You are currently viewing Reflection AI Raises $2 Billion to Build Open-Source Frontier Models, Valued at $8 Billion

Reflection AI Raises $2 Billion to Build Open-Source Frontier Models, Valued at $8 Billion

Prime Highlights

  • Reflection AI is positioning itself as an open-source alternative to AI giants like OpenAI and Anthropic, aiming to democratize access to frontier AI models.
  • The startup plans to release a frontier-scale language model in early 2026, trained on tens of trillions of tokens.

Key Facts

  • Reflection AI has raised $2 billion in new funding, bringing its valuation to $8 billion—up from $545 million just seven months ago.
  • Its 60-person team includes top researchers and engineers from DeepMind and OpenAI, with backers like Nvidia, Sequoia, and Eric Schmidt.

Background

Reflection AI, a startup founded in 2024 by former DeepMind researchers Misha Laskin and Ioannis Antonoglou, has raised $2 billion. The company is now valued at $8 billion, one of the highest for any AI company this year. Just seven months ago, its value was $545 million, so this marks a huge jump.

The company, which first focused on autonomous coding agents, is now positioning itself as an open-source alternative to major AI labs like OpenAI and Anthropic. It also aims to become a Western counterpart to Chinese AI firms such as DeepSeek and Qwen.

Reflection AI has assembled a 60-member team of researchers and engineers, many from DeepMind and OpenAI. The startup plans to release a frontier-scale language model trained on tens of trillions of tokens early next year.

“We built a large-scale platform capable of training advanced Mixture-of-Experts models, something once possible only in the top labs,” said CEO Misha Laskin. He added that open-source AI is key for the U.S. and its allies to remain competitive globally.

The company’s approach to “open” focuses on releasing model weights for public use while keeping datasets and training pipelines private. Laskin said the business model targets enterprises and governments seeking control, customization, and cost efficiency in AI systems.

Major investors in the latest round include Nvidia, Sequoia, Lightspeed, GIC, Citi, B Capital, and former Google CEO Eric Schmidt.

According to Laskin, the funds will be used to expand computing infrastructure and scale model training. Reflection AI’s upcoming model will initially focus on text-based tasks, with multimodal capabilities planned for future releases.

Read Also : Beauty Tech Group Raises £29m in £300m IPO on London Stock Exchange