Magic, an AI startup that creates models to automate code generation and software development, has landed $320 million in its latest funding round. Led by former Google CEO Eric Schmidt, this investment saw contributions from Elad Gil, Nat Friedman, Daniel Gross, Jane Street, Sequoia, which backed Fireworks AI and Statsig, and Atlassian, which recently invested in Allstacks
This round follows the $117 million Series B round raised a few months back from NFDG Ventures, Alphabet’s CapitalG, and Elad Gil. With the latest raise, the total funding raised by Magic accounts for $465 million.
Partnership with Google Cloud
In addition to the investment, Magic also announced a partnership with Google Cloud to construct two AI supercomputers. The first one, the Magic-G4, will utilise NVIDIA H100 GPUs, while the more advanced Magic-G5 will incorporate NVIDIA’s next-generation Blackwell chips. This collaboration aims to scale up to tens of thousands of GPUs over time, transforming AI model training and inference.
Amin Vahdat, VP and GM of ML, Services, and Cloud AI at Google Cloud, highlighted the partnership’s potential: “Google Cloud’s end-to-end AI platform provides high-growth, fast-moving companies like Magic with complete hardware and software capabilities for building AI models and applications at scale.”
What does the company do?
Founded in 2022 by Eric Steinberger and Sebastian De Ro, Magic develops AI-driven tools designed to help software engineers code, review, debug, and plan changes. These tools work like an automated pair programmer, trying to understand and learn about the context of different coding projects.
The company has trained LTM-2-mini, a model capable of handling contexts up to 100 million tokens. It is equivalent to about 10 million lines of code or 750 novels. This advancement could improve code synthesis by allowing models to consider vast amounts of code, documentation, and libraries during inference.
Magic believes that pre-training has limitations and that inference-time computing represents the next frontier in AI. The company aims to enable developers to spend minimal time and resources on an issue and reliably receive high-quality pull requests for entire features.
Competition
The US startup faces stiff competition from other AI coding startups like Cursor, Codeium, Cognition, and Augment, and market leaders like GitHub Copilot. Given that Magic’s focus is on ultra-long contexts, it could be key to setting it apart in the industry.
