Keysight Technologies is using ECOC 2026 to showcase testing, simulation and validation technologies aimed at one of the biggest challenges in AI infrastructure: moving increasingly large volumes of data between accelerators, servers and storage. The demonstrations span photonic integrated circuits, AI networking, 1.6T optical links, 3.2T research and AI-enabled test automation.
As AI clusters grow, the performance of the underlying network increasingly determines how effectively expensive compute resources can be used. At ECOC 2026 in Málaga, Spain, Keysight Technologies is demonstrating an end-to-end portfolio designed to help engineers develop and validate the optical and interconnect technologies needed for that next phase of data-center expansion.
The company will showcase the technologies at booth 1154 from September 21–23, covering photonic design, high-speed characterization, AI infrastructure benchmarking, optical-link validation and production testing.
One focus is the connection between photonic simulation and physical device testing. Keysight says its automated photonic integrated circuit (PIC) test capabilities can be combined with characterization reaching 220 GHz, creating a workflow that spans design validation through manufacturing. Its new N4378A Lightwave Component Analyzer is part of that demonstration, supporting electro-optic measurements of high-speed photonic components.
The shift toward higher-speed optical connectivity is becoming increasingly important as AI clusters expand. IDC reported that the worldwide Ethernet switch market reached $18.9 billion in the second quarter of 2026, up 43.4% year over year. Data-center Ethernet revenue increased 64.5% to $12.3 billion, with AI training and inference infrastructure identified as a major driver. 800GbE ports accounted for 41.2% of data-center Ethernet revenue during the quarter.
Keysight’s ECOC demonstrations move beyond today’s 800GbE environment toward 1.6T optical interconnects and research for 3.2T systems. Its 1.6T transmitter and receiver validation workflow combines bit-error-rate testing, reference transmitters and high-speed optical analysis to assess signal quality and performance margins.
That validation layer matters because increasing link speeds can amplify design and manufacturing uncertainties. Testing must establish whether optical components and interconnects maintain performance under increasingly demanding signaling conditions before systems reach volume production.
The company is also positioning network testing as part of the AI infrastructure stack itself. Its AresONE 1600GE test system and KAI Data Center Builder software are being demonstrated for benchmarking AI fabric performance under emulated real-world workloads. Keysight is also showing validation of AI transport technologies, including Ultra Ethernet-related mechanisms, alongside inference-scale testing.
This reflects a broader transition in AI infrastructure. IDC estimates worldwide AI infrastructure spending reached $318 billion in 2025 and forecasts spending to rise to $487 billion in 2026, with the market expected to exceed $1 trillion by 2029. The spending encompasses accelerated compute, servers and networking infrastructure, increasing the importance of tools that can validate increasingly complex systems before deployment.
For AI infrastructure providers, the challenge is no longer simply supplying more GPUs. Data must move efficiently between compute nodes, and network bottlenecks, optical losses, interoperability problems or manufacturing defects can undermine the performance of entire AI clusters.
Keysight is therefore connecting several traditionally separate engineering workflows: electronic and photonic design, physical measurement, interconnect testing, workload emulation and production automation. The approach is relevant to hyperscale data centers as well as semiconductor, optical-component and networking equipment manufacturers developing next-generation AI infrastructure.
AI is also being incorporated directly into the testing process. Keysight says its AI-enabled automation technologies can assist with photonics test development, troubleshooting and data analysis, potentially reducing the manual effort involved in interpreting large volumes of engineering measurements.
The ECOC demonstrations ultimately point to an infrastructure trend extending beyond faster individual components. As AI systems scale, optical links and network fabrics have to be designed, validated and manufactured as part of a much larger system. Companies including NVIDIA and other AI infrastructure vendors are driving demand for increasingly high-bandwidth networking, while suppliers across the optical ecosystem are racing to keep pace.
Keysight’s role is at the validation layer: providing the simulation, measurement, emulation and test infrastructure needed to determine whether those technologies perform as expected before they become part of large-scale AI deployments.
Market Landscape
AI infrastructure investment is pushing data-center networking toward higher bandwidths and more sophisticated validation. IDC’s latest Ethernet data shows particularly strong growth in data-center switching, while its AI infrastructure forecast points to continued expansion through the end of the decade.
The resulting opportunity extends across GPUs, Ethernet switches, optical transceivers, silicon photonics, cables, network fabrics and test equipment. At the same time, higher signaling speeds increase the need for automated characterization and production testing.
Top Insights
- Keysight is linking photonic design, optical measurement and automated testing to address the engineering complexity of next-generation AI infrastructure.
- Its ECOC demonstrations span 1.6T validation, 3.2T optical research, AI fabric benchmarking and high-speed interconnect testing.
- Data-center Ethernet growth is accelerating as hyperscalers expand infrastructure for AI training and inference workloads.
- AI-enabled test automation could help engineering teams analyze complex photonics measurements and troubleshoot high-speed infrastructure faster.
- The company’s approach reflects a broader shift toward validating AI infrastructure as an integrated compute, networking and optical system.
Power Tomorrow’s Intelligence — Build It with TechEdgeAI
