Xsight Labs

Kicking off our Networking Field Day replay series, Eric Vallone explores open software in Segment 1. By pairing standard Linux with an accessible SDK, engineering teams can build, test, and deploy network apps without reinventing the wheel. Watch Part 1: youtube.com/watch?v=mCcNynA3… #Linux #NetworkingFieldDay #OpenSource #DataPlane #CloudNetworking #NFD41

From #programmable data planes to ultra-efficient #silicon, Yoav Galon and Eric Vallone showed why the chip inside your #switch matters more than ever at @TechFieldDay session at #NFD41. Check out the full talk: youtube.com/watch?v=l-uD4E-G… #Networking

We had a fantastic time presenting at Networking Field Day with @NetworkingNerd. Eric Vallone and Yoav Galon, took the stage to dive into how Xsight Labs is accelerating next-generation #datacenter architectures, driving #silicon efficiency, and pushing the boundaries of high-performance #networking. Huge thanks to the delegates, the @TechFieldDay team, and everyone who tuned in online and engaged in deep-dive technical discussions with us. ICYMI, watch the recording of the live broadcast and catch up on the sessions here: 👉 techfieldday.com/appearance/… #NetworkingFieldDay #NFD41 #DataCenter #Networking #Semiconductors #Innovation #TechFieldDay

Great insights from Eric Vallone and Yoav Galon at #NFD41!

Xsight Accelerate adds packet manipulation and stateful mirror-on-drop to the X2 Network Packet Broker! Here is what this means for your network visibility: ⚡ Instant Line-Rate Visibility: Truncate, encapsulate, and mirror dropped packets at full line rate with complete state awareness. 🎯 Zero-Latency Jitter: Perform advanced header modifications, slicing, and metadata tagging without sacrificing throughput. 💰 Optimized TCO: Ultra-dense NPB performance at industry-leading power efficiency, fully integrated with open NOS and #SONiC. Eliminate monitoring blind spots without paying a performance tax. Pameely Rashid and Gagandeep Singh share more: xsightlabs.com/blog/xsight-a… #NetworkPacketBroker #NPB #NetworkObservability #DataCenter #MirrorOnDrop #SONiC #Networking @Gagan4344

Last week, we announced the expansion of the Xsight Labs E-Series #DPU portfolio with the #E1L, bringing full, programmable DPU offload starting at just 25 Watts because power and cooling constraints shouldn’t mean compromising on full DPU offload.🔌 Built on the exact same architecture as our flagship 800G E1: ⚡ Extreme Efficiency: High-performance programmable acceleration for edge & power-constrained environments. 🔄 Software Consistency: Write your offload once, run it seamlessly from core to edge. 🛠️ Full Programmability: Stay in complete control of your pipeline. Dive into the architecture with Yoav Galon: xsightlabs.com/blog/introduc… #DPU #EdgeComputing #Networking #HardwareInnovation #Silicon

Our own @TheRealJKaplan had a chance to sit down with James E. Carroll from @NextGenInfra_io and Roy Chua from @AvidThink at #OCP2026 to talk about the massive growth Xsight Labs has seen since the last #OCP. From the #datacenter to #outerspace, we've literally been everywhere in the world and beyond. Check out the key takeaways at: nextgeninfra.io/2026-ocp-glo…

"Xsight Labs is just different." That's how @Patrick1Kennedy at @ServeTheHome summed up the brand new E1L. While much of the industry is racing toward 800G and beyond, we took our E1, the industry's first 800G DPU, and scaled it down to 200G at just 15W typical draw. Bringing #DPU offload where power matters most. 🔋 ⚡ Vanilla Arm, on purpose. The E1’s Neoverse Arm N2 cores are standard enough that STH installed out-of-the-box #Ubuntu LTS on it. The E1L carries the same approach: Linux and DPDK on Arm cores, with no proprietary environment. ⚡ Switch-ready form factor. The E1L comes on a COM Express Type 7 module, and Patrick’s first thought was “that works in network switches.” It also suits #NVMe and disk storage servers. ⚡ Seamless scaling. Same architecture and software stack as our 32-64 core 800G E1—meaning your optimisation path carries straight over to the 8-22 core E1L. Dive into the full review from #ServeTheHome: servethehome.com/xsight-labs… #EdgeComputing #Networking #Hardware

Heading into the final day of #OCP2026. Our customer and partner suite is booked solid and the booth has been buzzing. Head over to Booth #T21 to talk to our team and check out our #programmableswitch and #DPU.

Day 1 of #OCP2026 is in the books, and the floor is buzzing! From nonstop traffic at our booth to deep-dive strategy sessions with customers in our suite, the energy around open compute innovation this year is next-level. Huge thank you to everyone who stopped by today to share insights, talk shop, and explore what’s next for the #openecosystem. If we haven’t connected yet, let’s change that. @OpenComputePrj

All set up at #OCP2026 and excited to see everyone in our Suite and Booth #T21! Who is coming?! #networking #dpu #switch @OpenComputePrj

Introducing the E1L: Expanding our E-series DPU portfolio with fully programmable offload starting at just 25 Watts. As #AI shifts from answering questions to #autonomousagents taking action, background infrastructure workloads, networking, storage, and security, are increasingly competing with AI applications for valuable host CPU resources. Bringing #DPU offload into power-constrained #edge and embedded environments used to mean accepting proprietary software stacks or leaving offload out entirely. Not anymore. Powered by Arm ®Neoverse® N2 cores and TSMC’s N5 process, the new E1L brings fully programmable DPU capabilities up to 200G within a 25W–40W TDP power envelope (with typical draw as low as 15W). 🔹 Unified Architecture & Software: Runs on the exact same open, vendor-neutral software stack as our 800G E1 DPU—write once, deploy anywhere from 100G edge to 800G core. 🔹 Open Standards: Powered by standard Linux and integrated with XDP, DPDK, and SPDK. No proprietary software licensing or lock-in. 🔹 Versatile Form Factors: Available as a discrete device, COM Express module, and 1RU server. 🔹 Robust Compute & Security: Offers 8, 16, or 22 Arm Neoverse N2 cores, up to 20 PCIe Gen5 lanes, DDR5-5200 with inline encryption, and line-rate AES-GCM IPsec offload. Read the full announcement here: prnewswire.com/news-releases…

We are heading to the OCP Global Summit, and our team is ready to connect. Discuss custom solutions with our engineering and leadership teams. Grab some exclusive swag! Whether you want to talk #AI, dive into product demos, or discuss the future of data center {and space} #infrastructure, our experts will be on-site and ready to chat. 🗓️ When: October 12–15 📍 Where: San Jose Convention Center | Booth T21 Book a 1-on-1 meeting ahead of time to lock in a dedicated slot with [email protected] See you in San Jose! #OCPSummit @OpenComputePrj @OCP_Tweets #OpenComputeProject #DataCenter #DataCenterTech #Networking #HardwareInnovation #TechConferences

Welcome the E1L!

Yoav Galon and Eric Vallone are heading to Networking Field Day 41 #NFD41 with @TheFuturumGroup. As #AI workloads push hashtag#datacenter infrastructure to its limits, traditional #networking can’t keep up. Tune in as we dive into how we’re reimagining #programmableinfrastructure: 🔹 Linux-Native Data Planes: Unified control and data plane processing built for true feature velocity. 🔹 End-to-End Scale: Ultra-efficient switching and programmable DPUs from cluster fabrics to the server edge. 🔹 New Sockets: Extending open data plane architecture into tight power and form-factor budgets. And we may have some big news as well 📰 #Semiconductor #DPU #AIInfrastructure #Linux

We are excited to be on a panel next week at the @OpenComputePrj Foundation Global Summit where we will discuss: ▶️ x86 Bottlenecks: How Xsight’s E1 800G 64-core Arm #DPU powering an ultra-dense OFP storage system reduces power and rack space footprint. ▶️ 10x Density: How pairing OFP with @Hammerspace_Inc parallel file system enables AI companies, research institutions, and #neoclouds to scale exponentially. ▶️ Real-World Insights: Featuring open-standards research and deployment insights from @LosAlamosNatLab. We hope to see you there: 2026ocpglobal.fnvirtual.app/… #OpenFlashPlatform #AIStorage #ExabyteScale #DataCenter #HighPerformanceComputing