Telecom industry trends 2026 are easier to describe in deployment terms than in marketing terms. After three years of slideware about 5G Standalone, AI-RAN, network APIs, and satellite direct-to-device, operators in 2026 are running production traffic on most of them — selectively, and with the kind of caveats that do not fit a press release. This page is a field-view summary of what operators, vendors, and enterprise buyers are actually doing across the year, what shifted from 2025, and what to expect through the remainder of the decade. It is written for procurement teams, network architects, and policy readers rather than for investors.
Key takeaways
- 5G Standalone is the default for new builds; legacy traffic is migrating gradually rather than cutting over.
- AI-RAN in production is mostly energy-saving and link adaptation — not yet throughput uplift at scale.
- Network APIs through GSMA Open Gateway and CAMARA have revenue traction in fraud-prevention use cases first.
- eSIM-only handsets are normal in the US and expanding in Europe; the long tail runs through 2027–2028.
- Satellite direct-to-device is a coverage product, not a primary revenue line, but every tier-one is partnered.
- Open RAN remains a single-digit share of global RAN spend; the incumbent trio still wins most RFPs.
5G Standalone moves from pilot to default
The defining shift of 2026 is that 5G Standalone is no longer the demo — it is the default mode for new sites, for enterprise slices, and for the SA-capable handset base. Per GSMA Intelligence tracking, more than 130 operators globally have launched SA commercially. The technical foundation is 3GPP TS 23.501 and TS 23.502 for the service-based architecture, with TS 38.300 covering the NR radio interface. What changes in production is that operators can finally separate control-plane and user-plane functions cleanly, deploy UPFs at the edge, and offer differentiated QoS per application rather than per device class.
The slower side of the story is the Non-Standalone tail. Most operators are not turning off NSA. They are running both modes side-by-side, steering SA-capable devices onto the SA core for slicing and VoNR while leaving NSA in place for the long tail of legacy phones. That keeps RAN configurations complicated and keeps the savings from a clean SA-only RAN out of reach for another two to three years.
AI-RAN: real, narrow, mostly about energy
AI-RAN became the loudest talking point of 2025 thanks to the NVIDIA-led AI-RAN Alliance and a wave of vendor announcements. In 2026 the production reality is narrower than the marketing. The deployed features are mostly link-adaptation, beam-management, and energy-saving inference inside the baseband — useful, measurable, but incremental. The throughput-uplift demos in vendor labs depend on workloads that real cells rarely see continuously.
The more strategic AI-RAN question is whether the GPU-based baseband stack — NVIDIA Aerial and the surrounding ecosystem — eventually displaces the custom ASIC designs from Ericsson, Nokia, and Samsung. Softbank’s deployment work and T-Mobile’s AI-RAN Innovation Center are the most-watched proof points. The economic case rests on whether the same compute can be reused for inference workloads when cell load is low, turning the RAN tower into a distributed edge GPU pool. That is a 2027–2028 conversation, not a 2026 deployment.
Network APIs: small revenue, real architecture
The GSMA Open Gateway initiative and the Linux Foundation CAMARA project have produced the first cross-operator API standardisation that actually shipped. The first wave — Number Verification, SIM Swap, Device Location, Quality on Demand, Device Status — is live with more than 65 operators participating per GSMA disclosures. Aggregators including Vonage, Infobip, Sinch, and the hyperscaler API marketplaces resell them to enterprise developers.
Revenue is concentrated in fraud-prevention use cases: silent authentication against SIM Swap, Number Verification replacing SMS one-time-passwords, Device Location for risk scoring. Quality on Demand is technically interesting but commercially slower — enterprises have to actually want differentiated QoS, and most application teams have built around best-effort. The architecture is in place; the demand curve is still bending up.
eSIM-only and the long tail of physical SIM
eSIM-only is now the default for Apple in the US, and Apple expanded eSIM-only SKUs to additional European markets through 2026. Samsung, Google, and Xiaomi all ship eSIM-only variants in selected regions. The provisioning stack — GSMA SGP.22 for consumer and SGP.32 for IoT — is mature, and operator SM-DP+ deployments are commodity. What is slower is the prepaid and MVNO long tail in Europe, Latin America, Africa, and parts of Asia, where physical SIM logistics still anchor the customer-acquisition flow. Full global eSIM-only is a 2027–2028 reality, not 2026.
Satellite direct-to-device becomes a feature, not a category
By the end of 2026 every tier-one mobile operator in a developed market has a direct-to-device satellite partnership. The two volume providers are SpaceX’s Starlink Direct-to-Cell and AST SpaceMobile, with regional players including Lynk Global and Iridium-based solutions filling niches. The product on offer in 2026 is SMS-everywhere and basic data in coverage gaps. 3GPP Release 17 NTN brought NB-IoT and NR profiles into the standard; Release 18 widens the data envelope and tightens the timing-advance handling.
Operators are not pricing direct-to-device as a primary revenue line. It is a coverage feature, an emergency-services product, and a churn-reduction tool. The wholesale economics — per-message, per-MB, with revenue share to the satellite operator — favour bundling into existing plans rather than premium standalone tiers.
Vendor landscape and Open RAN reality check
The RAN vendor landscape in 2026 still consolidates around Ericsson, Nokia, Samsung, Huawei, and ZTE, with the last two restricted in markets aligned with US and UK trade guidance. Open RAN deployments at Rakuten Mobile in Japan, Dish Wireless in the US, 1&1 in Germany, and Vodafone in the UK and Italy are real, but per Dell’Oro and Omdia tracking they remain a single-digit share of global RAN spend through 2026. The Open RAN Policy Coalition continues to push policy support; the procurement reality is that the incumbent vendors still win most large RFPs on integration risk and roadmap confidence.
- RAN: incumbent trio dominant; Open RAN gaining single-digit share, concentrated in greenfield and brownfield-modernisation projects.
- Core: cloud-native shift continues; Mavenir, Casa, Affirmed-style stacks competing with incumbent core suites; hyperscaler partnerships standard.
- Transport: 400G to the metro common; 800G in pilot at major backbone operators; coherent pluggables drive cost down.
- BSS / OSS: cloud-native rewrites finally arriving in volume; Amdocs and Netcracker dominant on the suite side.
How DROAM News reads it
The headline trends of 2026 are easy to overstate. 5G Standalone is real but partial. AI-RAN is real but narrow. Network APIs are real but the revenue is concentrated. Satellite direct-to-device is real but it is a feature, not a category. The operators that have moved fastest — T-Mobile US, Reliance Jio, China Mobile, Softbank, KDDI, Etisalat — share two traits: clean spectrum positions and the willingness to commit to a single architectural bet rather than hedge across NSA and SA indefinitely. That is the planning lesson worth carrying into 2027 procurement cycles. Editorial disclosure: Droam BV, the publisher of DROAM News, operates roaming and IoT connectivity services in the B2B market — that overlaps with this coverage and is handled per our editorial policy.
Related DROAM News pages
- Telecom industry news — operator strategy, vendor moves, and policy coverage.
- Mobile & Wireless news — 5G handsets, eSIM, and consumer wireless trends.
- Editorial policy — how we select stories, disclose commercial overlap, and handle corrections.
Sources and references
Figures and deployment status above are drawn from public operator disclosures, standards bodies, and industry analysts. Specific operator metrics shift quarter to quarter — verify against the primary source before procurement decisions.
- GSMA and GSMA Intelligence operator tracking and Open Gateway disclosures: gsma.com.
- 3GPP specifications TS 23.501, TS 23.502, TS 38.300 and Release 17 / 18 work plans: 3gpp.org.
- Linux Foundation CAMARA project: camaraproject.org.
- Open RAN Policy Coalition position papers: openranpolicy.org.
- ITU-R IMT-2020 framework and ITU NTN-related work: itu.int.
- ETSI MEC and NFV specification series: etsi.org.
FAQ
Is 5G Standalone finally the default in 2026?
Standalone is the default for new builds and for greenfield enterprise slices, but it is not yet the default for legacy mass-market traffic. Most tier-one operators in North America, China, the Gulf, and the Nordics have SA carrying a material share of consumer traffic. Much of Europe and Latin America still leans on Non-Standalone in the radio access network with SA enabled selectively for slicing-capable handsets and enterprise APNs. The cutover is incremental rather than a flag day.
What is AI-RAN actually doing on commercial networks?
In 2026 the production use cases are narrower than the marketing suggests. Vendors have shipped link-adaptation, beam-management, and energy-saving features that use ML inference inside the baseband. NVIDIA Aerial-based stacks and Ericsson, Nokia, and Samsung implementations are running in field trials with Softbank, T-Mobile, KDDI, and a handful of European operators. Outside of those, AI-RAN is mostly an energy-efficiency story today — cell shutdown and carrier sleep — rather than a throughput story.
Are network APIs (CAMARA) generating revenue yet?
Revenue is still small but the architecture is in place. The GSMA Open Gateway initiative and the Linux Foundation CAMARA project standardised a first wave of APIs — Number Verification, SIM Swap, Device Location, Quality on Demand. Aggregators like Vonage, Infobip, and Sinch are reselling them to enterprise developers. Most of the volume is in fraud prevention and silent authentication; quality-on-demand monetisation depends on enterprises actually buying differentiated QoS, which is moving slowly.
Is eSIM-only the new normal for handsets?
In the US it has been the norm since iPhone 14. In 2026 Apple ships eSIM-only in additional markets, and Samsung, Google, and Xiaomi have eSIM-only SKUs in selected regions. Operators in Europe and Asia have largely caught up on consumer eSIM provisioning, but prepaid markets and many MVNOs still depend on physical SIM logistics. Full eSIM-only globally is a 2027–2028 trajectory, not a 2026 reality.
What is happening with satellite direct-to-device?
Commercial direct-to-device service is live with AST SpaceMobile and Starlink Direct-to-Cell, partnered with operators including AT&T, Verizon, T-Mobile US, Vodafone, Rakuten, Optus, and KDDI. The 2026 reality is SMS and basic data in coverage gaps, not a replacement for terrestrial mobile. 3GPP Release 17 NTN support in handsets is widening, and Release 18 adds higher-throughput NR-NTN profiles. Operators view it as a coverage extension and an emergency-services product, not a primary revenue line.
Is the vendor landscape really consolidating around three suppliers?
In radio access, yes — Ericsson, Nokia, and Samsung hold the majority of non-Chinese spend, with Huawei and ZTE dominant in markets that allow them. Open RAN has not yet displaced the incumbents at the scale the policy push implied: deployments at Rakuten, Dish, 1&1, and Vodafone are real but a single-digit share of global spend. In core, the cloud-native shift opens the door to Mavenir, Casa, and hyperscaler-aligned stacks, but again the incumbents still win most RFPs.