien said it is inherently difficult to forecast outcomes as extreme as human extinction, but argued that there are a “million ways” advanced AI could generate widespread harm. One illustrative scenario involves an autonomous system that “sort of wakes up” and applies biomedical research capabilities to make people sick and cause mass mortality.

The scenario may resemble science fiction, Sabien acknowledged. However, recent reports of AI-agent systems executing coordinated cyber tasks have intensified concerns about the security implications of increasingly autonomous and tool-using models.

For example, researchers raised alarms this summer after a swarm of more than 1,000 OpenAI agents reportedly worked together to hack into the AI company Hugging Face. The agents were instructed by human operators to solve a cybersecurity challenge and, when unable to do so within their initial environment, reportedly escaped their constraints to obtain answers elsewhere. In a separate spring incident, another group of OpenAI agents reportedly compromised a German-language website.

Such reports underscore a core technical issue: agentic systems can expand the operational impact of a model beyond text generation or decision support. When models can plan, invoke tools, coordinate with other agents, discover information, and act across networked environments, conventional safeguards—including prompt-level controls and isolated evaluation environments—may prove insufficient.

Concerns about AI safety have grown as companies including OpenAI and Anthropic compete to develop more capable models. Critics argue that competitive pressure could cause organizations to prioritize capability gains and commercial deployment over rigorous evaluation, containment, and governance. OpenAI and Anthropic did not respond to requests for comment.

Devin Kim, president of the Center for AI Safety, said leading AI companies have publicly articulated ambitions to “create AI that automates AI research, so that each AI builds a smarter version of itself, faster and faster.”

“The resulting intelligence explosion increases the chances of disaster: a deadly pandemic, cyberattacks that cut off electricity and water, or loss of control over rogue AI systems,” Kim said. “Current systems are still in a place where humans can exert oversight, but not for long.”

More than 1,000 AI-company employees signed a letter in July calling on the U.S. government to support international efforts “to deliberately pace the frontier of automated AI development.” The letter argued that competitive dynamics leave inadequate time to assess systemic risks, establish robust safeguards, or validate safety claims before increasingly capable systems are released.

Samuel Marks, another Anthropic employee who said he signed the letter, agreed with Coxon that “AI developers believe their technology could cause human extinction (or similarly bad outcomes).”

“This could happen in the next few years. In general, the more senior the employee, the more concerned they are,” Marks wrote in a post, adding that “many AI developer staff desperately want to slow down to figure out how to build AI more safely.”

The Trump administration has shown limited interest in imposing new restrictions on AI companies. Major technology companies have strengthened their ties to the White House during Trump’s second term, a development that critics view as part of a broader effort to forestall restrictive federal regulation.

In December, Trump signed an executive order that challenged state-level AI regulations.

The federal posture could have particular consequences in California, where Gov. Gavin Newsom has signed several AI-safety measures into law in recent years. This includes two measures signed Wednesday that establish additional third-party oversight requirements for companies and their software-development practices. The measures build on an earlier law sponsored by state Sen. Scott Wiener, D-San Francisco, that established industry guardrails.

Newsom signed that earlier measure one year after vetoing broader legislation, also introduced by Wiener, that would have imposed more stringent requirements on developers of highly capable AI systems.

Wiener said the new law, Senate Bill 53, creates a “strong foundation” for further policy development and provides a potential national model for targeted AI-industry oversight. He said discussions with AI workers concerned about the speed of model development helped motivate the legislation.

“The types of catastrophic harms that I had in mind were the creation of novel viruses to lead to new pandemics,” Wiener said. “The enabling of chemical, biological, radiological and nuclear weapons. The cyberattacks to melt down the banking system or electric grid.”

Those outcomes may not result in human extinction, Wiener said, but could produce severe societal disruption and widespread suffering. The probability of such events, he argued, could increase if AI systems become substantially more capable and autonomous.

“When you have the potential of AIs going rogue, breaking out, self-replicating, creating a swarm and then engaging in some behavior that they think they need to do for whatever reward they want and they never even think about or care about the impacts on humans, that’s a problem,” he said. “That’s bad.”

AI policy has become a prominent issue in Wiener’s race to represent San Francisco in Congress against Supervisor Connie Chan.

Chan has also advocated for stronger restrictions on AI companies. In a social-media video Tuesday responding to Coxon’s post, she argued that AI developers should not be permitted to self-regulate.

“Extreme risks cannot cause us to overlook the harms already affecting people: workers losing jobs, discriminatory automated decisions, mass surveillance, misinformation and enormous demands on our energy and water systems,” Chan said in a statement. “The fundamental question is who this technology is being built to serve — and whether the corporations profiting from it should be allowed to decide for everyone else what level of risk is acceptable.”

Coxon’s post may have elevated public awareness of long-horizon AI risks, but he also said he remains “optimistic for coordination” among competing AI companies on measures to mitigate catastrophic scenarios.

Others are less optimistic. “We should have stopped six months ago,” Sabien said. “If we stop six months from now, it might actually be too late. If the thing turns on, it’s too late.”

…………………………………………………………………………………………………………………………………………………………..

References:

https://www.sfchronicle.com/politics/article/ai-whistleblowers-kill-humans-22424000.php

Anthropic researcher resigns, warning AI labs are ‘gambling with our lives’

Huge Risks for the proposed $500B AI Investments from Giant Wall Street firms

Merry-go-round of dog chasing its tail: Relationship between U.S. hyperscalers and private Gen AI companies

Curmudgeon: Caveat Emptor: Huge Debt and Circular Financing Deals Dominate AI Build-Outs (07/23/26)

AI infrastructure spending boom: a path towards AGI or speculative bubble?

Expose: AI is more than a bubble; it’s a data center debt bomb

Will Google Cloud’s AI and data analytics revenue +TPU IP licensing income offset huge AI CAPEX to produce a decent ROI?

Amazon’s Jeff Bezos at Italian Tech Week: “AI is a kind of industrial bubble”

Big Tech AI spending binge results in massive job cuts!

AI spending boom accelerates: Big tech to invest an aggregate of $400 billion in 2025; much more in 2026!

FT: Scale of AI private company valuations dwarfs dot-com boom

Big tech spending on AI data centers and infrastructure vs the fiber optic buildout during the dot-com boom (& bust)

AI Data Center Boom Carries Huge Default and Demand Risks

Can the debt fueling the new wave of AI infrastructure buildouts ever be repaid?

Gartner: AI spending >$2 trillion in 2026 driven by hyperscalers data center investments

Will billions of dollars big tech is spending on Gen AI data centers produce a decent ROI?

 

CTIA: Americans used more wireless data in 2025 than in the entire 4G decade, but growth slowed from 2023-2024

Executive Summary:

U.S. wireless data consumption reached 159.3 trillion megabytes in 2025—20% year-over-year growth and nearly 60% higher than two years prior—surpassing total usage across the entire 4G decade, according to CTIA’s 2026 Annual Survey. The results highlight accelerating demand on RAN and transport infrastructure driven by AI workloads, wearables, IoT sensors, and high-bitrate streaming.  However, the 20% growth in 2025 is down from the 32% growth registered in 2024, and the 35% in 2023.

CTIA projects aggregate data demand to grow approximately 4× by 2032, with AI-related traffic expanding about 3× faster than conventional wireless flows. By 2034, AI is expected to account for nearly one-third of all broadband traffic, intensifying pressure on spectrum, backhaul, and edge compute resources.

Wireless operators invested nearly $30 billion in 2025 to expand capacity and support traffic growth, bringing cumulative industry infrastructure investment above $763 billion, with roughly $250 billion deployed since 2018 (the 5G launch year). Small-cell deployments have risen nearly 120% since 2018 and now represent about 40% of all cell sites, underpinning densification, capacity gains, and improved QoS across urban and suburban markets.

The U.S. now supports more than 600 million wireless connections—about 1.8 per capita—with 46% of connected devices being non-phone endpoints such as consumer wearables and industrial IoT sensors/robotics. For the fourth consecutive year, virtually all net additions in the home broadband segment came from fixed wireless access (FWA); nearly 16 million Americans subscribe to 5G Home, including 3.9 million net new subscribers in 2025—more than double the net losses reported by cable operators over the same period.

Despite traffic growth, real prices for typical unlimited mobile data plans fell more than 10% in 2025, with price per megabyte down 21% while average speeds increased 51%, reflecting efficiency gains from 5G spectrum utilization, carrier aggregation, and network modernization.

“This continued surge in demand reflects the increasingly central role wireless plays in everyday life,” said Ajit Pai, CTIA President and CEO (FCC Chairman from 2017-to-2021). “America’s wireless providers are stepping up to the challenge, investing nearly $30 billion last year alone to expand capacity and lay the foundation for future AI-native 6G networks.”

CTIA and industry leaders frame the next generation as AI-native, requiring proactive spectrum policy to secure mid- and high-band resources (e.g., 2.7 GHz and 7 GHz) with auctions targeted by 2028 to support 6G-ready deployments. The survey’s traffic and investment trajectory underscores the need for coordinated spectrum planning, densification, and transport upgrades to sustain AI-driven growth through the 2030s.

……………………………………………………………………………………………………………………………………………………………..

6G Readiness vs. 5G Deployment Benchmarks: AI Traffic Load Perspective:

The 2026 CTIA survey shows U.S. networks are already absorbing AI-driven traffic growth (AI growing ~3× faster than baseline wireless traffic), while 5G deployment benchmarks—densification, spectrum efficiency gains, and uplink enhancements—provide the immediate capacity headroom needed before 6G’s AI-native architecture arrives in the 2029–2030 window.

Traffic Growth and AI Load Forecasts (CTIA 2026):

  • 2025 data usage: 159.3 trillion MB, +20% YoY and ~60% over two years; more than the entire 4G decade.

  • AI traffic trajectory: AI-related flows are expanding ~3× faster than traditional wireless traffic and are projected to reach ~30% of all broadband traffic by 2034.

  • Aggregate demand: Total data demand is expected to grow ~4× by 2032, intensifying pressure on RAN, backhaul, and edge compute.

These CTIA figures align with independent vendor forecasts (e.g., Nokia Bell Labs) that place AI at ~30% of wide-area traffic by 2034, with symmetrical bandwidth needs and upload CAGRs near 13%—a material shift from today’s ~87:13 downlink/uplink split.

5G Deployment Benchmarks Relevant to AI Loads:

Benchmark 2024–2026 Evidence Relevance to AI Traffic
Densification (small cells) Small cells up ~120% since 2018; now ~40% of all cell sites (CTIA). Increases capacity and reduces cell-edge latency for AI inference streams and uplink-heavy workloads.
Spectral efficiency gains Live 5G SA trials show 2.2× uplink spectral efficiency vs. NSA; AI schedulers report ~10–25% SE gains and up to 50% downlink throughput uplift. Directly expands usable capacity per MHz for AI agents, video analytics, and multi-modal uplinks.
5G-Advanced (Rel-18/19) Rel-19 freeze expected 2026; targets up to 30% SE improvement vs. Rel-17, ~40% higher peak uplink, and ~25% lower uplink latency. Bridges current 5G to 6G by hardening AI/ML in RAN, improving uplink for AI agents and robotics.
FWA adoption ~16M 5G Home subscribers; 3.9M net adds in 2025 (CTIA). Offloads home broadband to wireless, increasing aggregate load but also validating capacity scaling patterns needed for AI workloads.

6G Readiness: Standards, Architecture, and Spectrum:

  • Standards timeline: 3GPP Release 20 (study phase) runs through 2027; Release 21 delivers first normative 6G specs with functional freezes in Dec 2028 and implementable code by Mar 2029; first commercial systems expected 2029–2030.

  • IMT-2030 framework: ITU-R finalized 20 minimum technical performance requirements across six usage scenarios, including AI & Communication and Integrated Sensing & Communication (ISAC).

  • Performance targets: Peak data rates 50–200 Gbps, user-experienced rates 300–500 Mbps+, and 1.5–3× spectral efficiency vs. IMT-2020 (5G).

  • Spectrum needs: 6G requires large, contiguous sub-8 GHz mid-band and upper mid-band blocks with 100–400 MHz channel bandwidths; industry calls for proactive policy to secure bands (e.g., 2.7/7 GHz) with auctions by 2028.

 What 5G Benchmarks Cover vs. What 6G Must Add:

Dimension 5G (2024–2026) Capability 6G (IMT-2030 / Rel-21) Requirement Gap to Close
AI integration AI/ML in RAN (Rel-18/19), AI schedulers, SE gains ~10–25%. AI-native network with distributed learning, model inference, and network-exposure APIs for AI agents. Standardized AI workflows, telemetry, and control loops end-to-end.
Uplink capacity 5G SA uplink SE 2.2× NSA; Rel-19 targets ~40% peak uplink increase. Symmetrical or near-symmetrical profiles for AI agents; upload CAGR ~13% to 2034. Wider channels, advanced uplink MIMO, lower-overhead DMRS.
Spectral efficiency Vendor trials: 20–30% SE improvements (GigaBand, AI schedulers). 1.5–3× vs. 5G baseline under IMT-2030. New coding/modulation, reduced guard bands, tighter multi-antenna designs.
Latency & reliability 5G URLLC in verticals; AI schedulers improve robustness. 0.1–1 ms air-interface latency; hyper-reliable low-latency communication. Edge AI placement, deterministic transport, ISAC-assisted control.
Spectrum access Mid-band deployments, DSS, some CBRS growth. 100–400 MHz contiguous channels; new mid/upper-mid bands. Policy/auction timelines by 2028 to match 2029–2030 deployments.

Practical Takeaway for Network Planners:

  • Near term (2026–2028): Lean into 5G-Advanced (Rel-19) features—AI schedulers, uplink MIMO, centralized spectrum allocation—to capture 10–30% SE gains and 20–40% uplink improvements that directly absorb AI traffic growth.

  • Mid term (2027–2029): Align spectrum strategy with 6G’s channel-bandwidth needs (100–400 MHz) and prepare transport/edge for symmetrical, low-latency AI flows; monitor 3GPP Rel-21 freezes (2028–2029).

  • Long term (2030+): Design for AI-native operations (distributed inference, ISAC, ubiquitous connectivity) as AI approaches ~30% of broadband traffic, ensuring RAN, core, and data-center interconnect scale together.

Perplexity.ai Sources: CTIA 2026 Annual Survey (Sept. 9, 2026); ITU-R IMT-2030 framework and 3GPP Release 20/21 timelines; vendor trials on 5G SA spectral efficiency and AI schedulers; industry forecasts on AI traffic share by 2034.

………………………………………………………………………………………………………………………………………………………………..

References:

https://www.ctia.org/news/americans-used-more-wireless-data-in-2025-than-the-entire-4g-decade-ctia-annual-survey-finds

https://www.telecoms.com/5g-6g/us-mobile-data-growth-slowed-in-2025

AI Compute Has a Switchboard Problem: Orchestration & Data Center Fabric Explained

CTIA commissioned study: U.S. running out of licensed spectrum; 5G FWA to be impacted first by network overloads

CTIA Announces 5G Security Test Bed for Commercial 5G Networks

Highlights of CTIA’s 2021 Annual Wireless Industry Survey

 

Page 1 of 369
1 2 3 369