- Huawei is attempting to define the control plane for autonomous networks
- The A2A-T protocol addresses the coordination problem at the heart of multi-agent telecom operations
- If carriers and standards bodies adopt it, competitors may have little choice but to follow
The most significant announcement at TM Forum's DTW Ignite in Copenhagen this week wasn't another AI agent, another autonomous network feature or another demonstration of what generative AI might someday do for telecommunications. It was a protocol that defines how thousands of future AI agents might communicate with one another inside autonomous networks. Judging by much of the media coverage emerging from the event, the significance of the announcement appears to have been largely overlooked.
Coordination, not capability
While much of the industry's attention was focused on individual AI and autonomous capabilities, Huawei used one of the most prominent stages at the event to discuss something considerably more fundamental. During a keynote delivered by Yang Chaobin, Huawei's EVP and CEO of the company's ICT Business Group, the company outlined its vision for A2A-T, or Agent-to-Agent for Telecom.
In explaining the concept, Yang argued that "a unified agent communication standard serves as the universal language to enable efficient interconnection and large-scale multi-agent collaboration across layers and systems."
That statement may ultimately prove more important than many of the AI announcements that dominated the event.
The delivery mechanism matters. Companies do not devote one of the most important keynote slots at DTW to a minor engineering feature. Keynotes are used to signal strategic intent, and Huawei's intent appears increasingly clear. The company is not simply trying to build AI agents. It is attempting to help define the architecture through which future AI agents communicate, coordinate and collaborate.
Why A2A-T could function as a control plane
At the center of that effort sits A2A-T, an emerging framework developed through the TM Forum Autonomous Networks community and championed by Huawei, China Mobile and other industry participants. At first glance, A2A-T sounds like just another telecom acronym destined to disappear into the industry's ever-expanding alphabet soup. In reality, it addresses what may become the defining challenge of autonomous networking: how thousands, and eventually perhaps millions, of autonomous systems communicate, coordinate and collaborate with one another.
The significance of Huawei's announcement is that the company did not simply discuss a protocol. It outlined the foundations of what could become a control plane for multi-agent telecom operations.
That distinction is important because history suggests that the technologies exerting the greatest influence over industries are often not the applications themselves but the protocols and architectures that govern how ecosystems operate. TCP/IP enabled the Internet. HTTP enabled the Web. Kubernetes became the control plane of cloud computing. If autonomous networking becomes reality, competitive advantage may accrue not merely to those building the most capable agents, but to those helping define how those agents communicate.
The challenge A2A-T seeks to address is substantial. Future autonomous networks are unlikely to be managed top-down by a single AI system. Instead, they will be managed by thousands of specialized agents responsible for radio networks, transport infrastructure, cloud resources, customer experience, security, service assurance, energy management and business operations. Those agents will come from multiple vendors, run on different platforms, use different data models and be built using different AI architectures. Some may come from Huawei, others from Ericsson, Nokia, Cisco, Ciena, hyperscalers, software vendors or the carriers themselves.
This is where what I have previously described as Agentic Chaos Theory becomes relevant. As the number of autonomous systems increases, complexity increases with it. Each agent may be behaving rationally according to its own objectives, policies and data, yet the interactions between thousands of such agents can create feedback loops, conflicts, unintended consequences and forms of emergent behavior that become increasingly difficult to predict.
An old networking problem, at far greater scale
A2A-T attempts to head off that problem by providing a framework through which agents can exchange intent, goals, context, state information, requests, recommendations and decisions while also providing mechanisms for negotiation, reconciliation and conflict resolution when competing objectives inevitably arise.
In some respects, it resembles a highly evolved descendant of the original CSMA/CD protocol that governed early Ethernet networks. Ethernet's challenge was determining which node could transmit data without creating collisions. Autonomous networks face a similar problem, albeit at vastly greater scale: determining how thousands or millions of intelligent agents can make decisions without creating operational collisions. The underlying challenge remains coordination. Only the complexity has exploded.
This is where Huawei's strategy becomes particularly interesting. Rather than attempting to drive adoption through proprietary technology, the company appears to be following a familiar Huawei playbook built around standardization, ecosystem development and carrier support. By working through TM Forum and encouraging broad operator participation, Huawei is attempting to move A2A-T beyond a vendor initiative and toward an industry framework.
Huawei's standardization playbook
That approach matters because adoption creates gravity. During conversations at DTW, China Mobile demonstrated A2A-T operating within a live pilot environment, suggesting that the concept is already moving beyond theory. If operators begin implementing A2A-T at scale, and if TM Forum ultimately incorporates similar concepts into the broader autonomous networks architecture, competing vendors may find themselves adapting to the framework regardless of whether they originally supported it. Not because Huawei demands it, but because interoperability increasingly requires it.
The repeated emphasis on open collaboration, multi-agent communication, open-source interfaces and cross-domain orchestration is therefore no accident. Huawei understands something fundamental about telecommunications. No major operator runs a single-vendor network, and autonomous networking only works if it functions across heterogeneous environments involving multiple technologies, domains and suppliers.
Whether A2A-T ultimately becomes the dominant framework for agent communication remains an open question. What already seems clear, however, is that much of the industry spent DTW talking about intelligence while Huawei was talking about coordination.
The really interesting question is not whether Huawei announced a protocol.
The really interesting question is whether Huawei just announced the beginnings of a control plane for multi-agent telecom operations.
Stephen M. Saunders MBE is a communications analyst and USPTO-registered inventor examining how digital infrastructure — 5G, cloud, and AI — is reshaping industry, power and society, as well as underpinning the emerging, ubiquitous global digital economy. As anchor of FNTV and a longtime industry insider, he focuses less on growth narratives and more on execution, risk and how hyperscale technology is distorting markets, governance and society at scale.
Opinion pieces from industry experts, analysts or our editorial staff do not represent the opinions of Fierce Network.