2025 Showcase
Welcome to the NextGenInfra Data Center Networking for AI and Cloud Showcase!
As the world invests multiple hundreds of billions to trillions of dollars in data center build out for AI, what's the role of networking? How should networking serve both AI and Cloud workloads? Will the shift from pre-training to post-training to test-time/inference-time scaling change the data center computing and networking requirements? So many questions, but do we know the answers?
To help our readers, we've captured insights from the leading thinkers in the data center networking ecosystem here in our showcase. The content in the videos and our report highlight the state-of-the-art in networking scale-up, scale-out, and scale-outside. What's scale-outside? Download the report to find out!
Thanks to our sponsors for making the report possible!
Data Center Networking for AI and Cloud
Welcome to the NextGenInfra Data Center Networking for AI and Cloud Showcase!
As the world invests multiple hundreds of billions to trillions of dollars in data center build out for AI, what's the role of networking? How should networking serve both AI and Cloud workloads? Will the shift from pre-training to post-training to test-time/inference-time scaling change the data center computing and networking requirements? So many questions, but do we know the answers?
To help our readers, we've captured insights from the leading thinkers in the data center networking ecosystem here in our showcase. The content in the videos and our report highlight the state-of-the-art in networking scale-up, scale-out, and scale-outside. What's scale-outside? Download the report to find out!
Thanks to our sponsors for making the report possible!
Sponsor Spotlight Alphawave Semi
Use Cases for Chiplets in AI Clusters
Tony Chan Carusone, Chief Technology Officer at Alphawave Semi, discusses how AI demand and maturing die-to-die interface standards are accelerating chiplet adoption. He outlines three main applications: connecting compute chiplets with high bandwidth and low latency, linking compute cores to IO chiplets for extended connectivity, and interfacing compute chiplets with optical components for long-distance AI cluster communication.
Sponsor Spotlight Arrcus
AI Infrastructure Evolution
Shekar Ayyar, CEO of Arrcus, outlines how networking infrastructure supports AI's progression from training to distributed inference applications through their multi-platform networking fabric and operating system. He discusses Arrcus's IPsec offloading for AI data centers and cost management solutions, while highlighting Ethernet's expanding role in connecting GPU stacks to edge computing.
Sponsor Spotlight Astera Labs
Four Connectivity Fabric Innovations for AI Clusters
Paroma Sen, Vice President of Corporate Marketing at Astera Labs, showcases four major data center innovations including PCI Gen 6 solutions, Cosmos software suite, and Leo CXL smart memory controllers 5. She shares the company's expanding technology portfolio while highlighting their leadership in UALink development.
Sponsor Spotlight Baya Systems
Breaking the AI Data Bottleneck
Sailesh Kumar, CEO of Baya Systems, leads the development of Network on Chip (NoC) technology that addresses data movement challenges in AI compute architectures. Baya has successfully delivered its chiplet-ready foundational product to multiple customers and is expanding into automotive and data center markets.
Sponsor Spotlight DriveNets
Scheduled Ethernet Fabric for Data Center Infrastructure
Dudy Cohen, VP, Product Marketing at DriveNets, outlines how data centers can support AI workloads through advanced networking solutions that eliminate jitter and packet drop while maintaining low latency. His showcases the DriveNets Network Cloud AI solution, which uses scheduled Ethernet fabric to connect thousands of GPUs through a single hop while operating on cost-effective white box hardware.
Sponsor Spotlight Juniper
Ethernet Matches InfiniBand for AI Clusters
Amit Sanyal, Senior Director of Data Center Product Marketing at Juniper, shares how Ethernet performance matches InfiniBand in AI clusters through benchmarks from Juniper's AI lab and major hyperscalers. His analysis shows how leading AI providers select Ethernet for its extensive ecosystem and cost advantages, while highlighting their focus on power efficiency through liquid cooling and advanced switch designs.
Sponsor Spotlight Marvell
Speeding Forward with PCIe Gen 6 and 7
Annie Liao, Product Management Director at Marvell, presents PCIe Gen 6 retimer solutions and PCIe Gen 7 technology critical for scaling up AI infrastructure. The demonstrations include a three-board PCIe Gen 6 setup with retimer capabilities, and a PCIe Gen 7 system achieving 128G transfer speeds with improved bit error rates through TSMC's 3nm process.
Sponsor Spotlight Marvell
Active Electrical Cables for AI Server Architecture
Winnie Wu, Senior Director of Product Marketing at Marvell, describes how data center architectures are adapting to AI server needs through changes in switch placement and rack configurations. She outlines their 7-meter 28-gauge cables supporting 800G breakout connections and their 1.6T product with 32-gauge cables optimized for GPU-to-GPU connectivity.
Sponsor Spotlight Multilane
Testing 1.6T Cables at 200GB/Lane
Hani Daou, Business Development Manager at Multilane, explains how AI and cloud providers are pushing the boundaries for faster, higher-density network connectivity. He showcases Multilane's new 200 GB per lane bit error rate tester with a 3M 1.6 terabit active electrical cable, helping vendors advance their high-speed interconnect solutions for AI clusters.
Sponsor Spotlight Nokia
AI Data Centers and the Hyperscale Model
Mike Bushong, Vice President of Data Center at Nokia, explains that AI infrastructure goes beyond just hyperscalers and spans multiple markets with distinct geographical constraints and requirements for power and space. He outlines how Nokia is addressing these needs through partnerships, multivendor management, and comprehensive networking solutions.
Sponsor Spotlight NVIDIA
Network Architecture for Scaling AI
Kevin Deierling, Senior Vice President of Networking at NVIDIA, outlines the company's scale-up and scale-out networking technologies, focusing on their GB200 NVL72 architecture, InfiniBand offerings, and Spectrum-X Ethernet platform. He discusses scale-up and scale-out scenarios, highlighting how their optimized Ethernet offering can support 100K+ GPU clusters.
Sponsor Spotlight UALink Consortium
Networking Massive GPU Clusters
Kurtis Bowman, Chairman of UALink Consortium, is leading the development of an Ultra Accelerator Link technology that enables connecting up to 1,024 GPUs for enhanced AI capabilities. The initiative, supported by major tech companies and over 75 consortium members, aims to create an open standard supporting 200-800 Gbps data rates with low latency communication between GPUs.
Courtesy of our sponsors!
Complimentary Report Download
2025 Data Center Networking
UNLOCKED REPORT — no personal info needed to download
Download 2025 Data Center Networking Report
By downloading this report, you agree to our Terms of Use and Privacy Policy.
Analysis by AvidThink
What the 2025 report finds
Across a dozen interviews, the 2025 report captures an AI-driven networking market splitting into two problems, scale-up inside the rack and scale-out across the cluster, with Ethernet mounting a credible challenge to InfiniBand and a wave of new interconnect standards and physical-layer parts racing to keep GPUs fed.
Ethernet closes the gap on InfiniBand
Juniper's AI-lab and hyperscaler benchmarks show Ethernet matching InfiniBand performance in AI clusters, and buyers increasingly pick it for ecosystem breadth and cost. Even NVIDIA leans in with Spectrum-X, positioned to support 100K+ GPU clusters.
Scale-up gets its own open standard
UALink emerges as an open scale-up fabric targeting up to 1,024 connected GPUs at 200-800 Gbps, backed by 75+ consortium members. Astera Labs stakes out an early leadership position in UALink development.
The physical layer is the bottleneck
Feeding GPUs is now a copper-and-silicon problem: Marvell pushes PCIe Gen 7 to 128G on TSMC 3nm and ships 1.6T active electrical cables, while Multilane tests at 200 GB per lane. Chiplets and die-to-die links (Alphawave, Baya) attack data movement inside the package.
AI infrastructure spreads beyond hyperscalers
Nokia argues AI buildout spans many markets with distinct power, space, and geographic constraints, not one hyperscale template. Arrcus echoes this, extending GPU-connecting Ethernet fabric out toward the edge.
Power and determinism become design centers
Liquid cooling and advanced switch designs recur as power-efficiency levers, and lossless behavior matters as much as raw speed. DriveNets' scheduled Ethernet fabric eliminates jitter and packet drop to link thousands of GPUs in a single hop on white box hardware.
Highlights from industry thought leaders
Video interviews
Use Cases for Chiplets in AI Clusters
- AI demand plus maturing die-to-die interface standards are the twin drivers accelerating chiplet adoption.
- Three chiplet use cases: compute-to-compute (high bandwidth, low latency), compute-to-IO for extended connectivity, and compute-to-optics.
- Compute-to-optical interfacing targets long-distance communication across AI clusters.
Abstract
Tony Chan Carusone, Chief Technology Officer at Alphawave Semi, discusses how AI demand and maturing die-to-die interface standards are accelerating chiplet adoption. He outlines three main applications: connecting compute chiplets with high bandwidth and low latency, linking compute cores to IO chiplets for extended connectivity, and interfacing compute chiplets with optical components for long-distance AI cluster communication.
AI Infrastructure Evolution
- AI is progressing from training toward distributed inference, and networking fabric must follow the workload.
- Arrcus offers IPsec offloading and cost-management for AI data centers atop a multi-platform fabric and OS.
- Ethernet's role is expanding to connect GPU stacks all the way out to edge computing.
Abstract
Shekar Ayyar, CEO of Arrcus, outlines how networking infrastructure supports AI's progression from training to distributed inference applications through their multi-platform networking fabric and operating system. He discusses Arrcus's IPsec offloading for AI data centers and cost management solutions, while highlighting Ethernet's expanding role in connecting GPU stacks to edge computing.
Four Connectivity Fabric Innovations for AI Clusters
- Astera Labs spans four connectivity innovations: PCIe Gen 6, the Cosmos software suite, and Leo CXL smart memory controllers.
- CXL smart memory controllers position Astera in the memory-expansion layer of AI clusters.
- The company is positioning itself as a leader in UALink development.
Abstract
Paroma Sen, Vice President of Corporate Marketing at Astera Labs, showcases four major data center innovations including PCI Gen 6 solutions, Cosmos software suite, and Leo CXL smart memory controllers 5. She shares the company's expanding technology portfolio while highlighting their leadership in UALink development.
Breaking the AI Data Bottleneck
- Baya's Network on Chip (NoC) technology targets on-die data movement bottlenecks in AI compute architectures.
- Its chiplet-ready foundational product has already shipped to multiple customers.
- Baya is expanding beyond its core into automotive and data center markets.
Abstract
Sailesh Kumar, CEO of Baya Systems, leads the development of Network on Chip (NoC) technology that addresses data movement challenges in AI compute architectures. Baya has successfully delivered its chiplet-ready foundational product to multiple customers and is expanding into automotive and data center markets.
Scheduled Ethernet Fabric for Data Center Infrastructure
- DriveNets Network Cloud AI uses scheduled Ethernet fabric to eliminate jitter and packet drop at low latency.
- The fabric connects thousands of GPUs through a single hop.
- It runs on cost-effective white box hardware rather than proprietary switches.
Abstract
Dudy Cohen, VP, Product Marketing at DriveNets, outlines how data centers can support AI workloads through advanced networking solutions that eliminate jitter and packet drop while maintaining low latency. His showcases the DriveNets Network Cloud AI solution, which uses scheduled Ethernet fabric to connect thousands of GPUs through a single hop while operating on cost-effective white box hardware.
Ethernet Matches InfiniBand for AI Clusters
- Juniper's AI-lab and hyperscaler benchmarks show Ethernet matching InfiniBand performance in AI clusters.
- Leading AI providers pick Ethernet for its broad ecosystem and cost advantages.
- Power efficiency is pursued through liquid cooling and advanced switch designs.
Abstract
Amit Sanyal, Senior Director of Data Center Product Marketing at Juniper, shares how Ethernet performance matches InfiniBand in AI clusters through benchmarks from Juniper's AI lab and major hyperscalers. His analysis shows how leading AI providers select Ethernet for its extensive ecosystem and cost advantages, while highlighting their focus on power efficiency through liquid cooling and advanced switch designs.
Speeding Forward with PCIe Gen 6 and 7
- Marvell demoed PCIe Gen 6 retimers and Gen 7 technology as critical enablers for scaling up AI infrastructure.
- A PCIe Gen 7 system hit 128G transfer speeds with improved bit error rates on TSMC's 3nm process.
- The Gen 6 demo used a three-board setup showcasing retimer capabilities.
Abstract
Annie Liao, Product Management Director at Marvell, presents PCIe Gen 6 retimer solutions and PCIe Gen 7 technology critical for scaling up AI infrastructure. The demonstrations include a three-board PCIe Gen 6 setup with retimer capabilities, and a PCIe Gen 7 system achieving 128G transfer speeds with improved bit error rates through TSMC's 3nm process.
Active Electrical Cables for AI Server Architecture
- AI servers are reshaping data center architecture via new switch placement and rack configurations.
- Marvell's 7-meter 28-gauge active electrical cables support 800G breakout connections.
- Its 1.6T product uses 32-gauge cables optimized for GPU-to-GPU connectivity.
Abstract
Winnie Wu, Senior Director of Product Marketing at Marvell, describes how data center architectures are adapting to AI server needs through changes in switch placement and rack configurations. She outlines their 7-meter 28-gauge cables supporting 800G breakout connections and their 1.6T product with 32-gauge cables optimized for GPU-to-GPU connectivity.
Testing 1.6T Cables at 200GB/Lane
- AI and cloud providers are driving demand for faster, higher-density network connectivity.
- Multilane's new bit error rate tester operates at 200 GB per lane.
- It was demonstrated with a 3-meter 1.6 terabit active electrical cable for AI-cluster interconnects.
Abstract
Hani Daou, Business Development Manager at Multilane, explains how AI and cloud providers are pushing the boundaries for faster, higher-density network connectivity. He showcases Multilane's new 200 GB per lane bit error rate tester with a 3M 1.6 terabit active electrical cable, helping vendors advance their high-speed interconnect solutions for AI clusters.
AI Data Centers and the Hyperscale Model
- AI infrastructure extends well beyond hyperscalers into multiple markets with distinct power and space constraints.
- Geography shapes AI data center requirements, not a single hyperscale template.
- Nokia's approach centers on partnerships, multivendor management, and comprehensive networking.
Abstract
Mike Bushong, Vice President of Data Center at Nokia, explains that AI infrastructure goes beyond just hyperscalers and spans multiple markets with distinct geographical constraints and requirements for power and space. He outlines how Nokia is addressing these needs through partnerships, multivendor management, and comprehensive networking solutions.
Network Architecture for Scaling AI
- NVIDIA spans scale-up and scale-out with GB200 NVL72, InfiniBand, and the Spectrum-X Ethernet platform.
- Spectrum-X, its optimized Ethernet offering, is positioned to support 100K+ GPU clusters.
- NVIDIA embraces both InfiniBand and Ethernet rather than betting on a single fabric.
Abstract
Kevin Deierling, Senior Vice President of Networking at NVIDIA, outlines the company's scale-up and scale-out networking technologies, focusing on their GB200 NVL72 architecture, InfiniBand offerings, and Spectrum-X Ethernet platform. He discusses scale-up and scale-out scenarios, highlighting how their optimized Ethernet offering can support 100K+ GPU clusters.
Networking Massive GPU Clusters
- UALink is an open standard for scale-up accelerator interconnect, targeting up to 1,024 connected GPUs.
- The spec aims for 200-800 Gbps data rates with low-latency GPU-to-GPU communication.
- The consortium counts over 75 members backed by major tech companies.
Abstract
Kurtis Bowman, Chairman of UALink Consortium, is leading the development of an Ultra Accelerator Link technology that enables connecting up to 1,024 GPUs for enhanced AI capabilities. The initiative, supported by major tech companies and over 75 consortium members, aims to create an open standard supporting 200-800 Gbps data rates with low latency communication between GPUs.
Reader feedback on our latest report: "Appreciate the broad coverage and unique insights." Download our 2025 Data Center Networking report now and send us your feedback!
Download our 2025 Data Center Networking Report
By downloading this report, you agree to our Terms of Use and Privacy Policy.
Continue exploring
More executive showcases across next-gen infrastructure.
2025
2025 MPLS World Congress
Leading executives at MPLS World Congress 2025 in Paris discuss latest in IP networking, AI networking, IPv6, SRv6, converged optics
2025
2025 OCP Global Summit
Exclusive videos from thought leaders covering data center networking (scale-up, scale-out, scale-across), photonics, liquid cooling, SONiC
2026
2026 Data Center Networking
Infrastructure leaders discuss data center networking and impact of AI/GenAI. Topics: InfiniBand, UEC, NVLink, UALink, ESUN