Nebius Group N.V. (NBIS) Stock Price, News & Analysis

Nebius Group N.V. (NASDAQ: NBIS) is a technology company focused on building full-stack cloud infrastructure for the global artificial intelligence (AI) industry. Classified in the Internet Content & Information industry within the Communication Services sector, Nebius positions its core business around an AI-native cloud platform designed specifically for intensive AI workloads. The company is headquartered in Amsterdam and is listed on Nasdaq, and it reports having a global footprint with research and development hubs across Europe, North America, and Israel.

Core business: Nebius AI Cloud

The company describes its primary business as an AI-native cloud platform built from the ground up for demanding AI training and inference. According to multiple company disclosures and press releases, Nebius AI Cloud combines proprietary software and hardware designed in-house to give AI builders access to compute, storage, managed services, and tools to build, tune, and run their models and applications. The platform is oriented toward intensive AI workloads and is used for large-scale model training, inference, and production deployment.

Nebius emphasizes that its infrastructure is full-stack, meaning it integrates purposefully designed and tuned proprietary software with hardware that the company designs in-house. This approach is intended to support high-performance AI workloads and to provide the building blocks needed across the AI lifecycle, from experimentation to production deployment.

AI infrastructure focus and global footprint

Across its news releases and regulatory filings, Nebius consistently describes itself as an AI infrastructure company serving the high-growth global AI industry. The company highlights that it operates R&D hubs in Europe, North America, and Israel, and that its AI cloud infrastructure is deployed across regions including the US, Europe, the UK, and the Middle East. This geographic spread is presented as a way to support regional data residency options and regulatory requirements for AI workloads.

Nebius also notes that it has been validated as an NVIDIA Exemplar Cloud and a Reference Platform NVIDIA Cloud Partner in various announcements, reflecting that its infrastructure is benchmarked against NVIDIA reference architectures and performance standards for large-scale AI training and inference. These validations are cited in relation to MLPerf Training and MLPerf Inference benchmark results on NVIDIA GPU systems such as GB200 NVL72, HGX B200, and other Blackwell Ultra configurations.

Platform evolution: Nebius AI Cloud 3.0 “Aether” and 3.1

Nebius has publicly detailed successive versions of its AI cloud platform. The Nebius AI Cloud 3.0 “Aether” release is described as delivering enterprise-grade security, compliance, and governance features for AI deployment at scale. The company highlights independently validated certifications, including SOC 2 Type II with a HIPAA section and ISO 27001, as well as alignment with frameworks such as NIS2, DORA, ISO 27032, ISO 27701, and ISO 27799. Aether also introduces enhanced observability, granular Identity and Access Management (IAM), updated user interfaces aligned with AI/ML workflows, and automated health monitoring with self-healing infrastructure.

The subsequent Nebius AI Cloud 3.1 release builds on Aether by adding next-generation NVIDIA Blackwell Ultra compute, including NVIDIA GB300 NVL72 and HGX B300 systems, and by deploying these systems on 800 Gbps NVIDIA Quantum-X800 InfiniBand interconnect. Nebius states that it is the first cloud in Europe to operate both GB300 NVL72 and HGX B300 systems in production and the first provider globally to run production GB300 NVL72 systems on this 800 Gbps fabric. Version 3.1 also introduces Capacity Blocks and a real-time Capacity Dashboard for visibility into GPU availability, as well as project-level quotas and lifecycle object storage rules for more granular resource and cost control.

Nebius Token Factory and AI inference

Beyond its core infrastructure, Nebius has launched Nebius Token Factory, which it describes as a production inference platform built on Nebius AI Cloud. Token Factory is presented as an enterprise-ready platform that brings together high-performance inference, post-training, and fine-grained access management into a single governed environment. According to company communications, it supports major open-source models such as Llama, Qwen, and other open models, and allows customers to host their own models.

The company states that Nebius Token Factory is optimized for efficiency, with features such as sub-second latency, autoscaling throughput, and a stated 99.9% uptime target for workloads that can reach hundreds of millions of requests per minute. It includes capabilities for fine-tuning and distillation, integrated deployment to production endpoints, and governance features such as Teams and Access Management, Single Sign-On (SSO), unified billing, and audit-friendly workspaces. Token Factory is also described as the next evolution of Nebius AI Studio, redesigned for enterprise readiness and full model-lifecycle management.

Partnerships and ecosystem

Nebius highlights several ecosystem relationships in its announcements. It notes a partnership with TD SYNNEX, under which TD SYNNEX launched an AI Infrastructure-as-a-Service offering in North America built on Nebius AI Cloud. This collaboration is described as enabling TD SYNNEX partners to access Nebius’s AI-native cloud infrastructure, including NVIDIA GPU capacity, with transparent pricing and options for reserved instances.

The company also references collaborations with platforms such as Anyscale, SkyPilot, Outerbounds, Saturn Cloud, Lightning AI, and SkyPilot Server, which are integrated with Nebius AI Cloud to support AI developer workflows. In the context of Nebius Token Factory, Nebius cites usage by organizations such as Prosus and Higgsfield AI, and mentions collaboration with Hugging Face to improve access and scalability for developers building on large open-source models.

Additional businesses within Nebius Group

In its financial reporting and group-level descriptions, Nebius Group N.V. explains that it has additional businesses operating under their own brands alongside its core AI cloud platform. These include:

• Avride – described as one of the most experienced teams developing autonomous driving technology for self-driving cars and delivery robots. Avride is characterized as a US-based developer of autonomous vehicles and delivery robots headquartered in Austin, Texas, with additional R&D hubs around the world.
• TripleTen – described as an edtech player in the US and certain other markets, focused on re-skilling people for careers in technology.

Nebius Group also reports that it holds equity stakes in other businesses, including ClickHouse and Toloka. Following an investment transaction in Toloka, Nebius ceased to hold majority voting power in Toloka and now reports its stake as an equity method investment, with Toloka’s prior-period results reclassified to discontinued operations.

Capital markets activity and major commercial agreements

The company’s SEC filings and press releases describe several capital markets and commercial developments. Nebius Group completed an underwritten public offering of Class A ordinary shares and concurrent private offerings of convertible senior notes due 2030 and 2032, with related indentures filed as exhibits to a Form 6-K. A subsequent press release notes that underwriters exercised an option to purchase additional Class A shares, and that the company’s registration statement on Form F-3ASR is automatically effective.

Nebius has also disclosed a commercial agreement with Meta Platforms, Inc., under which a wholly owned subsidiary, Nebius, Inc., will provide Meta access to two dedicated GPU infrastructure capacity clusters over a five-year term. According to the Form 6-K describing this agreement, the order has a total contract value of approximately $2.9 billion, with GPU services deployed in tranches and accompanied by associated storage and connectivity services. The agreement includes customary provisions such as service levels, termination rights, discounted fees for late delivery, representations and warranties, indemnities, and limitations of liability.

Technology positioning and sector classification

Across its disclosures, Nebius characterizes itself as building infrastructure for the high-growth global AI industry. It emphasizes AI-native cloud capabilities, full-stack infrastructure, and proprietary software and hardware as defining elements of its offering. The company’s classification in the Internet Content & Information industry and Communication Services sector aligns with its focus on cloud-based services and AI workloads delivered over networked infrastructure.

Nebius’s communications highlight that its platform is used across the AI lifecycle, from model training to inference and production deployment, and that it targets use cases in sectors such as healthcare, financial services, government, and other regulated industries, particularly in the context of its security and compliance certifications and governance features.

Corporate structure and reporting

Nebius Group N.V. files reports with the U.S. Securities and Exchange Commission as a foreign private issuer, using forms such as Form 20-F and Form 6-K. Its Form 6-K filings incorporate operating and financial reviews, unaudited condensed consolidated financial statements, equity distribution agreements, and press releases related to capital markets activity and commercial agreements. The company also describes the use of non-GAAP financial measures such as Adjusted EBITDA and Adjusted net income (loss), and provides reconciliations to the most directly comparable U.S. GAAP measures in its financial reporting.