Reflection AI, a new startup cofounded by former Google DeepMind researchers Misha Laskin and Ioannis Antonoglou just raised a staggering $2 billion in funding. This success has launched the company’s valuation to a staggering $8 billion. Reflection AI launched nationally in March 2024. It aspires to become a serious contender in the cutthroat AI scene, competing against well-known models such as DeepSeek.
That recent funding has more than doubled Reflection’s valuation to $3 billion. It exploded from $545 million seven short months ago, proving a lofty 15x that amount. This explosive trajectory is a testament to investor belief in the company’s vision and potential disruptive effect on the AI landscape.
Right now, Reflection AI has a full-time staff of 60, nearly all AI researchers and engineers. Tamas Laskin was instrumental in creating reward modeling for DeepMind’s Gemini project. At the same time, Antonoglou was one of the creators of AlphaGo, the historic AI that in 2016 beat the world champion at the board game Go.
Reflection AI has ambitious plans to make model weights publicly available. Its datasets and full training pipelines will remain proprietary. The firm is preparing to release a new next-generation large language model next year. This new model will be trained on an astounding “tens of trillions of tokens.” The first model will be text-based, but multimodal capabilities are anticipated to follow.
In order to make its model training easier, Reflection has gotten hold of a large compute cluster. The company plans to bring in revenue mainly from the big businesses that build products on top of Reflection’s models. Further increasing revenues, government contracts to develop “sovereign AI” systems will account for a significant portion of revenue expansion. Any researcher will have free access to Reflection’s models. This is in keeping with the company’s stated objective of ensuring the United States wins the AI race.
The company’s thoughtful approach is what earned them support from titans of the industry. Clem Delangue, co-founder and CEO of Hugging Face, remarked:
“DeepSeek and Qwen and all these models are our wake-up call because if we don’t do anything about it, then effectively, the global standard of intelligence will be built by someone else.” – Misha Laskin
Laskin acknowledged that only a limited number of companies can utilize the complete infrastructure stack necessary for advanced AI applications:
“So you can either choose to live at a competitive disadvantage or rise to the occasion.” – Misha Laskin
Reflection AI’s dedication to developing open-source AI reinforces its goal to promote innovation within the United States. As they prepare for their upcoming model release and secure partnerships with enterprises and governments alike, Reflection aims not only to compete but to lead in setting standards for future AI development.
“This is indeed great news for American open-source AI. Now the challenge will be to show high velocity of sharing of open AI models and datasets (similar to what we’re seeing from the labs dominating in open-source AI).” – Clem Delangue
David Sacks also noted the significance of Reflection’s initiatives:
“It’s great to see more American open source AI models. A meaningful segment of the global market will prefer the cost, customizability, and control that open source offers. We want the U.S. to win this category too.” – David Sacks
Laskin further explained the importance of making model weights accessible:
“In reality, the most impactful thing is the model weights because anyone can use them and start tinkering with them.” – Misha Laskin
Despite offering model weights for public use, Laskin acknowledged that only a limited number of companies can utilize the complete infrastructure stack necessary for advanced AI applications:
“The infrastructure stack, only a select handful of companies can actually use that.” – Misha Laskin
He also highlighted that larger enterprises inherently prefer open models:
“Once you get into that territory where you’re a large enterprise, by default you want an open model.” – Misha Laskin
Reflection AI’s commitment to advancing open-source AI underscores its mission to foster innovation within the United States. As they prepare for their upcoming model release and secure partnerships with enterprises and governments alike, Reflection aims not only to compete but also to lead in setting standards for future AI development.
Leave a Reply