The rapid proliferation of Artificial Intelligence has created an unprecedented performance challenge for the global networking infrastructure that was largely engineered before the AI revolution, leaving service providers at a critical crossroads. These providers face the complex task of capitalizing on their vast, pre-existing Ethernet investments while simultaneously adapting to the stringent, low-latency demands of modern AI workloads, a dilemma that pits legacy infrastructure against next-generation requirements. In response to this industry-wide challenge, Mplify, the alliance formerly known as the Metro Ethernet Forum (MEF), has unveiled a comprehensive overhaul of its certification program. This strategic relaunch introduces a dual certification framework intended not only to validate existing Ethernet services for today’s business needs but also to forge a clear, standardized pathway for certifying networks as “AI-ready,” bridging the gap between current capabilities and future demands as part of a broader strategic pivot toward Network as a Service (NaaS).
A Dual Certification Strategy for the AI Era
Evolving the Standard from MEF 3.0 to Carrier Ethernet for Business
The foundation of Mplify’s updated strategy is the “Carrier Ethernet for Business” certification, a program that serves as a direct rebrand of the widely adopted and respected MEF 3.0 standard. By maintaining technical requirements identical to its predecessor, this certification ensures complete continuity for the industry, allowing for a seamless transition that honors the decades of investment carriers have poured into standards-based Ethernet. This approach is critical, as it provides a stable and familiar baseline upon which the new AI-focused initiatives can be built. Service providers who currently hold MEF 3.0 certification are offered a straightforward migration path to the new Mplify certification at no additional cost. This immediate availability and easy transition are designed to preserve the value of existing certifications while aligning them with Mplify’s forward-looking vision, ensuring that the established ecosystem of certified services remains relevant and robust in a rapidly evolving technological landscape.
This transition from MEF 3.0 offers a significant practical advantage that extends far beyond a simple name change, empowering operators to refresh their market posture and demonstrate the ongoing competitiveness of their certified infrastructure. With some existing MEF 3.0 certifications being up to five years old, the opportunity to recertify services allows providers to obtain updated test records, presenting a fresh and compelling value proposition to customers without the need for new, costly infrastructure upgrades. Kevin Vachon, Chief Operating Officer of Mplify, emphasized that the goal was not to merely “put a different sticker on it” but to provide a tangible mechanism for operators to validate their network’s continued excellence. The certification testing process itself is modernized, conducted on live production networks using an automated platform that can be completed in just a few days following technical preparations. This efficient process, coupled with a financial model involving a one-time certification fee and predictable annual maintenance fees, makes it accessible for providers to keep their certifications active and competitive.
The New Frontier Carrier Ethernet for AI
The more groundbreaking component of the new strategy is the “Carrier Ethernet for AI” certification, which is not a standalone program but an advanced performance layer built upon the “Business” certification baseline. To qualify for this elite designation, a service provider must first achieve the “Carrier Ethernet for Business” certification and then prove its network can meet a set of additional, more stringent performance criteria specifically tailored for the unique demands of AI workloads. Mplify’s technical leadership, in collaboration with its membership, meticulously identified the critical performance parameters that differentiate AI transport needs from standard business traffic. Consequently, the validation for this certification focuses intensely on three key performance metrics: frame delay, inter-frame delay variation (jitter), and frame loss ratio. The testing protocol will utilize MEF 91 test requirements but will employ AI-specific traffic profiles and performance objectives that significantly exceed the thresholds established for conventional business services, thereby creating a new benchmark for AI-ready networks.
Initially, the “Carrier Ethernet for AI” certification, targeted for launch in the second quarter of 2026, will concentrate on validating performance for AI inference workloads, which are often more sensitive to latency and jitter than the more bandwidth-intensive AI training workloads. Vachon indicated that a future certification tier may be developed to address the distinct high-bandwidth requirements of AI model training. The program is designed to address three primary AI-related use cases: connecting subscriber locations that run AI applications to AI edge computing sites; interconnecting these AI edge sites to larger, centralized AI data centers; and facilitating high-performance interconnections between AI data centers. A key value proposition of this certification lies in its ability to validate transport performance across complex, geographically distributed, multi-provider environments. By leveraging the standards-based network-to-network interfaces maintained by hundreds of carriers, the certification ensures the interoperability that is crucial for AI workloads spanning multiple carrier domains.
Market Implications and Strategic Alignment
Monetizing Infrastructure and Enabling NaaS
From a market perspective, this new certification program is strategically positioned to help service providers monetize the extensive global footprint of the $57 billion annual Carrier Ethernet market for emerging and lucrative AI applications. For mid-tier operators, achieving these new Mplify certifications can serve as a powerful market differentiator, allowing them to demonstrably prove that their networks can meet the same high-performance standards as their larger, more established competitors. This levels the playing field and opens up new revenue opportunities in the rapidly growing AI sector. Conversely, larger carriers, who may already be confident in their network’s ability to meet these advanced standards, might evaluate the need for formal certification more selectively, perhaps pursuing it for specific services or customer segments where such validation is a key purchasing criterion. This flexible approach allows the certification to serve different strategic purposes for various players within the telecommunications ecosystem.
This entire initiative is deeply integrated with Mplify’s overarching “NaaS for AI” strategy, which aims to create a more agile, automated, and responsive networking environment for AI. The alliance’s Lifecycle Service Orchestration (LSO) API standards are a key technical enabler for this vision, supporting the automated provisioning and dynamic management of network services across multiple provider domains. The synergy between the new AI certification and the LSO framework could ultimately allow for sophisticated capabilities such as dynamic bandwidth allocation and other on-demand network adjustments that are essential for efficiently managing distributed AI workloads. By creating a standardized way to validate AI performance and an automated framework to manage it, Mplify is reinforcing its commitment to translating its strategic vision for NaaS into tangible, real-world solutions that directly address the pressing needs of the AI era and provide a clear path forward for the industry.
A Forward-Looking Validation Framework
The introduction of this dual certification program represented a pivotal moment for the industry, providing a structured response to the network performance demands ignited by AI. It established a clear framework that both honored past investments in Ethernet and provided a definitive path toward future-ready infrastructure. The “Carrier Ethernet for Business” certification successfully ensured the continued relevance of a global standard, while the “Carrier Ethernet for AI” certification created a new, higher benchmark for performance. This strategic pivot allowed service providers to articulate the value of their networks in the context of AI, creating new opportunities for monetization and differentiation. The initiative’s integration with the broader NaaS and LSO automation strategy demonstrated a cohesive vision for the future of networking, one where performance could be not only validated but also dynamically managed to meet the fluid requirements of next-generation applications.
