Reflection AI has signed a $1 billion compute agreement with European AI infrastructure company Nebius, securing access to Nvidia’s latest AI chips as the startup expands development of its open-weight artificial intelligence models.
The partnership follows a similar compute agreement Reflection AI recently reached with SpaceX, highlighting the growing competition among AI developers to secure the computing power needed to train and deploy advanced models. Access to high-performance GPU infrastructure has become a key priority as demand for generative AI continues to grow.
Reflection AI is among a new generation of companies focused on open-weight AI models. Interest in this segment has increased as debate continues over the role of proprietary AI systems, data retention concerns, and government involvement in access to advanced models. Last month, the Trump administration urged Anthropic and OpenAI to restrict access to some of their most powerful models. In contrast, increasingly capable open models from Chinese AI companies have drawn greater attention to the open AI ecosystem.
Founded in 2024 by two former Google DeepMind researchers, Reflection AI is currently valued at about $8 billion. The company has raised nearly $2.6 billion from investors, including Nvidia, Sequoia Capital, and Lightspeed Venture Partners, to support the development of its AI technology.
Nebius has also expanded its position as an AI infrastructure provider through several major partnerships. After receiving a $2 billion investment from Nvidia, the company signed a five-year infrastructure agreement with Meta worth up to $27 billion. It also secured a multiyear deal with Microsoft valued at up to $19.4 billion.
The Reflection AI agreement reflects the industry’s continued push to secure large-scale computing resources as AI companies invest in building increasingly capable models and supporting the growing demand for enterprise AI applications.
