- AI at MWC 2026 shifted decisively from experimentation to network‑native deployment
- AT&T emerged as a central force, backing open, telco‑grade AI models and AI‑ready network architectures
- Operators increasingly framed AI as core infrastructure, not an overlay or bolt‑on feature
MOBILE WORLD CONGRESS 2026, BARCELONA — AI was everywhere at Mobile World Congress 2026 in Barcelona — but the tone was markedly different from past years. Instead of conceptual demos or future promises, operators focused on how AI is being operationalized inside live networks, from core platforms to RAN architectures.
Across briefings and announcements, a consistent message emerged: AI is becoming a built‑in capability of telecom networks, designed to improve efficiency, resilience and automation rather than serve as a standalone product.
Orange trusts the future
European operators were especially vocal about this shift. Orange for example, framed connectivity itself as strategic infrastructure under its new “Trust the Future” strategy, emphasizing AI‑driven automation across fixed, mobile, submarine and satellite networks.
Orange’s MWC demonstrations highlighted AI embedded directly into network operations, including real‑time sensing, hybrid satellite‑terrestrial orchestration and AI assistants inside the core network. The giant European telco announced partnerships with AST Space Mobile and Satellite Connect Europe for direct-to-device connectivity and Tech Mahindra for enterprise end-to-end digital transformation. The operator positioned these capabilities as foundational to delivering predictable performance and trust at scale.
That platform‑centric framing echoed throughout the show, with AI increasingly tied to how networks are built and run, not just how they are marketed.
AT&T and telco‑grade AI
While European operators emphasized platform strategy, AT&T played a pivotal role in anchoring AI‑driven networking in open standards and production‑ready models.
On Day 1 of MWC, AT&T joined the GSMA in launching Open Telco AI, an industry initiative designed to accelerate the development of AI models purpose‑built for telecom workloads. AT&T is a founding contributor, providing a new family of open telco AI models alongside compute contributions from AMD and TensorWave.
The GSMA framed the effort as a response to a persistent gap: while general‑purpose AI models have advanced rapidly, they continue to struggle with telecom‑specific tasks such as interpreting network data and automating operations. Open Telco AI aims to close that gap through shared datasets, benchmarks and collaborative development.
From models to network architecture
AT&T reinforced that standards‑driven AI message with a series of Day 2 MWC announcements focused on network architecture.
In updates, AT&T highlighted progress across:
- Open RAN readiness, positioning disaggregated RAN as a prerequisite for automation
- Cloud RAN collaboration with Ericsson, tying virtualization directly to AI‑driven optimization
- Last‑mile connectivity initiatives with AWS, aimed at supporting AI‑intensive enterprise workloads
Taken together, the announcements underscored AT&T’s view that AI is shaping how networks are designed — from the edge to the core — rather than simply running on top of existing infrastructure.
Open, autonomous and efficient
A recurring theme across operator messaging was efficiency, particularly as AI workloads drive up energy consumption and operational complexity. AT&T and its peers repeatedly linked AI adoption to cloud‑native design, automation and software‑defined control as ways to manage cost and scale responsibly.
The emphasis on openness was equally notable. By backing Open Telco AI and open RAN architectures, AT&T positioned itself alongside operators arguing that no single vendor or proprietary model can solve AI for telecom alone.
DT moves AI into the fabric of comms services
Deutsche Telekom focused its news on how AI is moving directly into the fabric of communications services, unveiling a network‑integrated AI call assistant. Rather than relying on apps or specialized devices, the assistant is embedded in Deutsche Telekom’s network and can be activated by voice during a standard phone call, enabling features such as live translation, call summaries and real‑time assistance.
The bit of news underscored a broader operator push to make AI a native network capability, reducing friction for users while keeping intelligence centralized in the network layer rather than at the device edge.
The AI narrative at MWC also extended into early 6G planning, with DT and T‑Mobile US highlighting a new transatlantic collaboration focused on co‑developing 6G technologies that combine connectivity, sensing and intelligent compute to support emerging physical AI use cases, from autonomous systems to real‑time environmental awareness.
Intel makes the AI in telecom case
Intel used MWC to make the case that AI in telecom is quickly becoming a core‑network function rather than an edge or cloud‑only workload. In briefings at the show with Fierce Network's Dan Jones, the company argued that most operator AI use cases — such as anomaly detection, network performance optimization and real‑time analytics — can be handled by small and mid‑sized models running directly on existing 5G core infrastructure, without requiring GPUs or specialized hardware. Intel said carriers are prioritizing performance per watt and power efficiency over raw compute, with AI increasingly deployed to optimize live networks while keeping energy consumption in check.
The message aligned with a broader MWC theme: AI’s value in telecom will be determined less by model size and more by how effectively it can be integrated into production networks at scale.
In the 6G news realm, Intel announced a partnership with Ericsson aimed at accelerating ecosystem readiness for seamless transition to AI-native 6G deployments and use cases.
The bigger picture
By the end of MWC 2026, one conclusion was hard to avoid: AI‑driven networks are no longer a future concept. Operators are actively embedding intelligence into cores, radios and management systems, reshaping telecom infrastructure around automation and adaptability.
With initiatives like Open Telco AI and concrete architectural moves from companies like AT&T and Orange, the industry appears to be coalescing around a shared goal—making AI a native, trusted and scalable part of global communications networks.
Read all of our coverage from Mobile World Congress 2026 in Barcelona here.