Nokia SReXperts 26: Key Takeaways
In this blog
This year at SReXperts, World Wide Technology was named Nokia's 2026 U.S. Partner of the Year. It's a recognition of the work our teams do alongside Nokia for our customers, and I'm proud to be part of it. It also made this year's event especially memorable for those of us on the ground in San Diego.
Over the past few years, I've watched Nokia use their annual SReXperts conference to showcase significant evolutions in their hardware and software portfolio. This has included introducing the FPcx-based 7730 SXR family, modernizing the long-lived 7705 SAR platforms, and moving their software strategy towards model-driven architectures and SR-Linux.
This year Nokia celebrated the 20th anniversary of SReXperts and Wavelengths in San Diego, California—and there was a noticeable shift in emphasis. Rather than focusing on updated hardware or evolving software, Nokia spent much of the conference discussing how networks will need to adapt to the era of artificial intelligence.
Here are my biggest takeaways from SReXperts 2026.
AI Grid: The Big Picture
If there was one theme that dominated SReXperts this year, it was Nokia's vision for the AI Grid.
Most of us are familiar with the enormous infrastructure investments taking place inside AI data centers. High-performance GPUs, low-latency Ethernet fabrics and massive power requirements have become familiar topics across the industry.
But Nokia's argument is that AI infrastructure won't remain confined to centralized data centers.
As AI inference becomes more widely adopted, workloads will increasingly be distributed across regional facilities, metro locations, enterprise premises and edge computing environments. Connecting these resources will require high-capacity, low-latency and highly reliable networks.
That's where the carrier WAN comes into play.
Nokia sees an opportunity for carriers to move beyond simply selling bandwidth and begin offering differentiated connectivity services for AI workloads. These might include guaranteed performance SLAs, secure data center interconnect, multicloud connectivity and access to distributed AI computing resources.
One presentation described a progression from traditional best-effort IP connectivity all the way to AI infrastructure access, where service providers could potentially bundle connectivity with GPU resources through partnerships.
It's an interesting vision, particularly for carriers looking for new revenue opportunities as traditional connectivity services become increasingly commoditized.
Nokia is positioning its IP, optical and data center portfolios to support this AI Grid as a cohesive architecture, with the 7750 SR, 7250 IXR, 7730 SXR, 7220 IXR and coherent optical solutions all playing important roles.
New Routers, New Capabilities
As usual, one of the first things I wanted to see at SReXperts was what Nokia had been working on in the hardware department.
Last year's conference featured a substantial number of new router introductions, including the 2nd-generation 7705 SAR platforms, 3rd-generation 7250 IXR-e family, and the 7250 IXR-X4.
This year, Nokia provided updates on those products while introducing several additional hardware options.
7705 SAR-10x
One particularly interesting addition is the 7705 SAR-10x, which adds a fully modular chassis to Nokia's 2nd-gen SAR portfolio. Designed for mission-critical environments, it offers eight MDA card slots, redundant control and forwarding modules, and support for legacy interfaces. Its hardened, passively-cooled design makes it particularly well suited for utilities, transportation and other mission-critical use cases.
7730 SXR-R1
The new 7730 SXR-R1 is a compact 2RU router featuring Nokia's FPcx forwarding silicon, 4.4 Tbps of full-duplex capacity, high-speed QSFP-DD and SFP112 interfaces, and a redundant control plane.
I've been following the 7730 SXR family closely since its introduction at SReXperts 2023. The idea behind FPcx remains compelling: provide the advanced forwarding features and scale associated with custom routing silicon without requiring the high throughput of a core router. The SXR-R1 is a nice addition to that strategy.
Nokia also discussed the ongoing upgrade path from certain 7250 IXR-R6d/R6dl systems to FPcx-based 7730 SXR configurations. The ability to reuse portions of an existing hardware investment while transitioning between forwarding architectures remains one of the more interesting aspects of this product family.
Additional hardware developments
Nokia also added to the 7250 IXR lineup with the brand new IXR-s3x, which brings 800G QSFP-DD interfaces to the mixed-port aggregation scale (2x QDD-800, 4x QSFP28, 48x SFP56).
One notable data center platform is the 7220 IXR-H6, which offers up to 64 × 1.6TbE interfaces in a single system. While the 7220 is outside Nokia's traditional service-router portfolio, the increasing bandwidth requirements of AI data centers help explain why Nokia is investing so heavily in both data center fabrics and the networks connecting them.
AI Takes the Wheel — Carefully
Artificial intelligence has been a recurring theme at SReXperts for several years, but the AI developments showcased in 2026 go substantially further.
Nokia is working toward agentic network operations, where AI agents can do more than answer documentation questions or summarize telemetry. Network Services Platform (NSP) is being expanded with an agentic AI framework designed to assist with troubleshooting, network optimization and operational workflows. The architecture incorporates live network data, domain-specific knowledge, controlled execution and guardrails intended to make agent actions traceable and explainable.
For network engineers who have spent years building automation with Python, APIs and model-driven interfaces, it's fascinating to see the industry beginning to move toward systems that can reason about network conditions and coordinate those same tools.
There is still a meaningful gap between today's automation capabilities and the long-term vision of fully autonomous networking. Nokia's broader AgenticOps rollout extends into 2027, so this remains an evolving technology rather than a finished solution.
Nevertheless, this was one of the most significant developments at the conference.
Quantum-Safe Networking Continues to Evolve
Quantum-safe networking remains an important area of investment for Nokia. The challenge is familiar: encrypted traffic captured today could potentially be decrypted in the future if sufficiently capable quantum computers become available.
Nokia has been addressing this through a combination of post-quantum cryptography, platform security and high-performance network encryption. One technology that received considerable attention this year was ANYsec, Nokia's approach to extending high-performance encryption across more networking scenarios. Alongside traditional MACsec, ANYsec is intended to provide additional flexibility in protecting traffic as it traverses different network architectures and service boundaries.
Nokia also discussed post-quantum cryptographic algorithms for management and control-plane security, along with improvements to secure boot and platform trust.
For carrier networks, the opportunity extends beyond protecting their own infrastructure: high-performance encryption could also become a differentiated service offering.
What's Next for IP and Optical Network Convergence?
Nokia's optical portfolio has expanded greatly in the aftermath of the Infinera acquisition, and now Nokia seems to be aiming this expanded portfolio directly at the AI craze. The strategy is to bring IP, Optical, and Data Center technologies together to build networks that can support the future workloads of distributed AI inferencing.
Optical convergence via Digital Coherent Optics (DCO) is critical to that strategy. Nokia offers a wide range of DCO capabilities from 100G QSFP-28 ZR to 400G QSFP56-DD ZR/ZR+ to 800G QSFP112-DD ZR/ZR+. These capabilities include high launch-power versions and, in the case of 800G ZR/ZR+, an OSFP option coming soon.
While the DCO plugs provide IP and Optical network convergence across long distances, shorter-range optics such as 800G QSFP and 1.6T OSFP add high performance capability inside the data center, completing the vision of a unified, cohesive IP+Optical+DC network strategy.
As AI traffic continues to grow, improving the efficiency and operational coordination of these network layers will become increasingly important.
Is FP6 Around the Corner?
One question I've been particularly interested in is what comes next for Nokia's routing silicon.
The FP5 generation has been around for over five years now, and Nokia continues to invest in its capabilities. This year's conference highlighted support for 800GE interfaces, coherent routing, integrated security and several power-efficiency improvements. Some of those enhancements allow routers to reduce power consumption by deactivating unused resources or adjusting active fabric capacity.
In my opinion, this continued investment into FP5 highlights just how good Nokia is at investment protection. When FP5 was first announced in 2021, 100G was still extremely common so 400G was the logical sweet spot to target. Instead of settling for that, Nokia designed FP5 to be 800G-capable and that decision is paying dividends now. Two years ago they even developed the E5 chip to bring 800G speeds to FP4 routers.
FP6, whenever it does come, will come with a jump from 112G to 224G SerDes for higher throughput, a newer and smaller process node, improved power efficiency, and native 1.6T-class ports. These jumps certainly sound far out to me. But given how ahead of the curve the FP5 was, I expect them to come sooner rather than later.
Final Thoughts
Looking back at the past several SReXperts conferences, there's a clear progression.
In 2023, Nokia introduced the FPcx-based 7730 SXR and made a major push toward model-driven software.
In 2024 and 2025, the company expanded and refreshed its hardware portfolio while continuing to develop its automation capabilities.
In 2026, those investments are beginning to converge around a larger strategy: building networks that can autonomously support distributed AI workloads.
The new hardware remains interesting, but my biggest takeaway from SReXperts 2026 is that Nokia is increasingly focused on making the network itself more intelligent, secure and valuable.