Magic AI builds frontier code models designed to automate software engineering and accelerate AI research toward safe AGI. The company operates across multiple technical domains including generative models for code, large-scale pre-training, domain-specific reinforcement learning, ultra-long context windows, and inference-time compute.
The company has raised $515 million in funding from investors including Nat Friedman, Daniel Gross, CapitalG, and Sequoia. Magic AI operates substantial compute infrastructure, including 8,000 H100s to support its research and model development.
Magic AI is organised around a small team of engineers and researchers focused on fundamental research problems rather than incremental improvements. The company values integrity, hands-on building, and collaborative problem-solving, with an emphasis on hiring people who seek ownership of large components of meaningful challenges rather than executing predetermined roadmaps.
