Nebius Launches Token Factory to Boost Open AI Models

This article first appeared on GuruFocus.

Nvidia (NASDAQ:NVDA)-backed Nebius (NASDAQ:NBIS) is making a big move in the AI space with the launch of its Token Factory, a new platform built to help companies run and scale open-source AI models more efficiently and at a fraction of the cost.

Warning! GuruFocus has detected 6 Warning Signs with NVDA. Is NVDA fairly valued? Test your thesis with our free DCF calculator.

The Netherlands-based cloud firm describes Token Factory as a kind of all-in-one toolkit for developers. It combines inference, post-training, and access management into a single platform, making it easier to take AI models from testing to full production. It supports major open models like Nvidia's Nemotron, OpenAI's (OPENAI) GPT-OSS, Meta's (NASDAQ:META) Llama, DeepSeek, and Alibaba's (NYSE:BABA) Qwen and promises to cut inference costs and latency by up to 70%.

We didn't just want to build another model-serving platform, said co-founder Roman Chernin. We wanted to help teams actually scale faster, smarter, and more cost-effectively.

Developers are already using Token Factory for everything from chatbots and copilots to search and customer support tools. Nebius also teamed up with Hugging Face, whose CTO Julien Chaumond said the collaboration is helping developers deliver faster and more reliable inference on large open models.

View Comments