Nebius monthly digest: April 2025
Nebius monthly digest: April 2025
This month, we revealed some of the partnerships and collaborations we’ve been engaged in, such as with Meta and DDN. We also showcased work done by our customers like vLLM and Brave, as well as participants of our Research credits program — including Stanford, USC and others.
Revealing new partnerships and collaborations
- We are proud to be a partner of Meta, featured in the keynote at LlamaCon 2025 by Meta’s Chief Product Officer Chris Cox.
-pica-(1)-(1).jpg?cache-buster=2025-05-07T10:04:58.090Z)
- We sought a data intelligence platform to empower enterprises. DDN emerged as a clear choice — now its data architecture powers our cloud.
Showcasing customer success again, also in research
-
Using Nebius’ infrastructure, our customer vLLM — a leading open-source LLM inference framework — is testing, benchmarking and optimizing their inference capabilities in different conditions, enabling high-performance, low-cost model serving in production environments.
-
Brave Software, with over 80M users, develops a fast, privacy-focused browser and an independent search engine. Its Al-powered feature, Answer with AI, provides real-time, privacy-centric summaries for 1.3B queries per month, generating 11M+ answers daily — with Nebius behind the scenes.
-
Quantori’s cheminformatics department recently embarked on a research initiative to develop a molecular generation pipeline. After 1,500 training epochs over the dataset of 1.6M molecules, the model successfully generated chemically sound molecules that closely resemble real ones in shape.
-
It’s been just a couple of months since we launched the Research credits program — and we’re already thrilled to see participants using Nebius’ compute for breakthrough studies. We gathered the details on what some of them — including Stanford, USC and others — have been working on this year.
Job of the month: AI/ML Solutions Architect
Would you like to tackle complex challenges in distributed training and large-scale inference deployments? Collaborate with frontier AI labs around the world to help them build highly efficient pipelines using a state-of-the-art stack?
If so, the role of AI/ML Specialist Solutions Architect might be the perfect fit for you. By combining domain expertise in ML engineering, MLOps and research with a knack for solving practical problems, this position offers a chance to work at the bleeding edge of the industry.
Supercharging startups: Nebius accelerates AI-native innovation with NVIDIA
Nebius AI Cloud is collaborating with NVIDIA Inception, a leading program for AI startups with over 22,000 members, to accelerate AI startup innovation.
Our AI Lift program for startups offers eligible NVIDIA Inception members up to $150,000 in cloud credits for Nebius’ cutting-edge AI cloud infrastructure, as well as access to AI technical expertise and support, and a range of other exclusive benefits to accelerate customer success.
There’s always something to share in technical articles and platform news
In a recent research blog post, our AI R&D team explores how RL-finetuned reasoning language models can be used as a better alternative to regression-based critics when performing parallel trajectory search in test time, and how to train such models to prioritize precision over recall.
-
Meet Audit Logs in Nebius AI Cloud. This new feature enables you to improve your accountability, security and to ensure compliance with regulatory requirements.
-
You can now choose Nebius in dstack, an open-source container orchestration platform, and start managing dev environments, executing training jobs and deploying models on our AI infrastructure.
-
Moving large datasets between buckets is often challenging — especially across accounts or clouds. Our Solutions Architect Alex Kim shares a fast, fully open-source workaround using SkyPilot, s5cmd and our platform.
-
If you work with personal data and need to store it in the cloud, you have to choose a cloud provider that offers appropriate storage. If you’re wondering whether our cloud can store and process personal user data, the answer is yes. Nebius ensures the secure storage of personal data, in full compliance with applicable regulations.
You guessed it: tech docs have also been updated
-
The new Audit Logs service docs cover everything you need to know about events that happen to your Nebius AI Cloud resources: learn how to view, understand and interpret these events to improve your infrastructure’s security
-
Since Virtual Private Cloud is now its own service, it got its own documentation as well. The articles will help you isolate networks of resources from one another, manage public and private IP addresses and stay within the service’s quotas.
-
If you want to neatly organize your applications, workloads and experiments, take advantage of cloud projects. Our new guide provides simple instructions for creating and managing projects.
-
Our articles about Slurm and Soperator continue growing in number and scope. Soperator supports running containers with Apptainer, a runtime designed for HPC and scientific workloads. And if you want to dive deeper into Soperator, this article on cluster architecture provides a detailed yet clear overview of cluster components and their interactions.
-
Another fine addition to our collection of tutorials: a comprehensive guide on spinning up Managed Kubernetes clusters with Run:ai, an AI-tailored platform for dynamic GPU orchestration.
Explore Nebius AI Studio
See also
Nebius opens pre-orders for NVIDIA Blackwell GPU-powered clusters
Nebius opens pre-orders for NVIDIA Blackwell GPU-powered clusters
We are now accepting pre-orders for NVIDIA GB200 NVL72 and NVIDIA HGX B200 clusters to be deployed in our data centers in the United States and Finland from early 2025. Based on NVIDIA Blackwell, the architecture to power a new industrial revolution of generative AI, these new clusters deliver a massive leap forward over existing solutions.