When GSMA launched Open Gateway back in 2023, the pitch was simple enough to sound utopian: turn mobile networks into a programmable platform that developers could tap via standardized APIs, regardless of which country or carrier they were dealing with. No more bilateral operator negotiations, no more bespoke integrations, no more “write once, break everywhere.”

Three years later, the progress is genuinely hard to dismiss — even if the gap between what’s been announced and what’s been commercially deployed still deserves scrutiny.

The Scale Argument Is No Longer a Stretch

The initiative now counts 86 operator groups covering more than 300 networks and roughly 80% of global mobile connections. Alongside them, 60+ channel partners — including hyperscalers, CPaaS providers, and aggregators — are actively commercializing API access. What started with eight CAMARA APIs at launch has grown into more than 300 instances of 20 different CAMARA APIs commercially live across 65 markets, from Canada to Chile, from the US to New Zealand.

That’s not vaporware. For an industry that has historically struggled to coordinate across competing commercial interests, this kind of alignment is meaningful — even if implementation depth varies wildly market by market.

Where It’s Actually Doing Something

The use cases with real traction fall into a few categories, and fraud prevention is the clear frontrunner. Multi-operator deployments across dozens of countries are giving banks and retailers the ability to verify identity, detect SIM swap fraud, and protect transactions in real time. UK banks are already flagging SIM-swap scams at scale. Colombia, Poland, Italy, and New Zealand all announced new anti-fraud collaborations at MWC26 in Barcelona.

SIM swap detection matters for the travel tech ecosystem, too. It’s a direct attack vector for account takeovers — the kind of fraud that eSIM platforms and roaming providers are increasingly exposed to as they scale digital-first distribution. Any operator-level tooling that detects SIM swap events in real time has upstream value for identity verification across the broader connectivity stack.

Quality on Demand (QoD) is the other API that’s gaining real enterprise traction — enabling applications to request elevated network performance for payments, streaming, drone operations, and other latency-sensitive workloads. Tata Elxsi has called for standardized QoD APIs specifically for the automotive and drone industries. Skydio, the drone manufacturer, has similarly pushed MNOs to extend support for emergency-services use cases.

The Channel Partner Layer Is the Real Leverage

What often gets missed in the GSMA press cycle is the distribution story. The operators themselves aren’t the primary interface for most developers — it’s the channel partners who aggregate access and handle the commercial complexity. Orange and Shabodi are jointly working on manufacturing applications. China Mobile reportedly doubled in-app advertising conversion rates using CAMARA number verification. These are early proof points, but they reveal something important: the value isn’t just in the API itself, it’s in how it gets packaged and delivered downstream.

GSMA Fusion is attempting to solve a different version of the same problem — acting as a demand-side bridge so that enterprise sectors like fintech, automotive, and aviation can collectively define what they need from networks, rather than adapting to whatever a single carrier decides to expose. It’s a meaningful structural shift from supply-push to demand-pull.

What Agentic AI Changes

The most forward-looking announcements at MWC26 weren’t about new APIs — they were about the intelligence layer on top. Telefónica and Nokia are piloting Agent-to-Agent (A2A) protocols and the Model Context Protocol (MCP) to orchestrate tasks across AI agents, enabling automatic API discovery, intelligent capability selection, and goal-driven workflows without manual intervention.

In plain terms: instead of a developer manually chaining multiple API calls to build a fraud-detection flow, an AI agent does that autonomously — querying SIM-swap verification from multiple operators simultaneously, adjusting QoD parameters based on real-time network conditions, escalating to a human only when needed. Nokia and AWS, alongside du and Orange, have already demoed agentic network slicing. Mplify, Colt, Orange, and Google Cloud jointly demonstrated agentic connected experiences at the event.

This isn’t theoretical. The MCP standard — originally developed in the AI/developer tooling space — is now being used to make telecom APIs discoverable by AI agents. That is a significant architectural shift.

What This Means for the Connectivity Ecosystem

The honest assessment: Open Gateway has cleared the coordination problem that has stalled programmable network initiatives for 15+ years. The CAMARA standardization effort is the most credible attempt to date at making operator APIs interoperable at scale, and the channel partner ecosystem is building real commercial momentum around it.

For context, compare this to previous industry attempts: GSMA OneAPI (2010s), GSMA Wholesale Applications Community, various national API gateway projects — all fragmented, all stalled on commercial or political alignment failures. Open Gateway has outlasted most of those by building at the standards layer first and commercialization second.

The eSIM and roaming sector should watch this closely. As platforms like Airhub, Ubigi, and Yesim build API-native distribution architectures, the availability of standardized MNO-side APIs for identity, QoD, and fraud detection becomes infrastructure — not a nice-to-have. Providers running embedded connectivity at scale will eventually want to hook into network-layer signals for real-time fraud detection, session quality management, and identity assurance. That’s exactly what Open Gateway is building.

The risk is the usual telecom risk: momentum in press releases doesn’t always translate to developer adoption at scale. According to GSMA’s own Open Gateway State of the Market report, fraud prevention and mobile payments remain the top developer priorities, which at least suggests the use case roadmap is demand-aligned rather than operator-wishful. Whether the agentic AI layer accelerates adoption or adds complexity will depend heavily on how accessible those MCP integrations become for mid-market developers, not just Nokia and Telefónica.

The foundation is more solid than it looked three years ago. The next 18 months will determine whether the intelligence layer turns it into something developers actually build on.