Reflection AI Raises $2 Billion to Build Open Frontier Models

Following rapid growth from earlier financing rounds, Reflection AI Inc. has raised $2 billion at an $8 billion valuation. The funding round marks a major leap from its valuation of around $545 million earlier this year. The company, founded in 2024 by former Google DeepMind researchers Misha Laskin and Ioannis Antonoglou, aims to establish itself as an open frontier AI lab in the United States.

The startup plans to use the new capital to expand compute resources and train large-scale language models. It is developing systems based on Mixture-of-Experts architectures, which are used in advanced large language models. Reflection AI intends to release its first text-based model early next year, with multimodal capabilities to follow.

The company’s approach combines open access with proprietary infrastructure. It plans to make model weights publicly available while keeping datasets and training pipelines private. Reflection AI’s revenue model focuses on enterprise and government clients seeking customizable and controllable AI systems for internal use or sovereign development.

Investors in the latest round include Nvidia, Disruptive, DST, B Capital, Lightspeed, GIC, Sequoia, CRV, and individual backers such as Eric Yuan and Eric Schmidt. The company’s mission positions it as a Western counterpart to Chinese AI developers like DeepSeek, which has advanced open large language models at scale.