Industry Outlook: Telecoms & Connectivity — Week of July 27, 2026

Market Outlook

• AI data centers become core connectivity market. Verizon highlighted a $1 billion dark fiber deal with Google as the first of many AI data center connectivity wins, while SK Telecom created a dedicated AI data center subsidiary and AMD detailed large hyperscale AI deployments. Hyperscaler build plans still point to a doubling of US data center capacity in three years despite power and permitting constraints. Connectivity to AI infrastructure is shifting from an adjacency to a primary growth driver for large operators.
• Mobile gains offset broadband softness in US. Charter lost 172,000 broadband subscribers but added 406,000 Spectrum Mobile lines, and Comcast reported improved broadband losses alongside record wireless net adds. Verizon delivered its best Q2 in five years with 184,000 postpaid phone net adds and over 550,000 total mobility and broadband additions. Fixed-line pressure is persisting, but converged bundles and mobile growth are stabilizing overall revenue trajectories.
• Satellite rivalry pushes toward hybrid models. Charter is reported to be in talks with SpaceX about a potential partnership even as executives publicly sidestep Starlink speculation, while Comcast spent time on its Q2 call addressing satellite threat scenarios. T-Mobile remains tied to Starlink for direct-to-device (D2D) but is also advancing a terrestrial D2D joint venture with AT&T and Verizon, and Crown Castle reiterated that ground-based towers remain essential despite satellite hype. Market structure is moving toward integrated terrestrial plus non-terrestrial offerings rather than a pure displacement story.

Discussion: CTOs should treat AI data center connectivity and converged mobile-broadband bundles as core growth domains and start planning for integrated satellite-terrestrial propositions instead of viewing LEO purely as a threat.

Headwinds

• Consumer handset economics under mounting strain. Soaring memory prices are driving a global handset slowdown that is hitting budget brands hardest while advantaging Apple and Huawei. Recon Analytics found that handset deals are now the most disliked part of wireless, even more than billing support. Operators that depend on aggressive device promotions to drive subscriber growth face both margin pressure and growing customer dissatisfaction.
• Power and siting constraints for AI infrastructure. Synergy projects US data center capacity will double in three years, but notes growing power constraints and community opposition to new facilities. AI rack designs, such as AMD's Helios systems, are pushing power densities higher, which stresses both grid interconnects and telco edge facilities. Operators that want to win AI connectivity and edge hosting revenue must solve for power, cooling, and permitting, not just bandwidth.
• Workforce disruption as AI reshapes telco operations. Tata Communications is restructuring its workforce as it realigns talent around AI-driven networks, echoing a broader operator pattern. AT&T disclosed that its internal AI usage has crossed 1 trillion tokens per month and launched OTel 2.0 open source telco AI models, signaling rapid automation of planning, assurance, and customer operations. Engineering teams that do not reskill into AI-augmented workflows risk marginalization as vendors and hyperscalers move faster.

Discussion: CTOs should pressure test device-subsidy models, plan for power and site constraints in AI and edge roadmaps, and move quickly on AI skills, tooling, and organizational design before restructuring pressures arrive from the CFO and CHRO.

Tailwinds

• Mid-band spectrum policy opens 5G and early 6G runway. The FCC approved rules for an upper C-band auction and is moving toward a combined lower plus upper C-band 'Super Band', while also flagging a 2.7 GHz auction in 2028. A coherent mid-band plan gives US operators clearer visibility on capacity expansion for advanced 5G and early 6G use cases. Network planners now have firmer guardrails for long-term spectrum layering and small cell strategies.
• AI-centric networking drives vendor and operator upside. Nokia's AI and cloud networking business more than doubled in Q2, and the vendor still beat expectations while funding a significant restructuring. Verizon, AT&T, and SK Telecom all leaned into AI narratives, from AI data center connectivity to internal AI platforms and dedicated AI infrastructure subsidiaries. AI is no longer a slideware topic, it is starting to show up as measurable revenue and product lines in both vendors and operators.
• Fiber expansion and partnerships continue globally. US and Canadian operators such as altafiber, Spectrum, Comcast, Cogeco, and GCI reported fresh fiber and 5G buildouts in both urban and hard-to-serve regions, including Alaska road corridors. In Europe, Hyperoptic and a Belgian fiber collaboration are still securing regulatory approvals and capital, while Vecima partnered with Furukawa Electric to bring 10G-EPON FTTP gear into Japan. Fiber remains the anchor for backhaul, enterprise, and fixed wireless offload strategies, even as satellite and wireless grab headlines.

Discussion: CTOs should align spectrum roadmaps with the emerging Super Band and 2.7 GHz timelines, push AI-native networking initiatives from pilot to production, and keep fiber programs tightly linked to 5G, FWA, and edge compute strategies rather than treating them as standalone buildouts.

Tech Implications

• AI factories demand new transport and edge designs. Verizon's dark fiber deal with Google, SK Telecom's AI data center subsidiary, and AMD's Helios racks built on open Ethernet networking all point to a new class of 'AI factory' sites. These locations need deterministic low-latency transport, massive east-west bandwidth, and tight integration with cloud on-ramps and telco edge nodes. Traditional metro aggregation architectures will struggle unless operators adopt spine-leaf, open Ethernet fabrics, and automated optical provisioning.
• Satellite and D2D reshape radio and core integration. T-Mobile's continued Starlink exclusivity for D2D, combined with a parallel terrestrial D2D JV with AT&T and Verizon, creates a multi-path access environment that will need careful RAN and core integration. Charter's potential SpaceX partnership and Comcast's focus on satellite threat scenarios point to future hybrid CPE and device designs that can roam between terrestrial, LEO, and possibly GEO networks. Network architects will need to plan for non-terrestrial network (NTN) support in 5G standalone cores, updated roaming policies, and new billing and policy control flows.
• Operator AI platforms move toward open ecosystems. AT&T's OTel 2.0 open source telco AI models, trained on AMD silicon, show that operators are starting to externalize some of their AI tooling. Nokia's AI and cloud networking growth suggests that AI-based assurance, optimization, and slicing orchestration are gaining traction with carriers. Engineering teams will need to decide where to rely on vendor AI stacks, where to adopt open models, and how to govern data sharing across network APIs and AI platforms.

Discussion: CTOs should prioritize transport redesign for AI-heavy traffic, add NTN and D2D scenarios into 5G SA and device roadmaps, and define a clear architecture for AI platforms that balances vendor solutions with open and in-house models, including data governance and MLOps.

CTO Action Items

Revisit your three-year network plan with AI data centers as a named segment, including dark fiber, open Ethernet fabrics, and edge interconnect strategies, rather than treating AI traffic as generic cloud growth. Direct your architecture teams to model non-terrestrial integration, including D2D and LEO backhaul, in the 5G SA core, radio roadmap, and CPE/device specifications. Launch or accelerate an internal program on AI for network operations that evaluates platforms like AT&T-style open models versus vendor stacks, with a bias toward measurable opex savings within 12 to 24 months. Finally, align spectrum and fiber build priorities with the emerging US Super Band and 2.7 GHz timelines, and tighten mobile plus broadband bundling and device strategies in light of handset economics and rising consumer dissatisfaction with traditional handset deals.