Nebius recruits Clarify core team... advancing the "Token Factory" reasoning infrastructure upgrade

Dutch artificial intelligence infrastructure company Nebius Group N.V. has brought in the core engineering team of AI orchestration software company Clarifai. At the same time, it has acquired a patent portfolio and licensing for reasoning and computational orchestration technologies, accelerating the upgrade of its hosted reasoning services.

The transaction amount was not disclosed. However, the industry views it as a form of “talent acquisition,” leaning more towards “gaining talent” rather than a pure acquisition. Clarifai founder and former CEO Matthew Zeiler will join Nebius as Senior Vice President of the Research Department. Additionally, some researchers and engineers who spent decades building commercial-grade AI infrastructure will also transition to the company.

Clarifai has consistently offered a full-stack AI platform capable of leveraging both structured and unstructured data to develop intelligent applications. Its features include large-scale data lakes, automated data labeling, search, and indexing tools, supporting developers in training AI models on specific datasets.

What particularly draws market attention is Clarifai’s “computational orchestration” capability. This technology enables companies to centrally manage their on-premises servers and resources rented from public clouds through a unified portal. It can optimize AI resources based on performance and cost conditions, making it a foundational technology for large-scale inference service operations.

This technology complements Nebius’s “Token Factory.” Unlike providers like Amazon Web Services, Microsoft, and Google Cloud that offer general infrastructure, Nebius is categorized as a “Neocloud” service provider focused on large language models. This is also why it, along with CoreWeave, is regarded as a major player in the AI-specific cloud market.

Token Factory is a dedicated inference infrastructure service designed to help pre-trained models operate stably in real-world service environments. As AI model inference scales up, issues related to latency, cost, and resource allocation become more complex. Nebius simplifies these challenges through an architecture that vertically integrates hardware and software.

Recently, Nebius has rapidly increased investments in this area. Last month, it acquired inference software startup Eigen AI for $643 million. If Eigen AI is responsible for “model-level optimization,” making models more efficient, the addition of the Clarifai team is interpreted as Nebius gaining “system-level expertise” in system design and inference operations.

Nebius believes that integrating Clarifai’s technology into Token Factory will bring the company closer to a “full-stack inference platform.” The company states this will help improve “Token efficiency”—reducing costs per word generated or per image—and support more stable multimodal processing, agent-based inference, and long-context memory functions. The goal is to handle both text and visual data simultaneously while minimizing latency growth.

Nebius co-founder and Chief Business Officer Roman Chernin said, “To achieve efficient inference at scale, model optimization, system design, and compute orchestration must be organically integrated. The combination of Clarifai’s team and technology will help clients operate models more stably and cost-effectively in real-world environments.”

This talent acquisition indicates that the competition in AI infrastructure has shifted from simply ensuring GPU availability to “inference efficiency” and “operational automation.” In the context of inference demands growing faster than training needs, Nebius’s ability to demonstrate strong execution around Token Factory will be a key factor influencing the future AI cloud market landscape.

TP AI Notice: This article has been summarized using a language model based on TokenPost.ai. The main content of the original text may be omitted or inconsistent with facts.