The gap between providers quietly carving out a bigger slice of telecom’s multi-billion dollar revenue cake and those slowly bleeding ground isn’t technology, and it isn’t capital. It comes down to one nifty word: interoperability.

The providers pulling ahead have learned to dance with the rest of the ecosystem—threading terrestrial with satellite, 5G with legacy, their own stack with someone else’s—and deliver something end-to-end that a customer actually wants to buy as a whole. The ones still running exorbitant pipes to nowhere haven’t.

Here’s where interoperability shows its teeth across 2026’s sharpest industry movements.

• The end of the single-operator architecture
• GenAI enters telecom operations
• The stack owners are losing to orchestrators
• LEO satellites enter the enterprise stack
• 5G is unfinished business: there’s still money on the table

The end of a single-operator architecture

Key takeaway: Multi-domain, multi-provider service delivery has moved from a premium add-on to a baseline expectation. Operators who can’t automate across vendor boundaries aren’t losing on features, but on cost and speed; and they’re losing quietly.

If you’re not already offering on-demand, zero-touch provisioning across terrestrial, satellite, and cloud domains, you’re probably already losing ground to someone who is. Not dramatically, but consistently, contract by contract.

Geographical tensions, spectrum constraints, and mounting natural disasters are amplifying demand for resilient, diversified infrastructure. Redundant paths across vendors and domains aren’t nice-to-haves anymore.

The technical floor is well established

TM Forum APIs, Mplify LSO standards, and open northbound and east-west interfaces enable consistent exposure of capabilities (QoS slicing, location, device status, etc.) across operators, consumable as if from a single logical network. What’s still missing in most deployments is a mature orchestration layer that handles lifecycle management, intent translation, and closed-loop automation across multi-vendor environments without humans in the loop at every federation point. 

Without it, seamless delivery becomes impractical

Integration overhead, custom glue code, manual troubleshooting, and vendor lock-in inflate opex while slowing time-to-revenue. Operators who’ve closed that gap are compressing service delivery and protecting margin. Those still stitching domains together semi-manually or in isolation are running an expensive operation in a market that’s stopped rewarding infrastructure ownership and started rewarding those who can fulfill the fastest.

For telcos to get their fair share of that opportunity, they will have to develop a supply of fully interoperable APIs that work seamlessly across global networks. And only by collaborating at an unprecedented level—both among themselves and with a range of third parties, including vendors, system integrators, and enterprise developers—can they have a chance to do so.  

 

McKinsey | The future of telcos: Mapping the routes to renewed success

GenAI enters telecom operations

Key takeaway: GenAI isn’t a chatbot experiment sitting next to your OSS. It’s moving into the operational core, shortening integration cycles, cutting resolution times, and pointing toward a future where coordinated AI agents handle network decisions without a human in the loop.

AI in telecom is a sprawling conversation, from edge inference, GPU-as-a-service, and AI express lanes, to network-aware compute routing. These are big concepts with serious infrastructure implications. We’ve mapped those out separately. 

But for most operators right now, the more immediate entry point is generative AI: narrower in scope, faster to deploy, and already pulling weight in production environments. 

Customer experience is where it shows up first and most visibly, in things like context-aware AI agents handling service ordering through self-service portals, carrying state across sessions, translating intent into provisioning actions, and closing the loop without a human handoff. (We’re currently building this directly into our intelligent nBrace orchestration: a customer submits a request, the agent interprets it, and the workflow automatically fires across terrestrial, satellite, and cloud resources.) 

Internally, the gains are less visible but compound fast

Dev co-pilots are shortening integration cycles, generating orchestration scripts, policy rules, and multi-vendor adapter code that would otherwise eat engineering sprints. Support teams are running GenAI over their own trouble ticket history, vendor docs, and runbooks, cutting resolution times without scaling headcount. 

The less obvious use cases are worth a look too. Many of them are already in production:

• Automated root-cause analysis that stitches structured logs together with unstructured vendor notes. 
• Synthetic data generation for model training in privacy-constrained environments.
• Intent translation layers that map business objectives directly to orchestration rules without a manual intermediary.

Fragmented GenAI use cases won’t be enough

The direction of travel is toward the agentic network, i.e., environments where coordinated AI agents handle operational decisions dynamically, operating against intent rather than static human-defined workflows. That only works if the underlying orchestration, telemetry, and data architecture are mature enough to support it. GenAI makes it easier to sell, deliver, and get paid for telecom services, but only if the stack underneath is ready.

“Integrating Gen-AI into OSS/BSS isn’t just an upgrade—it’s a transformative overhaul of telecom operations. By automating workflows, personalizing customer interactions, and delivering real-time, actionable network insights, Gen-AI propels growth and improves customer satisfaction.” 

 

Ericsson | How multi-agent AI is transforming telco product configuration

The stack owners are losing to orchestrators

Key takeaway: Vertical integration made sense when scarcity was the product. Carriers that owned the full stack could set the terms. But that window has closed. The enterprise contracts worth chasing now thread through hyperscaler compute, third-party transport, regional access, satellite failover, and security overlays that no single operator controls, and the ones winning them are doing it through open APIs, standardized interfaces, and clean federation across domains they don’t own.

For decades, owning the stack was the strategy. Transport, access, provisioning, OSS/BSS, and customer delivery were kept within a single operational boundary and controlled end-to-end. It was a defensible position when the network itself was the scarce asset. It isn’t anymore.

The ecosystem has grown too large, too fast, and too tangled for any single operator to own the whole chain

A typical enterprise service today threads through hyperscaler compute, third-party transport, regional access, satellite failover, security overlays, and policy engines spread across different administrative domains. Nobody controls all of that. The operators are still trying to find out the hard way: 52% of telecom CEOs in PwC’s global survey believe their companies would no longer be economically viable in a decade if they continued on their current paths. 

Vertically integrated operators are being valued at the lower end of the spectrum…

…closer to asset-light retail companies, despite owning and operating significant fixed infrastructure like fiber networks and cell towers. The operators gaining ground have stopped trying to own every layer and started figuring out how to operate cleanly within someone else’s. Telefónica is restructuring around open, standardized interfaces designed to work across operator boundaries. In late 2024, Ericsson and eleven global tier-1 operators—AT&T, Deutsche Telekom, Orange, Verizon, Vodafone—formed Aduna: a joint venture explicitly built to drive a global API marketplace.

These aren’t partnerships of convenience

They’re a structural bet that the next margin pool lives in horizontal coordination, not vertical ownership. And again, the key is open, standard APIs and failsafe service orchestration carrying the promise of build-once, deploy-everywhere portability across networks.

“The future of the telecoms industry lies in a horizontal structure that minimises infrastructure costs and provides open and equal access for service providers to compete not just with other CSPs, but with the digital services providers (DSPs) too.”

 

Cerillion | Changing the telco paradigm from vertical integration to horizontal integration

LEO satellites enter the enterprise stack

Key takeaway: Satellite has gone from a maritime backhaul curiosity to a geopolitical priority and an active enterprise infrastructure. The challenge isn’t capacity, but lifecycle management. Without a common orchestration layer, satellite stays a bespoke integration project with bespoke economics to match.

For years, satellite was the thing telcos kept at arm’s length—mostly useful for maritime and remote backhaul. That’s over. Satellite has moved from adjacent infrastructure to an active geopolitical priority, and the window to treat it as someone else’s problem is closing fast.

The EU’s IRIS2 program: 290 satellites across LEO and MEO orbits, backed by a €10.6 billion budget, was launched because European governments concluded they could no longer afford strategic dependence on a single private operator. Russia’s invasion of Ukraine sharpened that calculation. Starlink’s role in that conflict made it impossible to treat satellite connectivity as a civilian infrastructure question. NATO militaries now treat low-latency satellite access as a defense asset, and China’s Guowang constellation, with 13,000 satellites, has begun deployment. 

For CSPs and wholesale providers, the strategic case maps cleanly onto commercial ones

LEO backhauls terrestrial cells in remote industrial sites, island networks, and cross-border corridors, provides continuity during storms or cable cuts, and serves enterprise verticals that centralized architectures can’t reach: logistics operators needing unbroken tracking across ocean freight lanes; energy companies running offshore platforms, or public sector clients requiring resilient comms that don’t route through a foreign commercial provider. 

The challenge is that satellite capacity doesn’t slot neatly into a terrestrial service chain. It’s a lifecycle management problem spanning multi-orbit provisioning, automated failover, cross-domain assurance, and billing abstraction, all while hiding complexity from the end customer. The consistent finding is that without a common orchestration layer to handle cross-domain dependencies, satellite remains a bespoke integration project. 

“LEO satellite broadband mega-constellations are emerging as a core component of global communications infrastructure… transforming satellite broadband from a complementary service into a core element of the global internet ecosystem.” 

 

Science Direct | OrbitShare: The policy imperative for satellite internet mega-constellation sharing

5G is unfinished business: there’s still money on the table

Key takeaway: The world set its sights on 6G before finishing 5G—and operators are paying for it in unrealized revenue. Dynamic slicing, programmable APIs, and edge computing are not inherently linked to 6G. They’re up for grabs today. The bottleneck isn’t the technology. It’s the lack of orchestration.

Fair warning: 6G isn’t coming in 2027. Standards bodies are still debating the architecture, while China is already running trials with technology that doesn’t formally exist. So, rather than jumping at technology that hasn’t crystallized, here’s what’s truly worth paying attention to—what the 6G conversation exposes about where 5G’s potential is still going untapped.

Most standalone 5G core deployments are still incomplete

Network slicing is partial. Orchestration is manual or missing. The cloud-native foundations that were supposed to make 5G programmable and composable never fully materialized in production. Consumer willingness to pay more for “faster mobile internet” plateaued, and operators’ investments in the network often turn no return. Not because the technology lacks potential, but because the operational layer to exploit it was never properly built. 

The untapped upside is substantial, and it doesn’t require waiting. Dynamic network slicing for latency-sensitive verticals like manufacturing, logistics, and autonomous mobility is architecturally possible right now, but operationally stranded without automated service lifecycle management. Wholesale programmable connectivity via CAMARA and TM Forum frameworks is ready. Edge compute services for industrial IoT and real-time analytics are deployable today.

The question is whether your orchestration layer can execute

Before the next capex cycle kicks in, it’s worth asking whether you’re extracting value from the infrastructure you already have. An operator still running siloed OSS/BSS, manual provisioning, and vendor-specific integrations in 2025 won’t absorb 6G complexity gracefully in the future, because 6G amplifies, not simplifies, fragmented operations. Worse, they’re leaving real revenue on the table from slicing, APIs, and edge services that are deployable today. Chasing the next network generation before closing the gap that exists today will compound the problem, with a heavier bill and a more complex stack to untangle on the other side.

“Deploying 5G is not the same as delivering business results… network capability must be matched by organizational readiness and cross-domain integration. Indeed, the reality is that network slicing is growing fast, but its rapid-growth phase has been delayed by a few years because of real-world practical considerations.”

 

5G Americas | The State of 5G: Growth, Challenges, and Opportunities in 2025

FAQs

1. What is interoperability, and why does it matter for telecom margins in 2026?

Interoperability in telecoms means the ability to deliver services seamlessly across multiple operators, vendors, and domains—terrestrial, satellite, and cloud—using standardized interfaces and APIs. In 2026, it’s become a direct margin driver: operators who can automate cross-domain service delivery reduce operational costs and win enterprise contracts that single-operator architectures can no longer compete for.

2. How is GenAI different from traditional AI in telecom operations?

Traditional AI in telecom has focused on predictive analytics and pattern detection—useful but narrow. GenAI goes further: it handles natural-language service ordering, generates orchestration code, automates root cause analysis across unstructured data sources, and increasingly acts as an autonomous agent capable of executing operational decisions without human intermediaries. The direction of travel is toward agentic networks, where AI systems manage service lifecycle end-to-end.

3. Why are vertically integrated telecom operators losing ground?

Vertical integration made sense when network infrastructure was a scarce asset. Today, enterprise services span hyperscaler compute, multi-vendor transport, satellite failover, and security overlays across multiple administrative domains—no single operator controls the whole chain. Operators who’ve restructured around open, standardized interfaces and horizontal coordination are gaining ground. Those still trying to own every layer face inflating opex and declining enterprise competitiveness.

4. What’s driving LEO satellite adoption in enterprise networks?

Three forces are converging: geopolitical pressure (IRIS2, Guowang, NATO defense priorities making satellite a strategic infrastructure question), commercial demand (logistics, energy, and public sector verticals needing resilient connectivity beyond terrestrial reach), and technical maturity (multi-orbit constellations like O3b mPOWER enabling the low-latency performance enterprise use cases require). The bottleneck now is orchestration: integrating satellite into terrestrial service chains without bespoke integration overhead.

5. What is 5G network slicing, and why isn’t it widely deployed yet?

Network slicing allows operators to carve a single 5G infrastructure into multiple virtual networks, each with guaranteed performance characteristics—low latency for manufacturing, high throughput for video, resilient connectivity for critical infrastructure. It’s architecturally possible right now. The barrier to deployment is operational: most operators lack the automated service lifecycle management needed to provision, assure, and bill slices at scale without manual intervention at every step.

6. How should operators prepare their infrastructure for 6G?

By fixing what 5G left unfinished. 6G will require autonomous, intent-driven network operations across terrestrial, satellite, and edge compute domains. Operators arriving at 6G with siloed OSS/BSS, manual provisioning, and vendor-specific integrations will face the same monetization failure as 5G, at higher capex. The preparation that matters now: deploying automated, API-driven, cross-domain orchestration; completing network slicing in production; and building the open northbound interfaces that make programmable connectivity commercially viable.