CoreWeave reported much larger quarterly revenue and more disclosed operating capacity than Nebius in the latest results covered here, but the figures are not a clean comparison of AI-cloud-only sales or directly interchangeable profitability measures. Nebius Group’s results include businesses beyond its AI cloud, while both companies sell specialist AI infrastructure and are expanding their platforms. For investors, the key distinctions are reporting scope, deployed versus contracted capacity, and execution risk; for cloud customers, the decision depends on current GPU availability, region, software, service terms, and workload performance.
How do their reported financial results compare?
The latest results covered here are for the three months ended June 30, 2026. The companies reported different measures and scopes, so the table is a guide to reported scale, not an apples-to-apples comparison of AI-cloud revenue or profitability.
| Measure | Nebius Group | CoreWeave | How to read it |
|---|---|---|---|
| Q2 2026 revenue | $582.3 million consolidated revenue, up 454% year over year. Nebius Group’s results include its AI cloud business, Avride, and TripleTen. | $2.575 billion revenue. | CoreWeave reported higher quarterly revenue. Nebius’s figure is group-wide, not AI-cloud-only revenue, so the difference does not establish the relative sales of the two cloud businesses. |
| Q2 2026 net income (loss) | Not stated in the Q2 financial figures summarized here; Nebius Group Q2 2026 results, August 12, 2026. | $626 million net loss. | CoreWeave’s reported revenue came with a GAAP net loss. Do not infer Nebius’s GAAP result from its adjusted EBITDA. |
| Q2 2026 adjusted EBITDA | $236.2 million. | $1.510 billion. | Both are company-defined non-GAAP measures. Their definitions may differ, so the figures should not be treated as directly equivalent or substituted for GAAP net income. |
| 2025 annual results | Not stated here on a comparable basis. | $5.1 billion revenue and $1.2 billion net loss for 2025. | CoreWeave’s annual figures provide historical context, but its latest quarter is more current. They do not establish a comparable annual AI-cloud-only revenue figure for Nebius. |
The most important qualification is scope. Nebius Group consolidates Avride and TripleTen alongside its AI cloud business and also holds equity stakes in other businesses. CoreWeave’s reported quarterly revenue is the company’s revenue, but the figures presented here do not provide a like-for-like segmented revenue comparison between the two AI cloud operations.
Adjusted EBITDA can help describe operating performance, but it is not a standardized substitute for GAAP earnings or cash flow. Each company’s definition and adjustments matter; compare the reconciliations in its results materials before using the metric to assess profitability.
#1 Best Overall
- NVIDIA Volta GV100 Architecture — 4,608 CUDA Cores, 640 1st-Gen Tensor Cores delivering 14 TFLOPS FP32 and 112 TFLOPS deep learning performance for AI training, inference, HPC, and scientific computing workloads
- 32GB HBM2 ECC Memory — 900 GB/s Bandwidth — High-bandwidth memory on a 4096-bit bus with ECC error correction provides the memory capacity and throughput required for the largest AI models, simulations, and datasets
- PCIe 3.0 x16 Interface — 250W TDP — Standard PCIe Gen3 connectivity with passive cooling designed for enterprise rack server deployment in HPE ProLiant, Dell PowerEdge, and Supermicro platforms with adequate chassis airflow
- NVLink — Scale to 96GB Unified Memory — Connect two V100 GPUs via NVLink at 300 GB/s bi-directional bandwidth to scale GPU memory from 32GB to 96GB for larger AI training and HPC workloads
- Multi-Precision Computing — Supports FP64 (7 TFLOPS), FP32 (14 TFLOPS), FP16 (112 TFLOPS) and INT8 precision modes for flexible deployment across training, inference, and scientific simulation workloads
What do backlog and capacity disclosures show?
| Disclosure | Nebius | CoreWeave | What it means |
|---|---|---|---|
| Backlog or commitments | A comparable backlog figure was not established in Nebius Group’s Q2 2026 release figures summarized here. | Approximately $104 billion revenue backlog at June 30, 2026. CoreWeave said the figure excluded more than $25 billion in net new commitments added in early Q3 2026. | CoreWeave’s backlog is an estimate of future revenue recognition under committed contracts, subject to delivery and availability requirements. It is not current revenue, cash on hand, or guaranteed delivery. |
| Active power | A same-date active-power figure comparable with CoreWeave’s Q2 disclosure was not established in the Nebius materials summarized here. | 1.5 GW at June 30, 2026. | Active power describes a different deployment stage from contracted power; it should not be combined with future capacity as if all were already operating. |
| Contracted power | The July 2026 partner announcement describes a model for adding partner infrastructure, but does not establish a same-date comparable active-power total. | Approximately 3.7 GW at June 30, 2026. | Contracted capacity may require time, equipment, power, and capital to become usable compute. |
| Earlier CoreWeave capacity disclosure | Not stated on a comparable basis. | Over 850 MW active and approximately 3.1 GW contracted at December 31, 2025, as reported in its 2025 Form 10-K. | This is an earlier dated snapshot, not a substitute for the June 2026 figures. |
Capacity numbers are meaningful only with their status and reporting date attached. Active, contracted, and planned power describe different stages of deployment. A cloud customer still needs to confirm whether a suitable GPU configuration is available in the required region at the time it is needed.
How do the cloud platforms differ?
Nebius: a full-stack AI cloud with new public-preview features
Nebius describes its AI cloud as a full-stack platform for work from data and model training through production deployment. Its Nebius AI Cloud 3.5 announcement on March 26, 2026, included serverless AI features in public preview, NVIDIA RTX PRO 6000 Blackwell Server Edition availability, a data-transfer service, Managed Soperator updates, Kubernetes observability improvements, billing APIs, and application-marketplace changes.
Rank #2
- Powered by Radeon AI PRO R9700 - Supercharge you workflow with the cutting-edge RDNA 4 Architecture and 2nd-gen AI Accelerators.
- 32GB GDDR6 with 256-bit memory bus - Tackle larger, more complex projects without limits.
- PCIe Gen 5 - Unlock lightning-fast data transfers with PCIe Gen 5 support.
- GIGABYTE TURBO Fan Cooling System - Indented metal cover and blower fan increase airflow intake, while the vapor chamber, all copper heat sink, and metal frame offer efficient heat dissipation. Optimized airflow design allows for easy multi-GPU scalability.
- Double Ball Bearing Fan - Delivers superior heat resistance and rotational efficiency for better performance and a longer lifespan compared to conventional sleeve fans.
Public-preview status matters: it is not the same as a generally available, production-ready service. Feature availability can also vary by location and change after an announcement. Customers should confirm the current service status, regions, and GPU inventory directly before planning a deployment.
CoreWeave: infrastructure and proprietary software for AI workloads
CoreWeave describes an AI-focused cloud stack combining infrastructure with proprietary software for model training, inference, agentic AI, agent development, and specialized workloads. Its Q2 2026 release highlighted CoreWeave Interconnect, initially with Google Cloud; SUNK Anywhere; LOTA Cross-Cloud; and agentic AI tools including ARIA and Sandboxes.
Rank #3
- [Local AI Inference & 70B Model Ready] Equipped with the AMD Ryzen 7 PRO 8845HS processor, NEXUS is engineered for heavy local AI workloads. With a full-size GPU bay, it runs 70B LLMs natively without an internet connection. Ideal for AI developers and tech enthusiasts who need private environment for coding and model testing.
- [132TB Mass Storage with ZFS Integrity] Features a hybrid storage architecture (3×NVMe + 4×3.5" HDD) supporting up to 132TB. Utilizing the enterprise-grade ZFS file system and ECC memory, it prevents data corruption and bit rot—a must-have for professional photographers and video editors safeguarding 4K/8K RAW footage.
- [OpenClaw-Driven Automation Workflow] The built-in OpenClaw execution layer allows complex automated tasks to be processed locally. Even when offline, your backup schedules and AI file organization continue seamlessly. Say goodbye to monthly cloud subscriptions and high latency.
- [Dual 10GbE & USB4 Ultra-Connectivity] Experience server-class speeds with dual 10GbE ports and a 40Gbps USB4 interface. It enables multi-user real-time collaboration on large project files directly from the NAS, ensuring zero-lag editing for creative studios and production teams.
- [Open-Source ZimaOS for Total Privacy] Running on the fully open-source ZimaOS, NEXUS ensures your data stays physically on-premise with no backdoors. It acts as a "Digital Fortress" for privacy-conscious families and small businesses who demand absolute data sovereignty.
These are company-disclosed product capabilities, not independent evidence that one platform is faster, more reliable, or more cost-effective. A useful comparison requires testing the same workload with matched GPU class, software, region, networking, storage, and service terms.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.What is different about Nebius’s partner-capacity model?
In its July 15, 2026 announcement, Nebius described a model in which infrastructure partners finance, own, and operate facilities, while Nebius provides architecture and its cloud software and services and brings capacity to customers. The company characterized it as an “asset-light” model. That description is management’s framing, not independent evidence about partner returns or the economics of individual deployments.
Rank #4
- Professional AI & Creator Workstation: AMD Radeon AI PRO R9700 GPU with 32GB GDDR6 is engineered for AI development, professional content creation, and compute-intensive workloads.
- Massive 32GB Memory Capacity: 32GB of GDDR6 memory on a 256-bit bus provides ample bandwidth for large AI models, 8K video editing, and complex 3D rendering.
- Advanced RDNA 4 with AI Accelerators: 64 Compute Units with 3rd Gen Ray Tracing and dedicated 2nd Gen AI Accelerators for groundbreaking AI performance and visual computing.
- Professional Blower Cooling: Efficient single blower design exhausts heat directly out of the chassis, ideal for multi-GPU workstation and server configurations.
- Enterprise-Grade Thermal Solution: Vapor chamber heatsink with industrial Honeywell PTM7950 thermal interface material ensures reliable cooling under sustained professional loads.
The arrangement changes who owns and operates some infrastructure; it does not eliminate the need to build and deliver usable capacity. Partners still need to secure sites, power, equipment, and operational capability, while Nebius depends on the infrastructure and delivery arrangements needed to serve customers. CoreWeave’s filings describe its own buildout alongside risks involving power access, component supply, capital requirements, customer concentration, competition, and execution. These are different operating models, but both face infrastructure-delivery and demand risks.
What should a cloud customer compare before choosing?
Neither company’s disclosed features establish a universal winner. Compare a specific workload rather than platform descriptions in isolation, and ask for current terms for the exact configuration under consideration.
- Compute: Confirm GPU model, memory configuration, quantity, availability date, and whether the capacity is on demand or reserved.
- Location: Check the region where the workload and data must run, and verify current inventory there.
- Software: Verify support for the required frameworks, orchestration, managed services, and deployment workflow. For preview features, confirm their status and suitability for production.
- Data and networking: Compare storage options, network topology and bandwidth, data-transfer charges, and any egress terms relevant to the workload.
- Commercial terms: Request comparable pricing for the same region and configuration, including reservation duration, minimum commitments, support, and transfer charges. No equivalent public price schedule for matching GPU configurations was established in the disclosures covered here.
- Validation: Run a representative workload under comparable conditions and measure the outcomes that matter to the project, such as time to train, inference throughput, or end-to-end cost.
What should an investor take away?
The disclosures show a difference in reported scale, but not enough to declare one company the better investment. CoreWeave reported substantially higher Q2 revenue and disclosed detailed active and contracted power figures. Nebius’s consolidated revenue includes multiple businesses, and its partner-capacity strategy assigns ownership and site operations to third parties for some infrastructure. Those differences make revenue scope, capacity stage, capital needs, and delivery execution central to any comparison.
For financial analysis, separate GAAP results from company-defined adjusted measures; distinguish backlog from recognized revenue; and distinguish operating capacity from commitments that may take time to deploy. CoreWeave’s backlog figure is conditional on future delivery and availability, while Nebius’s Q2 group-wide figure should not be presented as AI-cloud-only sales. The information above is company-reported and does not include independent market statistics or head-to-head cloud performance tests.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
