IP Infusion's OcNOS 7.0 Aims to Power AI and 800G Network Evolution

📊 Key Data
  • 600+ global customers using IP Infusion's OcNOS platform
  • 40-60% TCO savings with disaggregated solutions vs. legacy vendors
  • Support for 400G/800G ZR/ZR+ optics for high-capacity transport
🎯 Expert Consensus

Experts agree that OcNOS 7.0 represents a significant step toward open, disaggregated networking, offering cost savings and flexibility for AI and high-capacity transport needs.

2 months ago
IP Infusion's OcNOS 7.0 Aims to Power AI and 800G Network Evolution

IP Infusion's OcNOS 7.0 Targets AI and 800G Network Evolution

BARCELONA, Spain – February 25, 2026 – At Mobile World Congress 2026, IP Infusion unveiled OcNOS® 7.0, a significant update to its open network operating system poised to address two of the industry's most pressing challenges: the explosive networking demands of Artificial Intelligence and the complex migration to 400G and 800G optical transport. The launch signals a deepening industry shift away from traditional, proprietary hardware toward open, disaggregated solutions that promise greater efficiency, flexibility, and substantial cost savings.

With a base of over 600 global customers, IP Infusion is positioning its flagship platform as the foundation for a new generation of network infrastructure, spanning service provider backbones, emerging neo-cloud data centers, and high-capacity data center interconnects (DCI).

"OcNOS 7.0 has become the Service Provider’s strategic evolution platform of choice," said Miguel Alonso, Chief Product and Marketing Officer of IP Infusion, in a statement accompanying the announcement. "We are providing operators with the field-proven tools to accelerate the evolution of legacy infrastructure, secure the internet edge, and automate the entire lifecycle of their networks."

Powering the AI Revolution with Open Fabrics

The rise of large-scale AI and machine learning workloads has placed unprecedented strain on traditional data center networks. The distributed nature of GPU clusters, which require constant, high-throughput communication for training models, is highly sensitive to latency and packet loss. A delay of mere microseconds can lead to underutilized, multi-million-dollar GPU assets and significantly extended training times.

Historically, this high-performance niche has been dominated by InfiniBand, a specialized networking standard known for its ultra-low latency. However, OcNOS 7.0 is part of a growing movement championing a highly optimized Ethernet as a powerful and more cost-effective alternative. The new release streamlines AI fabrics by delivering low-latency, lossless networking designed to maximize GPU efficiency. This is achieved through robust support for technologies like Remote Direct Memory Access (RDMA) over Converged Ethernet (RoCEv2), which requires a meticulously managed network environment.

OcNOS 7.0 implements critical features such as Priority Flow Control (PFC) and Explicit Congestion Notification (ECN) to prevent the packet loss that can cripple AI workloads. By creating a guaranteed lossless fabric on open, white-box hardware, IP Infusion argues that operators can build AI-ready networks without being locked into a single vendor's ecosystem. This open approach, combined with new cloud-native capabilities like Docker and Kubernetes support on the router itself, allows for greater automation and the ability to run custom monitoring or security tools directly within the network fabric.

The Economic Imperative of Disaggregation

Beyond the technical specifications, a core driver of the shift toward open networking is economics. IP Infusion claims that by combining OcNOS 7.0 with commodity white-box hardware from a variety of manufacturers, operators can achieve Total Cost of Ownership (TCO) savings of 40-60% compared to closed, vertically integrated solutions from legacy vendors.

This claim is supported by broader industry analyses, which point to significant cost reductions in both capital and operational expenditures when adopting disaggregated models. The savings stem from several factors. First, separating network software from hardware breaks vendor lock-in, fostering a competitive hardware market that drives down initial purchase costs (CapEx). Operators are free to choose the best-in-class switches and optics that fit their specific price and performance needs.

Second, operational savings (OpEx) are realized through streamlined automation, reduced power consumption, and simplified inventory management. The move toward integrated architectures like IP over DWDM (IPoDWDM), a key feature of OcNOS 7.0, further amplifies these benefits by reducing the physical footprint and energy needs of the network. By eliminating entire layers of equipment, such as separate optical transponders, operators can dramatically lower costs associated with power, cooling, and rack space.

Simplifying the Leap to 400G/800G with IPoDWDM

As internet traffic continues its exponential growth, service providers and data center operators are racing to upgrade their transport networks to 400G, 800G, and beyond. The OcNOS 7.0 release directly addresses this evolution with a major focus on IP over DWDM (IPoDWDM), an architecture that integrates coherent optical transceivers directly into IP routers.

This convergence of the IP and optical layers promises to radically simplify network design and management. Instead of managing separate IP routing and optical transport systems, each with its own specialized hardware and operational teams, IPoDWDM creates a single, unified packet-optical network. OcNOS 7.0 facilitates this by supporting the latest 400G and 800G ZR/ZR+ pluggable optics from multiple vendors, giving operators true control over their network evolution.

To manage the increased complexity at the control plane, the release introduces a comprehensive Segment Routing feature set. Capabilities like Segment Routing Traffic Engineering and Flexible Algorithm (Flex-Algo) allow for the creation of deterministic, service-aware paths across the network. This simplifies traffic management and eliminates the complexities of older protocols, providing a more scalable and efficient foundation for high-capacity transport. While integrating traditionally siloed IP and optical domains presents organizational challenges, the long-term benefits of a simplified, cost-effective, and power-efficient architecture are proving to be a powerful incentive for the industry.

Ecosystem in Action: From Theory to Deployment

Underscoring the maturity of its open networking approach, IP Infusion and its partners are showcasing tangible implementations of these technologies at MWC 2026. At the UfiSpace booth, OcNOS is powering a range of disaggregated carrier solutions, from high-density cell site gateways to aggregation switches designed for 5G-Advanced and edge AI applications.

A particularly compelling demonstration is being held with Fujitsu, which is showcasing a total IPoDWDM solution that integrates its 1FINITY™ optical platform with OcNOS. This collaboration is presented within the context of the Innovative Optical and Wireless Network (IOWN) Global Forum, an initiative focused on building next-generation, photonics-based infrastructure. The joint solution highlights the practical benefits of IP and optical integration, aiming to deliver significant reductions in latency and power consumption—key goals of the IOWN framework.

These real-world showcases demonstrate that the vision of open, disaggregated networking is moving beyond theoretical benefits and into practical, deployable solutions. By fostering a robust ecosystem of hardware and software partners, IP Infusion is building a compelling case that the future of networking will be built on principles of openness, choice, and automation, providing the critical infrastructure needed to support the next wave of digital transformation.

Event: Industry Conference
Theme: Artificial Intelligence Machine Learning Automation ESG
Metric: Financial Performance
Sector: Software & SaaS AI & Machine Learning Fintech
Product: ChatGPT
UAID: 18170