Activity
6K followers
-
Matt Saxe shared thisWord is quickly spreading across the AI Infra ecosystem that the most demanding, innovative AI Clouds and Inference providers are building with vCluster. Why? We enable teams to productize raw GPU capacity end to end: vMetal rapidly provisions and orchestrates bare-metal servers, while vCluster layers isolated, self-service Kubernetes tenant clusters on top—delivering a dedicated-cluster experience from a pooled fleet, faster time to revenue, stronger tenant controls, and materially higher utilization. And at a time in the industry full of bold claims, we see each of our wins as an opportunity to deliver with a world class team of customer engineers and FDEs to ensure success. Come see us at the AI Infra Summit in Santa Clara this week and join us Weds with our partner Netris for a great evening. https://lnkd.in/grx2p6rz
-
Matt Saxe shared thisFirst time in Amsterdam. What an incredible city! I look forward to catching up with everyone at Kubecon. Don't miss our event Tuesday night to hear from Felipe Del Piccolo from JPMC, Everett Lacey from NVIDIA, and our own Jonathan Fullam on running Kubernetes at Enterprise scale in the age of AI. Register at the link below. https://luma.com/28cq1bmbRunning Kubernetes at Enterprise Scale in the Age of AI · LumaRunning Kubernetes at Enterprise Scale in the Age of AI · Luma
-
Matt Saxe shared thisGreat summary from vCluster Head of Customer Engineering Jonathan Fullam on our conversations at NVIDIA GTC with key customers and partners. If you missed the talk with Adam Tetelman, Lukas Gentele, and Alex Saroyan on how to build a software defined multi-tenant NVLinked cluster, the recording is live once you login to the site (will post in comments). Now onto Kubecon Amsterdam!Matt Saxe shared thisJust wrapped an incredible week at NVIDIA GTC and I left more convinced than ever that the AI infrastructure race is real, urgent, and vCluster is uniquely positioned to continue delivering value to our customers. 14 customer meetings, no high-level meet and greets, but real conversations with teams who are already building, and feeling the pressure to show ROI on their AI infrastructure now. The message was consistent across all of them: provisioning bare metal Kubernetes fast enough to utilize GPU clusters is the bottleneck. And the moment we showed how vCluster collapses that complexity, giving platform teams and AI workloads full isolation, multi-tenancy, and near-instant K8s environments without the sprawl, teams wanted to get started. Our booth was also a live proof point. Platform engineers, software developers, AI specialists, and executives lined up for discussions and demos. Many weren't just curious but had existing challenges we could address. The conversation wasn't "what is this?" It was "how fast can we move?" What impressed me most wasn't the traffic. It was our team. Watching our marketing, sales, and customer engineering colleagues move fluidly between deep Kubernetes / vCluster technology and problem solving in the same breath. What a world-class team! Happy to connect if you're thinking through your AI/Developer/K8s platform strategy or if interested in joining the team. #NVIDIAGTC #vCluster #AIFactory #Kubernetes #PlatformEngineering #NeoCloud #AIInfrastructure
-
Matt Saxe shared thisOur collaboration with the NVIDIA team really took off around GTC Paris last year. Fast forward 9 months and its amazing to see how quickly vCluster has become the standard for Kubernetes in AI Infrastructure. From our work with leading Neoclouds to AI factory deployments across every industry, our team is focused on delivering for our customers. I look forward to seeing everyone in San Jose this week. RSVP for a great Happy Hour event we are hosting with our partners at Netris on Tuesday https://luma.com/yq98vkek Dont miss the panel discussion with our CEO Lukas Gentele, Adam Tetelman from NVIDIA, and Netris CEO Alex Saroyan Stop by booth #206 to say Hello to our team https://lnkd.in/eQD_47Sr
-
Matt Saxe shared thisWe just wrapped a record year of growth at vCluster Labs. A big thank you to our customers, partners, and team for their support. The latest CNCF survey highlights what we see every day: "Kubernetes has become the common operating layer that organizations rely on to run modern, production-grade systems and bring AI into real-world environments." The teams we are working with are building for the future. They need to accelerate development across diverse workloads (varying isolation and security needs across CPUs and GPUs), improve utilization, control costs, and maintain portability across private and public clouds. These are not easy problems to solve. vCluster delivers with a breakthrough architecture that unlocks business value by empowering teams to truly “own” their compute strategy — up to 75% faster provisioning, 25–50% cloud cost reduction, and support for any tenancy model. Metrics and the DNA behind our success: • 290% YoY growth and 174% NRR, driven by a focus on rapid time-to-value and long-term partnership • Battle-tested, highly scalable deployments across fast-growing startups, large enterprises (global fin serv, healthcare), federal customers, and leading neo cloud providers • Close collaboration with the NVIDIA DGX, Run:AI, and Netris teams to deliver a best-in-class Kubernetes experience for the NVIDIA ecosystem (see you at GTC San Jose in March) • A leadership team committed to a “one team” mindset across engineering and GTM, grounded in being customer-obsessed • An R&D and engineering team that innovates at high speed while maintaining quality • A GTM team of cloud-native infra experts who act as strategic partners- diagnosing needs, recommending the right solution, and delivering (having fun doing it too) Proud of of this team. Excited for an even bigger 2026. Virtualize clusters not machines! Joyce Zhang Diaa Khalil John Karant Nick Kirchner Kurt Madel Chris Kuech Benjamin Walker Brandon J. John Barfield Bill DeCoste Andrew Brooks Laurie Maclachlan Oleg Matskiv Declan McAleese Stephen Smith Simon Wolfenden Andy Griffin Roy Schultheiß Daniel Welch Lukas Gentele Fabian Kramm Denise Schannon Paul Hinz Daniel Thiry Brendan Orf Tom Brightbill Erin Palestini
-
Matt Saxe shared thisInformative summary from Diaa Khalil on the key takeaways from our panel event last week at Kubecon. Special evening with a group of folks solving hard problems to deliver on the business value of AI.Matt Saxe shared thisLast night I had the privilege of moderating a panel on "The Future of AI and Kubernetes" at KubeCon Atlanta. With AI workloads exploding in complexity and scale, we dove deep into how they're reshaping Kubernetes ecosystems. Our panelists, included Joshua Bucknor from JPMorganChase, Everett Lacey from the NVIDIA DGX SuperPOD team and Fabian Kramm CTO and Co-Founder of vCluster, brought invaluable insights. A quick recap of our conversation: 🔹 Challenging Kubernetes Norms: We explored surprising ways AI workloads upend traditional assumptions, like the need for dynamic resource allocation beyond static pods, handling massive data pipelines that strain etcd, and ensuring fault-tolerant orchestration for long-running training jobs. 🔹 GPU Scheduling Pain Points & Innovations: The conversation highlighted current bottlenecks in GPU workloads on Kubernetes, such as inefficient bin-packing and NUMA-aware scheduling. Exciting innovations like DRA and it's effect on optimizing sparse GPU resources. 🔹 Resource Sharing & Multi-Tenancy: With AI chips being premium assets, organizations are leaning into namespace isolation, quota enforcement via Kubernetes Resource Quotas, and custom schedulers. Joshua shared how JPMC tackles multi-tenancy in large-scale clusters to balance security and efficiency. Everett discussed customer demands for Kubernetes atop DGX SuperPODs, including hybrid multi-tenant setups. 🔹 Fractioning GPUs & The VM vs. Bare Metal Debate: We debated the shift from VMs for traditional workloads to bare-metal for GPUs, driven by performance needs in AI. 🔹 Advice for Newcomers: If you're just starting with AI on Kubernetes, the consensus? Prioritize observability early, integrate Prometheus and Grafana to monitor GPU utilization and avoid under-optimized clusters. Grateful for the engaging dialogue and fresh perspectives! If you're running AI on K8s, what's your biggest challenge right now? Drop a comment below, let's connect and discuss. #AI #Kubernetes #GPUs #CloudComputing #TechLeadership
-
Matt Saxe posted thisDay 1 at Kubecon in Atlanta. Come by our booth and register to attend the panel on Weds night together with our team, JPMC, and NVIDIA. We just had the biggest release since open source vCluster was started and the reference guide we just published with the NVIDIA DGX team is a must read. Some of the developments you will learn about: vCluster Private Nodes and Karpenter-based vCluster Auto Nodes, to enable virtual clusters to dynamically autoscale GPU and CPU capacity across clouds, data centers, and bare metal environments. vCluster VPN, a Tailscale-based overlay virtual private network. Netris network isolation controller, for network isolation, giving each tenant its own dedicated network path. VNode Runtime to provide a container sandbox that helps prevent container breakouts.
-
Matt Saxe shared thisKubecon and Atlanta friends. We are hosting a happy hour on Weds, Nov 12th at Top Draft Sports Lounge which is a short walk from the event. Great opportunity to hear from leading platform engineers at JPMC and NVIDIA on how a flexible approach to Tenancy for Kubernetes and AI Infra is unlocking value. Register below. https://luma.com/xwbxheciThe Future of AI and Kubernetes - Happy Hour and Fireside Chat · LumaThe Future of AI and Kubernetes - Happy Hour and Fireside Chat · Luma
-
Matt Saxe shared thisThe rise of the tech industry in New York has always had deep roots in financial services. The competition and focus on excellence has sparked innovation and attracted talent from all over the world. I am fortunate to have been a part of the journey from the early days of electronic trading (catalyst for many breakthroughs in high performance computing) at 29West to AppDynamics and the rise of open source adoption Grafana Labs. Proud to have vCluster Labs sponsor the Open Source Finance Forum this week. If you are attending, don't miss this talk with our friends at J.P. Morgan and our CEO Lukas Gentele. Great opportunity to learn more about how vCluster unlocks major value in your compute strategy. https://lnkd.in/emP9ujZx
Matt Saxe commented on a post
1d
Lets go Nicole B.!
-
Matt Saxe liked thisIncredibly proud of this team. 🙌 $240.9M in new contracted value!!! What makes me proudest? Both of these customers were already working with us, and they chose to come back and go bigger. Repeat commitments like that don't come from a pitch deck. They come from a team that delivers, day in and day out. To everyone across sales, capital markets, engineering, operations and finance who made this happen: thank you. This one is yours! #AI #AIInference #NVIDIA #NeoCloud #QumulusAI #HyperSpeed QumulusAIMatt Saxe liked this"We're excited to keep growing alongside AI builders that are changing what this industry can do, and it says a lot that both of these customers came back to commit to three more years with us," said Mike Maniscalco, CEO of QumulusAI. "Inference platforms are planning capacity years out now, and they are backing those plans with cash. That lets us place GPUs against signed contracts rather than against forecasts." Read More >> https://lnkd.in/e-sQ5BHR #hyperspeed #AIinfrastructure
-
Matt Saxe liked thisMatt Saxe liked thisProud. Grateful. Inspired. I was honored to represent vCluster as a Platinum Sponsor of the Love Never Fails “Swing Against Human Trafficking” Golf Tournament. And to the great team at DDN of Andy LeSage, Timothy Termini and Stephen Burke for joining me in this amazing event, and for the laughs and your generosity! This was much bigger than golf. Love Never Fails helps survivors of human trafficking rebuild their lives — providing social and emotional support, restoring confidence, and creating pathways to economic independence. Through technology education and workforce development, individuals can build skills in areas such as cybersecurity and IT and turn opportunity into a career. That mission matters deeply to me. Technology is powerful, but its greatest impact is when we use it to change lives. I’m incredibly thankful to vCluster's Lukas Gentele (CEO) for believing in and supporting this cause, to Alvin Dacosta for making me aware of this cause almost 3 years ago, to the other sponsors who stepped forward including NVIDIA, Accenture, EY, Deloitte, Hitachi and especially to the Love Never Fails team, volunteers and supporters doing this work every day. THANK YOU! A beautiful day, an important cause, and a reminder of what we can accomplish when community, compassion and opportunity come together. Proud to have vCluster be a small part of something that matters this much. Thank you @ #LoveNeverFails #EndHumanTrafficking #vCluster #TechnologyForGood #Cybersecurity #AI #Community #GiveBack
-
Matt Saxe liked thisMatt Saxe liked thisI don’t post often, but this was a moment worth celebrating. I had the absolute privilege to be present as my dad, Steve Olson, accepted the Lifetime Achievement Award at the Orthopedic Trauma Association. Decades of work, research, and dedication to excellence led to his peers nominating him, and the association agreed. Dad, despite your tireless work ethic and demanding job, you made sure to show up for every important event in my life. It’s a privilege to be able to show up for yours. I’m so proud of you, and I love you. Duke Orthopaedic Surgery, ORTHOPEDIC TRAUMA ASSOCIATION
-
Matt Saxe liked thisNice end to end tutorial bringing up Dynamo on Nebius!Matt Saxe liked this🎯𝗡𝗩𝗜𝗗𝗜𝗔 𝗗𝘆𝗻𝗮𝗺𝗼 𝗼𝗻 Nebius 𝗺𝗮𝗻𝗮𝗴𝗲𝗱 𝗞𝟴𝘀 𝘄𝗶𝘁𝗵 𝘁𝗲𝗿𝗿𝗮𝗳𝗼𝗿𝗺🧑🏼🚀💚 👉 Read here: https://lnkd.in/gsuh3MDC by Kosseila H. New to NVIDIA #Dynamo? Here's what it's made of: an operator, a DynamoGraphDeployment, a 𝗙𝗿𝗼𝗻𝘁𝗲𝗻𝗱, 𝗮𝗻𝗱 𝘄𝗼𝗿𝗸𝗲𝗿𝘀(vllm). This stack wires the leanest topology that actually ships them, two models on one GPU behind a single Ingress, 𝗞𝗩-𝗮𝘄𝗮𝗿𝗲 routing on tap, on Nebius managed Kubernetes in 𝗼𝗻𝗲 𝘁𝗲𝗿𝗿𝗮𝗳𝗼𝗿𝗺 𝗮𝗽𝗽𝗹𝘆 🔥 🔎 𝗪𝗵𝗮𝘁'𝘀 𝗶𝗻 𝘁𝗵𝗲 𝘀𝘁𝗮𝗰𝗸? ✅ Architecture overview and why we chose our topology (single frontend) ✅ Two-layer deploy: Operator via helm, and Inference stack with manifests ✅ Multi-model per GPU: two models behind a single LLM endpoint ✅ Secure endpoints: HTTPS LLM serving (Nebius LB + Let's Encrypt certs) ✅ Full observability: Prometheus + Grafana + Dynamo's own dashboards ✅ KV-cache-aware routing: shared-prefix requests on cache-warm replica 🔥 𝗙𝘂𝗹𝗹 𝗖𝗼𝗱𝗲 𝗼𝗻 𝗚𝗶𝘁𝗛𝘂𝗯: https://lnkd.in/gBU5fpT3
-
Matt Saxe liked thisMatt Saxe liked thisGot a chance to meet up with some of the Pivotal team this week. It was incredible to spend time with such an amazing group, and it reminded me of how much that era shaped my thinking on teams, performance, and values. In a formative book for me, "The Fifth Discipline," Peter Senge explains that most of us, if we're lucky, get the opportunity to be part of a truly great team at some point in our careers. He describes these as groups where members deeply trust one another, individual strengths naturally compensate for weaknesses, and everyone is united by a shared vision far larger than any single person's ambition. We certainly weren't perfect, but I always think of my time at Pivotal in this way. I'm incredibly grateful to everyone who played a part in that experience! What team in your career was your "once-in-a-career" experience and why?
-
Matt Saxe liked thisMatt Saxe liked thisI had a chance to connect with The Information on how enterprises are approaching AI purchasing decisions in 2026, and it got me thinking about what separates a good vendor conversation from a forgettable one. At FICO, we don't evaluate AI capabilities in isolation. We ask the same questions we'd ask of any core system: does it fit our existing workflows, does it hold up on security and governance, and can we measure the outcome in weeks, not quarters. AI has moved fast enough that novelty stopped being a selling point a while ago. What matters now is whether a vendor can show how their AI capabilities can change how work gets done for our teams. Thanks to Laura Bratton for including FICO's perspective in the piece. Full story here: https://brnw.ch/21x5Wy5 #AI #CIO #EnterpriseAISoftware Firms Discount AI to Keep Customers From Anthropic, OpenAISoftware Firms Discount AI to Keep Customers From Anthropic, OpenAI
-
Matt Saxe liked thisMatt Saxe liked thisBig Tech has an army of lobbyists and endless cash. They don’t stand a chance against parents protecting their kids. Two and a half years ago, Mothers Against Media Addiction (MAMA) didn't exist. Today we have 55 chapters in 27 states and have delivered major wins for kids and families. In June, moms from across the country showed up on Capitol Hill and told Congress to choose: kids or Big Tech. In July, New Jersey passed the Kids Code Act. In August, a New Mexico court handed down a landmark ruling against Meta. In September, Governor Newsom signed a package of bills in California protecting kids from addictive social media and AI. And at the end of the summer Meta agreed to a $17 billion settlement with 52 U.S. states, territories, and the District of Columbia. It's the largest state consumer protection settlement outside the tobacco industry in the 1990s. None of this happened because tech companies volunteered to do the right thing. It happened because parents organized, testified, called their legislators, and refused to go away. I'm on MAMA's board and co-chairing our 3rd Annual Fall Benefit. Tuesday, October 20 48 Wall Street, NYC Cocktails at 6, program at 7 Speakers include Dana Suskind (author of Human Raised), Larz May of #HalfTheStory, Assembly member Alex Bores, actor and author Sheetal Sheth, and MAMA founder Julie Scelfo. Get your tickets here: https://lnkd.in/g4wxidNf Can't make it? You can sponsor a teacher's seat for $100 or donate on the same page. If you are in NYC, I'd love to see you there.
-
Matt Saxe liked thisMatt Saxe liked thisYesterday marked 7 years aboard this rocket ship called Grafana Labs 🎉🚀 Not going to write a long, sappy LinkedIn post. I’m just grateful to work with so many great colleagues and customers, spend my days on things I really care about, and play a small part in Grafana Labs’ success. And a particular highlight: morning runs with colleagues around the world during team offsites and customer visits. #letsgoooo
Other similar profiles
Explore more posts
Explore top content on LinkedIn
Find curated posts and insights for relevant topics all in one place.
View top content