- Combines Eigen AI’s industry-leading inference stack with Nebius’ global capabilities
- Collaboratively optimized endpoint achieved top ranking in artificial analysis across multiple models
- Eigen AI’s founding team, including researchers from the MIT HAN Lab, establishes Nebius’ Bay Area engineering and research hub
Amsterdam, May 1, 2026 — Nebius (NASDAQ: NBIS), an AI cloud company, today announced an agreement to acquire Eigen AI, a leading inference and model optimization company.
The acquisition strengthens Nebius Token Factory as a frontier managed inference platform for production AI, combining a battle-tested optimization stack with Nebius’ global computing capabilities and AI cloud platform, adding elite inference research talent to the company’s established in-house AI R&D capabilities.
Once finished, Eigen AI’s inference and post-training optimization layers will be integrated directly into Nebius Token Factory, providing enterprise-grade autoscaling endpoints and fine-tuning pipelines across all major open source models. Together, the companies are already delivering optimized implementations of leading open source models that rank among the fastest in artificial analysis.
The acquisition will also accelerate Nebius’ expansion in the United States. Eigen AI’s founding team, researchers who developed optimization techniques and tools used in the industry, will join Nebius to establish the Nebius engineering and research hub in the San Francisco Bay Area.
Roman Chernin, co-founder and chief business officer of Nebius, said: Said:
“We operate in a capacity-starved world where AI builders require optimized inference and infrastructure scale. The integration of Eigen AI’s optimization capabilities and founding team establishes Nebius Token Factory at the forefront of inference, delivering market-leading model performance and unit economics with massive compute power to our customers.”
Eigen AI’s founding team brings deep expertise from research that shapes how inference is deployed in industry today. Co-founders Ryan Hanrui Wang and Wei-Chen Wang are alumni of MIT’s HAN Lab, led by Professor Song Han, a pioneering researcher in AI computing and model efficiency.
Ryan’s pioneering sparse attention (SpAtten) work has been the most cited HPCA paper since 2020, while Weichen won the MLSys 2024 Best Paper Award for his activation-aware weight quantization (AWQ) quantization, which is now the standard for 4-bit models used in production environments. Co-founder Di Jin is an MIT CSAIL PhD who brings deep expertise in post-training and large-scale model tuning, contributing to Meta’s Llama 3 and Llama 4 post-training, and co-authoring the CGPO RLHF framework.
Ryan Hanrui Wang, Eigen AI Co-Founder and CEO; Said:
“We’re proud to join Nebius and work with the Token Factory team to push the boundaries of inference performance. Nebius has built a world-class AI cloud with a deep engineering culture that perfectly aligns with ours. Together, we’re removing the friction from customizing and deploying AI models, ensuring developers can reliably run models in production without managing the underlying infrastructure.”
Inference is currently the fastest growing area of AI and is predicted to account for approximately two-thirds of computing demand this year. Correspondingly, the use of open source models is also increasing. As more workloads move into production, the system optimization layer is becoming a critical infrastructure.
Efficiently running inference in production is inherently complex and requires deep expertise across the entire execution stack, from how to represent models to how GPU kernels run them to how to schedule workloads in real-time.
Open source models typically ship unoptimized, and new architectures such as Mixed Expertise (MoE), Compressed Sparse Attention (CSA), Inference, and Long Context Models pose additional challenges around memory, routing, and compute efficiency. Most teams don’t have the ability to solve these problems internally.
Eigen AI addresses this challenge with a full-stack optimization approach that spans the entire model lifecycle. From post-training and fine-tuning to production inference optimization, it runs across all major open source models in demand in production, including GPT-OSS, Gemma, Qwen, Llama, Nemotron, DeepSeek, GLM, Kimi, and MiniMax.
Nebius eliminates this bottleneck throughout the lifecycle by integrating Eigen AI’s optimization layer directly into the Nebius Token Factory. The system, model, and kernel-level techniques developed by the Eigen team are designed to extract significantly better performance from the hardware without incurring additional engineering overhead, resulting in higher throughput and lower cost per inference.
As a result, Nebius Token Factory customers benefit from faster time to production, significantly improved unit economics, and the ability to deploy new models more quickly. Existing Eigen AI customers will have access to Nebius’ global AI infrastructure and platform capabilities.
The transaction consideration will be paid in a combination of cash and Nebius Class A shares, with a total value at the time of signing based on Nebius’ 30-day weighted average stock price (adjusted) of approximately $643 million. The transaction is expected to close in the coming weeks, subject to certain customary conditions, including antitrust clearance.
About NeviusAbout Nevius
Nebius, the AI cloud company, is building a full-stack platform for developers and enterprises to own the future of AI, from data and model training to production deployment. Nebius was founded on deep in-house technical expertise and operates at scale with a rapidly expanding footprint around the world, serving startups and enterprises building AI products, agents, and services around the world.
Nebius is listed on Nasdaq (NASDAQ: NBIS) and is headquartered in Amsterdam.
For more information, please visit www.nebius.com.
contact addresscontact address
Media relations: media@nebius.com
Investor information: askIR@nebius.com
DisclaimerDisclaimer
Forward-looking statements
This press release contains forward-looking statements within the meaning of the Private Securities Litigation Reform Act of 1995 that involve risks and uncertainties. All statements other than statements of historical fact contained in this press release are forward-looking statements, including, without limitation, statements regarding our ability to complete the acquisition of Eigen AI, integrate the Eigen AI team, and achieve anticipated synergies and other benefits. The words “anticipate,” “believe,” “continue,” “estimate,” “expect,” “lead,” “intend,” “likely,” “may,” “will” and similar expressions and the negative thereof are intended to identify forward-looking statements.
These forward-looking statements involve risks, uncertainties and assumptions, some of which are beyond our control. Actual results may differ materially from those predicted or implied by such statements, and the results reported by the Company should not be considered as an indication of future performance. Potential risks and uncertainties that could cause actual results to differ from those anticipated or implied by such statements include, among others: risks related to acquisitions and business and team integration; Market, macroeconomic and geopolitical conditions. Technology development. our ability to attract and retain customers; our ability to secure additional capital to enable the growth of our business; Also included are the risks and uncertainties associated with the Company’s continuing operations included under the captions “Risk Factors” and “Operating and Financial Review and Prospects” in our Annual Report on Form 20-F for the year ended December 31, 2025, filed with the Securities and Exchange Commission on April 30, 2026.
All information in this press release is current as of the date of publication (unless otherwise noted). Except as required by law, we undertake no obligation to publicly update or revise any forward-looking statements, whether as a result of new information, future events or otherwise, after the date on which they are made or to reflect the occurrence of unanticipated events.
Additionally, statements such as “we believe” and similar statements reflect our beliefs and opinions regarding the subject matter concerned. These statements are based on information available to us as of the date of this document, and although we believe that such information forms a reasonable basis for such statements, such information may be limited or incomplete, and our statements should not be construed as an indication that we have conducted an exhaustive investigation or consideration of all relevant information available to us. Investors are cautioned not to place undue reliance on these statements as they are inherently uncertain.
