Magic co-founded by Eric Steinberger and Sebastian De Ro announced a significant milestone in its journey to push the boundaries of artificial intelligence. The company has successfully raised a total of $465 million in funding, including a recent investment round of $320 million from a mix of new and existing investors.
The latest funding round saw participation from high-profile investors such as Eric Schmidt, Jane Street, Sequoia, and Atlassian. Existing backers, including Nat Friedman, Daniel Gross, Elad Gil, and CapitalG, also contributed to this round, demonstrating continued confidence in Magic’s vision and potential.
Magic is pioneering the development of ultra-long context models, with a focus on their recently trained LTM-2-mini model capable of processing 100 million tokens. This breakthrough equates to approximately 10 million lines of code or 750 novels, marking a significant leap in AI’s ability to understand and process vast amounts of information.
The company’s ambitious goal is to make AI-assisted software development more efficient and accessible. “Imagine if you could spend $100 and 10 minutes on an issue and reliably get a great pull request for an entire feature. That’s our goal,” Magic stated.
To achieve this, Magic has developed its own training and inference stack from the ground up, including custom CUDA implementations. The company currently operates with a team of 23 people and an impressive array of 8,000 H100 GPUs, with plans to scale up to tens of thousands of GB200s in the future.
As part of its growth strategy, Magic is actively hiring across various roles, including Engineers, Researchers, Supercomputing and Systems Engineers, and a Head of Security. The company emphasizes the importance of responsible AI development, likening the sensitivity required in AI advancement to that of the nuclear industry.
With this substantial funding and a clear vision for the future of AI, Magic is poised to make significant strides in the field of inference-time compute, potentially transforming the landscape of software development and AI applications.
