Nebius

Nebius is excited to announce our role as a founding consortium partner of the Physical AI Leaderboard (PhAIL) by @Positronic, a platform for training and deploying any robot AI model on any robot. PhAIL uses real hardware to evaluate vision-language- action (VLA) models on bin-to-bin order picking—a high-volume, commercially representative task. We are glad to be part of the effort, and congratulations to the Positronic team! Read our blog to learn more: https://lnkd.in/eYXjiZJP

What does it really mean to have each other’s backs at Nebius? From stepping in during high-pressure moments to celebrating wins (big and small), the support here goes beyond collaboration. It’s about trust, ownership, and knowing you’re never solving challenges alone. At Nebius, having each other’s backs isn’t just something we say, it’s something we practice every day. Sounds like your kind of team? Explore open roles at: https://lnkd.in/eQEeCkFG #LifeatNebius

We’re excited to announce the AI Discovery Award 2026 — our annual award for startups using AI to transform Life Sciences and Healthcare. The award recognizes innovative companies applying AI in areas such as biopharma, genomics, digital health, medical devices, and medical imaging. Participants can receive: • up to $100K in GPU cloud credits • introductions to leading VCs • mentorship session with Nebius C level management • brand awareness, digital promotion, and industry recognition Last year, we received 257 applications from some of the most promising startups around the world, with support from leading industry partners and top-tier venture capital firms. Check out the timeline, prize details, project criteria — and apply today: https://lnkd.in/eTdCG2Cg

AI is moving from experimentation to production. What does it take to make that shift real? Join us at the AI Infra Summit webinar to hear how teams are moving from prototypes to production systems and what infrastructure decisions matter most along the way: 📌 AI in Practice: From Innovation to Market Reality ⏰ April 16, 9:00 AM PST Hear from leaders building and deploying AI at scale: • Anna Veronika Dorogush, Recraft • Savita Raina, Nebius • Ricardo Rei, Sword Health • Matthieu Toulemont, PhotoRoom Register here: https://lnkd.in/e-miggpJ

At our Nebius․Build hackathon just before #NVIDIAGTC, Benjamin Shyong from the Meta Ray-Ban Display team joined forces with Vishal Verma and Alex Shirazi, winning on their first attempt! We invited Ben to showcase the project at the #NVIDIAGTC expo, chat with attendees and share his experience. #GTC26

On April 14, we're running a hands-on workshop where we walk through the full picture from building and running OpenClaw on Nebius, to connecting it to Token Factory for fast inference, to knowing when to use managed inference vs serverless vs AI Cloud. No fluff. Just the decisions you'll actually face when moving from a prototype to something you can rely on. You'll walk away knowing how to: → Build and run OpenClaw on Nebius → Connect to Token Factory (OpenAI-compatible inference) → Bring in custom models for specific use cases → Deploy and run your own model workloads April 14 · 6:00 PM CEST / 9:00 AM PDT Speakers: Colin Lowenberg (DevRel) & Mikhail Rozhkov (Technical PM) 🔗 https://lnkd.in/gwW5ReX9

Season 1 of the Nebius for Startups Podcast is here. This isn’t just another AI podcast—it’s a front-row seat to how AI companies are actually being built and scaled today. Bringing together perspectives from across the ecosystem, including founders, engineers, researchers, investors, and operators, this 17-episode series explores what it really takes to scale an AI startup—From infrastructure and model development, to go-to-market, fundraising, and building world-class teams, each conversation connects the dots across the full AI scaler’s journey. If you're building, investing in, or simply trying to understand the future of AI you won’t want to miss this: 🎧 Listen to Season 1 now: Website: https://lnkd.in/gsiYhmMR YouTube: https://lnkd.in/gUASXMxa Spotify: https://lnkd.in/gKeJnxvK Apple: https://lnkd.in/gqRgxSUy

Last week, teams across Nebius came together in Amsterdam and tuned in from around the world to reflect on and celebrate what we’ve been building over the past few months It has been a big period for us, from a strategic partnership with NVIDIA and a $27 billion contract with Meta, to an incredible GTC. Behind every milestone are the people and teams whose work keeps the momentum moving forward every day. Moments like this give us time to recognise progress, share what’s next, and celebrate what we’re achieving together, as we continue to build as one team. #LifeatNebius

Nebius will be at Hannover Messe in Germany — one of the world’s largest industrial trade fairs — and together with NVIDIA, we’re hosting an exclusive Executive Briefing on Robotics & Physical AI on April 22. This closed-door session is designed for senior leaders and will focus on practical, end-to-end approaches to building and scaling physical AI — from training and simulation to testing and edge deployment in real-world environments. April 22 Radisson Blu (15-minute walk from Hannover Messe) Starting at 15:00 3-hour session + demo, followed by a networking dinner If you’re attending Hannover Messe and want to explore how Physical AI is moving into real-world applications, join us. Registration is required — space is limited and subject to approval. https://lnkd.in/ebf4BN3s #HannoverMesse #PhysicalAI #Robotics #AI #IndustrialAI #NVIDIA #Nebius #ExecutiveBriefing

NVIDIA RTX PRO 6000 Blackwell Server Edition is now available on Nebius AI Cloud. While recent AI infrastructure growth has been driven by large foundation models, another wave of AI is accelerating across robotics, scientific simulations, digital twins and multimodal systems. RTX PRO 6000 expands our GPU lineup to support these workloads with strong single-precision compute, advanced visualization capabilities and architecture designed for spatial computing. With 96GB of GPU memory, teams can run substantial models on a single card, simplifying system design and reducing cross-GPU communication overhead for production inference. The GPU also supports Multi-Instance GPU (MIG), allowing one card to be partitioned into multiple isolated workloads for efficient resource utilization across teams and applications. Read the breakdown: https://lnkd.in/e-TQbPhD #AIInfrastructure #GPUCloud #SpatialAI #MachineLearning