The numbers are in, and they tell a story the telecom industry has been building toward for years. Kaleido Intelligence’s latest forecasts put the number of 5G roaming consumers — across both NSA and SA deployments — at more than 340 million in 2026. Add IoT into the mix, and another 38 million 5G roaming connections come into frame. That’s not a niche experiment anymore. That’s infrastructure at scale.

But here’s the thing: most of those 340 million roamers will be riding on Non-Standalone architecture. And NSA, while functional and widely deployed, is essentially borrowed infrastructure — 5G radio layered on top of existing LTE cores. It works. It’s available in 130+ markets through 200+ operators. But it doesn’t unlock the full promise of 5G, and the industry knows it.

The real pivot — technically and commercially — is Standalone.

Why SA Changes the Equation

5G SA brings the cloud-native core into the picture. That’s what enables network slicing, service-level guarantees, and the kind of security architecture that enterprise clients actually care about. When operators talk about bringing the domestic 5G experience to roaming, they mean SA. NSA was always a transitional layer.

Kaleido currently tracks around 90 full commercial 5G SA launches across 45+ countries, with 190+ MNOs actively investing in SA domestic buildouts. Trials for SA roaming have been completed in 15+ markets. And in Kaleido’s Q4 2025 survey of roughly 70 tier-1 and tier-2 operators globally, at least 25 indicated plans to launch commercial 5G SA roaming services this year.

That’s a meaningful cluster of early movers — not the whole market, but enough to establish the commercial pattern that others will follow.

The Technical Lift Is Real

Don’t mistake momentum for ease. More than 80% of survey respondents said that testing new roaming network functions is either an immediate or near-term requirement. We’re talking SEPP (Security Edge Protection Proxy), HTTP/2 signalling, policy control, charging interfaces, and core security functions — none of which are trivial to coordinate across bilateral roaming partners.

The SEPP situation is particularly telling. Deployment strategies are still evenly split across models, with no clear dominant approach emerging. Many operators are in wait-and-see mode, running trials and building understanding before committing to architecture decisions. That’s rational behaviour given the investment involved, but it does mean that full ecosystem alignment is still some way off.

BCE — Billing and Charging Evolution, the new framework for flexible, real-time roaming monetisation — is another sticking point. Roaming partner readiness for BCE was flagged as the single most widespread challenge in SA deployments. It’s a classic chicken-and-egg problem: operators can’t fully commercialise SA roaming until their partners support the billing infrastructure, and partners won’t prioritise that investment until there’s commercial demand.

The Business Case Isn’t Universal

Here’s where the market gets complicated. Over a third of operators not planning SA roaming launches say there simply isn’t a sufficient business case for roaming-specific standalone investment. That’s a significant share of the market essentially sitting out the first wave.

It’s not irrational. For smaller operators with limited cross-border traffic, the ROI on SA roaming infrastructure is genuinely uncertain. The enterprise use cases — private network extensions, network slicing for mission-critical applications, SLA-backed connectivity for connected logistics — are compelling, but they require a customer base that not every operator has. The consumer case for SA roaming is murkier. Most travellers can’t tell the difference between NSA and SA in real-world conditions, which makes it hard to price the upgrade or build differentiated retail products around it.

More than 60% of respondents in Kaleido’s survey called out real-time analytics as an immediate requirement — alongside QoS assurance, interconnect security, and IPX capacity — which signals that operators know the service quality bar needs to rise, even if commercial frameworks haven’t fully caught up.

What This Means Beyond the MNO Layer

For the eSIM and travel connectivity ecosystem, the SA transition matters in ways that aren’t always obvious. eSIM providers — whether retail-focused operators like Airalo, Holafly, and Nomad, or wholesale/B2B infrastructure players — ultimately depend on the underlying roaming agreements and network capabilities that MNOs expose. Better QoS guarantees and network slicing at the SA layer could, in theory, enable differentiated data product tiers for eSIM users: guaranteed-speed packages, enterprise-grade connectivity, or slice-based offerings tied to specific use cases.

That’s not available today at scale. But the architecture being put in place now is what makes it possible in three to five years. The operators building SA roaming frameworks in 2026 are laying the groundwork for the commercial models that eSIM and MVNO players will eventually build products on top of.

Where This Is Heading

Ericsson’s own mobility data consistently points to 5G becoming the dominant global access technology before 2030. GSMA Intelligence tracks similar trajectories. What Kaleido‘s data adds is operational granularity — specifically, that the roaming layer is lagging domestic rollout by roughly two to three years, which is historically normal but worth tracking given enterprise demand timelines.

The operators most likely to move fast on SA roaming are the ones with heavy enterprise exposure: Deutsche Telekom, SoftBank, SK Telecom, and a handful of Gulf-region carriers where premium connectivity commands real pricing power. The operators most likely to delay are mid-tier carriers in markets where roaming volumes are lower and bilateral partner ecosystems are thinner.

2026 won’t be the year 5G SA roaming goes mass market. But it will be the year the commercial and technical frameworks either prove out or stall — and that verdict will shape the roaming industry’s trajectory for the rest of the decade. The foundation is being poured. What gets built on top of it depends entirely on whether the business case catches up with the engineering.