Nebius today introduced Token Factory, a managed inference platform that lets vertical AI companies and digital enterprises deploy and optimize open‑source and custom models at scale with enterprise‑grade reliability and control.
Token Factory delivers sub‑second latency, autoscaling throughput, and 99.9 % uptime even for workloads exceeding hundreds of millions of requests per minute. The service is built on Nebius’s full‑stack AI Cloud 3.0 Aether, which provides enterprise‑grade security, proactive monitoring, and consistent performance validated by MLPerf inference benchmarks.
The platform supports a broad portfolio of open‑source models, including NVIDIA Nemotron, DeepSeek, OpenAI’s GPT‑OSS 120 b and 20 b, Meta’s Llama, and Alibaba’s Qwen. Customers can also host their own models, giving them full control over data and deployment strategy.
Early adopters such as Prosus, Higgsfield AI, and Hugging Face are already using Token Factory to power chatbots, coding copilots, high‑performance search, and video‑creation workloads. Prosus reported up to 26× cost reductions compared with proprietary models, while Higgsfield AI has leveraged the platform’s autoscaling to support its expanding video‑generation pipeline.
The AI inference market is projected to reach $254 billion by 2030, driven by the rapid adoption of generative AI and large language models. Nebius’s focus on open‑source models and enterprise‑grade reliability differentiates it from cloud providers and hardware vendors, positioning the company to capture a larger share of this high‑growth market.
Investors responded positively to the launch, citing the platform’s potential to deliver significant cost savings and to meet the growing demand for scalable, low‑latency inference services.
The content on BeyondSPX is for informational purposes only and should not be construed as financial or investment advice. We are not financial advisors. Consult with a qualified professional before making any investment decisions. Any actions you take based on information from this site are solely at your own risk.