Reflection AI, a company developing open-source artificial intelligence models, has raised $2 billion on an $8 billion valuation. The funding round included participation from prominent investors such as Nvidia, Disruptive, DST, 1789, B Capital, Lightspeed, GIC, Eric Yuan, Eric Schmidt, Citi, Sequoia, and CRV.
With this move, Reflection AI has increased its valuation 15-fold in just seven months, positioning itself as an open-source alternative to major AI companies like OpenAI, Anthropic, and DeepSeek.
Founded in March 2024 by former Google DeepMind researchers Misha Laskin and Ioannis Antonoglou, Reflection AI aims to take AI development power out of the hands of a few large tech companies and make it more open and accessible.
Initially focused on AI systems capable of writing code autonomously, the company is now developing advanced language models that can think and communicate like humans.
Reflection AI plans to keep the training data and processes confidential, while making the models themselves openly available. This approach allows researchers, corporations, and governments to integrate, customize, and deploy the models on their own infrastructure.
Seven months ago, the U.S.-based startup was valued at $545 million. With the $2 billion investment, its valuation has soared 15 times, making it one of the most ambitious names in the AI sector.
After the investment, co-founder Misha Laskin said:
"At a certain point, once you become a large company, you naturally want to have an open model. You want something you can keep in-house, run on your own infrastructure, control costs, and customize for different applications. After all, AI already costs an enormous amount of money, and you want to use it as efficiently as possible. That’s exactly the market we serve."
Noting that models like DeepSeek and Qwen are a warning signal, Laskin added:
"If we don’t take action, the global AI standard will be set by others. It won’t be built in America."
