2025 Communications & Media Year-End Review and 2026 Outlook
Dear Clients & Friends,
2025 was the year big bets from the past decade reached a turning point. Fiber is a fixture. 5G fixed wireless is no longer an experiment. The lines between linear, streaming, and social media consumption have blurred. And AI, after two years of headlines, is no longer a slide in the strategy deck – it’s a line in the P&L.
For operators, investors, and policymakers, three questions repeatedly came up in our work:
1. What creates sustainable advantages in markets where most households now have genuine choice?
2. As products, pricing, and content become more complex, how do you build winning customer experiences and maintain trust?
3. How do you move AI from pilots and promises to scaled, governed, revenue-generating and cost-saving capabilities?
The following is a reflection on the past year: what we saw, what we learned alongside you, and what we believe it means for the future of the communications and media industry.
Networks: From Scarcity to Abundance
Viewed across markets, the connectivity story in 2025 is remarkable. In the US, the Fiber Broadband Association (FBA) reports that fiber is now passing 58% of US households – a 13% increase year over year. In Europe, the FTTH Council Europe reports fiber-to-the-home coverage reaches nearly 75% of households. In the UK, reports show fiber is available to 79% of residential premises, with Northern Ireland at 95%.
The UK’s fiber boom arrived early, flooding the market with alt-nets that are now engaged in price wars and entering a rapid consolidation phase. By contrast, the US fiber market is maturing more slowly. Fragmentation persists, retail price competition is only now intensifying, and many regional fiber providers remain in growth mode, delaying the consolidation reflex that has already gripped the UK.
At the same time, fixed wireless has rapidly gone from novelty to mainstream in the US, with subscriber growth outpacing fiber growth in the first half of 2025. The technology now represents over 10% of US broadband subscriptions with NPS scores better than cable.
Greater access changes the game competitively. We analyzed the FCC’s latest Broadband Data and found 93% of US households now have at least one fixed provider offering 100/20 Mbps or better – the FCC definition of broadband. In fact, 65% have two or more, and remarkably, 30% have three or more.
The era of uncontested local internet service provider (ISP) dominance is over.
An abundance of new broadband projects, however, does not mean they are inexpensive to build. Our most recent Fiber Deployment Cost Annual Report with the FBA highlights how much of each new-build dollar still goes into the ground – or up the pole. Labor accounts for 60-80% of deployment costs. Underground construction is roughly twice as expensive per foot as aerial, and local permitting and make-ready rules – key policy focus areas going into 2026 – remain critical friction points. While cost inflation looks to be moderating, the capital cycle has clearly become more discerning, particularly for overbuilders and private equity buyers.
As for technology, LEO constellations are redefining rural and mobile connectivity. Starlink now serves more than 6 million subscribers via a constellation of nearly 9,000 satellites, while Amazon Project Kuiper is beginning to scale and is testing terminals capable of roughly 1,000/400 Mbps speeds – a development terrestrial operators are watching closely given the success of fixed wireless.
Against this backdrop, we spent much of this year helping clients reconsider what they are actually building. Our Technology Neutral Framework for BEAD Restructuring Policy Notice gave state broadband offices a way to compare fiber, FWA, and LEO projects on “day one” capacity, upgrade paths, and scalability, rather than on technology labels. The same principle applies to operators and investors: value accrues not just to the networks that exist, but to those that are scalable and efficient.
Customers: Trust, Guarantees, and Invisible Journeys
If broadband supply in 2025 is defined by abundance, demand is about expectations. Customers expect real choice, and they are surprised – and vocal – when that choice still comes with confusion, hidden fees, or disruption. And too often, it still does.
We described how FWA’s mainstream adoption is reshaping competitive dynamics – and it is doing the same for customer expectations. Mobile network operators have gained millions of residential FWA subscribers by offering a simple proposition: next-day shipping with self-install, “instant on” plug-and-play hardware, month-to-month terms, transparent pricing, and a digital-first journey.
Customers are now starting to ask: “If my mobile company can make this affordable, reliable, and easy, why can’t you?”
It comes at little surprise, then, that operator pricing is increasingly focused on affordability, too. In Broadband Pricing Strategy: From Promo Pricing to Price Locks, we described how US ISPs are starting to retire the classic “teaser then cliff” price model in favor of multiyear guarantees. Spectrum, Verizon, and Xfinity all rolled out versions of price locks during 2024-2025 – often without contracts – trading some ARPU upside for trust and lower churn.
The benefits of guarantees extend well beyond pricing. In A New Era of Telecom Consumer Guarantees, we looked at how providers like AT&T and Spectrum are bringing enterprise-style Service Level Agreements (SLAs) into the residential market – promising uptime thresholds, rapid technician visits, proactive outage credits, and real financial consequences for the provider if they fall short.
These moves reflect a broader shift toward differentiation beyond speed and price, with ISPs leaning into propositions that are both easier for customers to understand and harder for competitors to fake.
This shift is increasingly unfolding through digital channels – Salesforce research found nearly half of respondents prefer to make communications-related transactions online. In Optimizing Digital Customer Experiences, we argued that a positive digital experience is now as critical to growth and retention as network performance – with better personalization and journey design reducing churn by 10-15% in some segments.
AI-driven search has also reshaped how customers begin digital journeys. In our work on AI-driven search and SEO in the AI-era, we showed how Google’s AI Overviews and LLM assistants are compressing the discovery funnel. Telecom related AI Overview results grew by 12% between late 2024 and early 2025, outpacing most other industries. Research suggests AI originated searches could overtake traditional search traffic by 2028. Organic search strategy is no longer just about Google rankings; it’s about ensuring LLMs have something accurate and compelling to say about the brand.
Video consumers, in particular, face a uniquely frustrating complexity. In Watching Sports Is Hard, we quantified just how fragmented sports rights have become in the US – with live sports spread across more than a dozen broadcasters and platforms, often costing hundreds of dollars per season. Our piracy report, Why Are Most Streaming Platforms Vulnerable?, highlighted how this fragmentation and cost pressure coincides with a rebound in illegal streaming, now accounting for more than 80% of online video piracy globally.
Across broadband, video, and digital discovery, the common thread is clear: the real competition is for trust and attention. Providers that simplify choice – without oversimplifying reality – are gaining share and loyalty. Those that rely on cleverness over clarity risk being left behind.
AI: From Pilots to Performance
No review of 2025 would be complete without addressing AI. Virtually every client initiative we saw this year touched on it, and the tone is shifting fast.
Our research points to a significant “AI value gap” in telecoms. Research shows 80% of executives now view AI as crucial to future success, and roughly half are actively adopting or assessing generative AI. Yet only 16% of CEOs report AI initiatives scaled enterprise wide, and only a quarter of AI programs have delivered expected ROI. This research suggests that scaling AI is more an organizational transformation challenge than a technology problem.
This year, leading operators began breaking out of the pilot trap by treating AI like any other major transformation.
As we supported operators in this transition, we identified six essential pillars to making it happen:
- Clear strategy and workflows
- Explicit ownership and governance
- Modern data and technology platforms
- Realistic talent plans
- Transparent communication
- Continuous improvement
It’s not glamorous, but it is what separates organizations that quietly scale dozens of AI use cases from those that announce dozens and scale none.
2025 also showed that you don’t have to be a global incumbent to compete. The AI Advantage for Small Operators argued that the very things that slow large players – fragmented data, sprawling organizational charts, legacy processes – can be advantages for smaller providers. With fewer systems, tighter teams, and clearer accountability, they can often move from decision to deployed AI in months, not years, if they focus on people and process as much as tools. We’ve helped several operators turn that agility into real competitive edge.
Our role has remained deliberately pragmatic. We’re not building foundational models; we’re helping clients accelerate outcomes, answering questions like:
- Where can AI measurably reduce churn, opex, or network incidents in the next 12-24 months?
- How do we design guardrails so AI assisted interactions build customer trust rather than erode it?
- How should teams, functions, and infrastructure be reimagined to support AI-centric operations?
- How do we scale new use cases quickly – without pilot sprawl?
In short: 2025 was the year “AI first” without thoughtful transformation plans started to look reckless.
Bringing It Home
For Cartesian, these industry shifts shaped both our project work and our own investments. We supported clients through hundreds of engagements this year and remain deeply grateful for the trust you place in us.
We also practice what we preach – investing in our own data science and knowledge platforms, especially AI tools and methodologies that supercharge what we can deliver. Work that felt out of reach just a few years ago is increasingly becoming the norm.
Above all, 2025 reinforced a long-held belief: in communications and media, complexity isn’t going away. New networks, rising customer expectations, evolving content models, and accelerating AI capabilities will continue to intersect in unpredictable ways. Our role is to help turn that complexity into clarity – and, ultimately, advantage.
On behalf of all of us at Cartesian, thank you for your partnership, collaboration, and candor.
We look forward to working with you in 2026.
With gratitude,
The Cartesian Team
Reflecting on Our 2025 Predictions
Each December, Cartesian publishes its outlook on the trends and innovations expected to shape the Communications and media sectors in the year ahead. Before presenting our predictions for 2026, here is a review of how our 2025 forecasts performed.
Prediction 1: Consolidation will continue across the industry as organizations look to enhance their footprint, network capabilities, and other operational efficiencies. Deals under review such as Verizon and Frontier, and T-Mobile and US Cellular are likely to be approved in 2025 and new deals will be announced.
Did this happen? YES.
Consolidation advanced significantly in 2025, with several high-profile transactions clearing key regulatory thresholds and reinforcing the momentum behind footprint expansion. Verizon’s $20 billion acquisition of Frontier secured FCC approval in May, Charter announced their $34.5 billion merger deal with Cox in May, and T-Mobile closed its $4.3 billion US Cellular transaction in August. Additional large transactions remain underway, with Verizon’s Starry acquisition and AT&T’s acquisition of Lumen both expected to be finalized in 2026.
It is worth mentioning that joint ventures have emerged as an important mechanism for network expansion alongside traditional M&A. Partnerships such as Verizon-Tilman and T-Mobile-Metronet/Lumos demonstrate how operators extend broadband reach without full acquisitions. Like AT&T and BlackRock’s Gigapower model, joint venture structures backed by private equity continue to accelerate infrastructure investment. Together, these dynamics demonstrated how operators are using both M&A and strategic partnerships to scale networks, enhance capabilities, and strengthen market positioning.
Prediction 2: More operators will announce plans for horizontal separation to unlock value from infrastructure-class assets, enabling them to focus on service innovation and customer experience.
Did this happen? YES.
Operators continued advancing horizontal separation strategies in 2025, leveraging structural realignment to monetize infrastructure-class assets and sharpen their focus on service innovation and customer experience. TPG Telecom accelerated this shift through the $5.25 billion sale of its fiber and fixed network assets to Vocus, while established InfraCo models in Europe – such as Italy’s FiberCop and Spain’s MasOrange/Vodafone FibreCo – illustrate how infrastructure separation is becoming increasingly embedded across major markets.
In the US, structural moves are also broadening. Comcast formalized the board structure for VERSANT, its media-operations spin-off from the core broadband business. Open-access initiatives also gained traction, including AT&T’s continued expansion of its open-access JV with BlackRock and T-Mobile’s trial of open-access fiber partnerships in select markets. Taken together, these developments underscore a steady, industry-wide move toward horizontal separation, even if adoption remains uneven across operators.
Prediction 3: As adoption of FTTP increases, broadband churn will fall. Future churn will be driven by home moves, poor pricing strategies, and unforced errors in customer service.
Did this happen? NOT YET.
Across major US fiber builders, FTTP penetration continued to rise in early 2025, with telcos reporting modest increases in take-up within their fiber footprints. Fiber remains highly attractive where available, and US providers collectively added more than 900K+ broadband customers in the first half of the year.
However, these gains have not yet translated into lower operator-level churn. Legacy copper losses and intensifying fixed-wireless competition – which added roughly two million subscribers over the same period – continue to offset fiber improvements, masking the stabilizing effect of FTTP expansion. As pricing rationalization and longer-term rate commitments take hold, we expect fiber-driven churn benefits to become more visible in 2026.
Prediction 4: Service providers will find new AI use cases beyond operational efficiency and customer service solutions. However, progress will be slower than hoped due to data quality and integration challenges.
Did this happen? YES.
New AI use cases continued to emerge across the sector in 2025, but – as predicted – progress has moved slowly as operators work through data quality, system integration, and governance challenges. While the many AI deployments still center on operational efficiency and customer service, the landscape is gradually broadening into areas such as network optimization, fraud detection, targeted marketing, and B2B AI platform offerings.
Leading operators such as SK Telecom and Deutsche Telekom are advancing the frontier with generative AI assistants and AI-as-a-Service initiatives. These developments reflect meaningful momentum – albeit at a measured pace – as the industry works to overcome data and integration hurdles.
Prediction 5: In 2025, concern will grow over the environmental impact of training large AI models and the power consumption and emissions of hyperscale data centers more generally.
Did this happen? SOMEWHAT.
Global electricity demand from data centers rose sharply in 2025 – up an estimated 16% year over year, with demand projected to double by 2030. Much of this increase is linked to AI and cloud expansion, reflected in rising emissions among leading players such as Google, whose emissions are up 50% from 2019, and Microsoft, up 23% from its 2020 baseline. Public concern also intensified: In Ireland, community groups in South Dublin mobilized around data-center growth; Amsterdam continued restrictions on new large facilities due to local grid constraints; and in Northern Virginia – the world’s largest data center market – legal and preservation disputes slowed progress on the Prince William Digital Gateway.
These pressures have not slowed the broader momentum of the sector. Data-center capex is expected to grow more than 30% in 2025, reflecting continued demand from cloud, AI, and network-driven workloads. The industry is beginning to navigate environmental and power-supply concerns in parallel with expansion, as discussed in prediction six.
Prediction 6: The growing imbalance between electric power supply and demand will force telecom service providers into increasingly challenging decisions about how to procure and manage power. Power forecasting will shift from being a budgeting exercise into a cross-functional strategic planning activity. Service providers will follow the example of hyperscalers by forming cross functional power strategy teams.
Did this happen? SOMEWHAT.
Power management emerged as a meaningful strategic constraint for telecom operators in 2025, driven by the rise of AI workloads, rising electricity costs, and localized grid saturation across several US regions. As a result, operators have begun treating power less as a routine operating expense and more as a core capacity-planning consideration. AT&T committed $14 billion to energy-efficient network upgrades, Verizon surpassed 1 GW in renewable power-purchase agreements, Vodafone UK secured long-term solar supply, and BT advanced its climate transition plan. These actions signal growing urgency around energy resilience and cost management.
However, energy strategy within most telcos remains distributed across operational silos rather than organized as a fully integrated, cross-functional strategy. By contrast, hyperscalers have started consolidating power procurement, forecasting, and grid interconnection planning into enterprise-level capabilities, reflecting electricity constraints that are increasingly leaving facilities empty. Telcos are moving in this direction, but the shift is still in its early stages and has not yet fully materialized.
Prediction 7: Commercial deployment of Open RAN will start to pick up next year after a lackluster 2024. Deployments will tend to be small in scale, as providers work through interoperability issues ahead of larger rollouts in future years.
Did this happen? SOMEWHAT.
Open RAN gained momentum in 2025, though adoption remained measured as operators continued to navigate operational complexity – particularly the challenges of multi-vendor interoperability.
Europe showed the strongest progress: Vodafone initiated a major Open RAN rollout in Germany in October 2025 with Samsung and Wind River, launching a live site in Hannover and outlining plans to scale across thousands of sites in its European footprint. In Spain, MasOrange entered a five-year Open RAN-ready modernization agreement with Ericsson. In the UK, SONIC Labs continued operating in 2025 as the central Open RAN testing program, while projects funded under the earlier £80 million Open Networks Ecosystem competition remained active.
The US picture was more mixed. The shutdown of EchoStar/Dish’s nationwide multi-vendor Open RAN network following its $23 billion spectrum sale to AT&T represented a notable setback. At the same time, NTIA continued advancing the ecosystem by awarding more than $117 million to nine vendors developing commercially deployable Open RAN solutions.
Prediction 8: Commercial Direct to Cell (D2C) satellite services will slowly begin to emerge in 2025, subject to regulatory safeguards to prevent interference with other satellite and terrestrial services.
Did this happen? YES.
Commercial D2C satellite services began to materialize in 2025, exceeding expectations in both technological readiness and regulatory progress. The FCC granted SpaceX and T-Mobile conditional approval to deliver “Supplemental Coverage from Space,” enabling mobile-band spectrum reuse and accelerating commercial deployment. By Q3 2025, roughly 600 Starlink satellites were dedicated to D2C service, supporting T-Mobile, Rogers, KDDI, Salt, and other global carriers, with next-generation satellites designed to operate on spectrum acquired from EchoStar.
Regulators also advanced formal D2C/D2D frameworks – most notably Ofcom – while additional players made significant progress. Apple continued to scale its Globalstar-powered SOS service; Lynk Global merged with Omnispace to leverage global S-band spectrum for enhanced D2D offerings; and AST SpaceMobile prepared for a 2026 US launch backed by commitments from Verizon and AT&T.
Prediction 9: Global concern over the harms of social media will increase causing governments to weigh their options in the wake of Australia’s move to regulate social media usage for children and teens.
Did this happen? YES.
Concerns about social media’s impact on young people continued to rise in 2025, reflected in heightened public scrutiny and high-profile legal action. New York City filed a federal lawsuit against Meta, Google, Snap, and TikTok, alleging that addictive design features contribute to anxiety and sleep disruption among adolescents.
Regulatory activity accelerated globally, as well. Australia’s Social Media Minimum Age amendment to the Online Safety Act – requiring platforms to prevent users under sixteen from holding accounts – will take effect in December 2025. In the United States, states including Nebraska, Virginia, and California enacted stronger online-safety rules, introducing age verification, parental-consent requirements, and limits on addictive design features for minors.
The UK also advanced protections: Ofcom added more than forty mandatory measures to its children’s online-safety codes, including robust age-assurance systems and restrictions on harmful content recommendations. Across Europe, national initiatives, such as the Netherlands advising children under fifteen to stay off social media, were accompanied by new EU guidance and a prototype age-verification system under the Digital Services Act.
Predictions for 2026: Future TMT Trends
Looking forward to 2026, these are the key trends we expect will shape the TMT landscape.
1. Bundling as a Platform Strategy
Operators will accelerate the shift toward integrated multi-play propositions – combining mobile, fixed wireless, and OTT streaming through more flexible architectures. Cable and satellite providers will double down on enhanced streaming bundles, content integration, and upgrade incentives to retain market share.
As these converged offerings mature, operators will gain richer customer insights, enabling more personalized, lifestyle-driven bundles. Loyalty, rather than acquisition, will become the primary driver of customer value, opening pathways to new monetization opportunities well beyond core connectivity.
2. Fiber Market Shakeout Intensifies
As FTTP coverage expands, churn will be driven less by technology gaps and more by pricing pressure, inconsistent customer experience, and performance shortcomings. Europe – particularly the UK – will face an accelerated wave of fiber consolidation in 2026 as oversupply and rising capital costs push weaker players toward mergers or failure. With alt-nets increasingly being acquired or merged, operators will also face rising operational complexity in 2026 — particularly around safely migrating customers across networks, systems, and support processes, an area where seamless execution correlates strongly with higher retention and faster monetization.
Growth in the US market will be fueled less by large-scale M&A and more through partnerships, wholesale arrangements, and open-access models. Major operators will increasingly rely on third-party fiber networks to extend reach and manage capital intensity, accelerating the shift toward a more distributed, partnership-led broadband ecosystem.
3. Satellite Innovation Goes Mainstream
Satellite connectivity – including direct-to-cell, non-terrestrial networks (NTN), and satellite-enabled roaming – will increasingly be integrated into operator strategies to enhance rural coverage, improve roaming resilience, and close coverage gaps. In 2026, a major broadband provider will launch an integrated terrestrial/LEO offering, signaling a shift from satellite as a niche solution to a core component of network design.
Early traction in mobile D2D satellite services will prompt operators to test similar models, particularly for coverage assurance and differentiated premium tiers. Meanwhile, the commercial ramp-up of Amazon’s Kuiper constellation provides the first large-scale alternative to Starlink, intensifying competition across the LEO ecosystem.
4. Agentic AI Enters the Deployment Era
Agentic AI will move from pilots to initial deployment in 2026, with early predictive, self-optimizing network capabilities and increasingly autonomous operations beginning to take hold. On the customer side, personalized experiences and AI-driven digital interactions will start to scale – reshaping how consumers research and select providers and forcing operators to rethink acquisition strategies and digital engagement.
The year will unfold on a dual track: rapid automation and innovation paired with heightened focus on governance, security, identity, and fraud risks. AI-driven energy optimization will also become a more strategic priority, helping operators meet accelerating decarbonization commitments and improving data-center efficiency as workloads expand. Having learned from past technology-adoption cycles, operators will invest heavily in safeguards to ensure AI delivers real value without introducing new systemic vulnerabilities.
5. Fragmentation of the Global Cloud Market
Regulatory, security, and performance pressures will accelerate the fragmentation of the global cloud landscape in 2026, driving increased adoption of sovereign, industry-specific, and AI-optimized cloud environments. While AWS will remain a leading hyperscaler, it will no longer serve as the universal default. Instead, operators’ edge platforms and emerging AI-specialist providers will capture high-value, latency-sensitive workloads, creating a more distributed and diversified cloud ecosystem.
As workloads become more regulated, more industry-tailored, and more AI-intensive, the era of a single global cloud standard will give way to a federated model built around specialization and performance.
