While it may seem like just another technology partnership, it’s a game changer: Demand for advanced AI chips is skyrocketing and supply is tight. By taking early action, Nebius ensures it can meet rising business needs for high-performance computing without delays. This early adopter status catapults Nebius ahead in the AI infrastructure race, potentially generating massive revenue as companies flock to its platforms for faster AI deployments.
The Rubin Edge for Enterprise AI
Integrating the Vera Rubin NVL72 directly into Nebius AI Cloud and Token Factory isn’t just about brute force – it’s about making AI practical for businesses. Companies won’t just rent computers; they will build and deploy agentic AI systems – smart agents that handle complex tasks autonomously – from start to finish. This end-to-end installation accelerates launch and scales operations across regions.
The NVL72 racks, with 72 GPUs and 36 CPUs, are built for heavy lifting when training massive models and efficiently performing real-time inference. Think of it as a supercharged engine that reduces costs by optimizing cluster performance, making AI more affordable at scale.
What sets Nebius apart is its rare place among giants such as AWS, Google, Microsoft (MSFT), and Oracle (ORCL) as an early Rubin partner. For a specialized AI cloud like Nebius, this signals Nvidia’s confidence in its ability to quickly handle high-stakes inferences. Previous Nvidia launches show that early movers win the lion’s share of business deals in the first one to two years. Agentic AI can increase hardware usage two to three times compared to basic models, turning investments into profits faster. Additionally, the combination of hardware with managed cloud and tokenized tools reduces production timelines by months.
With coverage in the US and Europe, Nebius addresses key barriers such as data privacy legislation and low latency needs, making it easier for global companies to adopt this technology.
Still, it’s not surprising that Nebius secured early access to Vera Rubin: Nvidia was an early backer of the AI cloud startup, investing $700 million through a private placement in late 2024. Nvidia owns 1.2 million shares of NBIS stock.
Why Rubin is changing the game
Nvidia’s Vera Rubin NVL72 – unveiled at CES 2026 – represents a leap forward with up to 5x better inference performance and potentially 10x lower cost per token compared to the previous Blackwell generation for certain workloads such as Mix-of-Experts (MoE) models. It is designed for the most demanding AI tasks: advanced reasoning, long context processing, and agentic systems that solve problems autonomously in multiple steps. Features like NVLink 6 for ultra-fast GPU communications and improved reliability (cable-free design, maintenance-free downtime) ensure clusters run efficiently at scale.
For Nebius, adding Rubin alongside existing Blackwell and GB200 capabilities creates a diversified lineup tailored to different AI needs – from training giants to high-throughput inference. As an NVIDIA Exemplar Cloud Partner, Nebius settings are pre-validated against Nvidia benchmarks, giving customers immediate confidence in its performance. This is not common with newer providers; it underlines Nebius’ rapid rise as a trusted scaler in the AI factory era.
In short
The market cheered the news and boosted NBIS stock by nearly 8% after the announcement, closing the stock at over $100 per share. Over the past year, the stock has more than tripled, up 228%, reflecting strong investor confidence in its AI pivot. With Rubin deployments beginning in the second half of 2026, Nebius has a clear path to sustainable growth, capitalizing on the next wave of the AI boom.
More popular stories from Money Morning
#Heres #Nebius #Group #ensured #dominate #enterprises