Magic, an AI startup focused on developing advanced models to generate code and automate a wide range of software development tasks, has secured a significant investment boost. The company announced on Thursday that it has raised $320 million in a fundraising round, with contributions from prominent investors, including former Google CEO Eric Schmidt, Alphabet's CapitalG, Atlassian, and other notable names such as Elad Gil, Jane Street, Nat Friedman & Daniel Gross, and Sequoia.
This latest round of funding brings Magic’s total capital raised to an impressive $465 million, positioning the company among the best-funded AI coding startups, alongside competitors like Codeium, Cognition, Poolside, Anysphere, and Augment. Interestingly, Schmidt is also a backer of Augment, another player in the AI coding space.
Magic had been reported by Reuters in July as seeking to raise over $200 million at a valuation of $1.5 billion. The round exceeded expectations, although the current valuation of the startup remains undisclosed. Back in February, Magic was valued at $500 million.
In addition to the fundraising news, Magic revealed a strategic partnership with Google Cloud to develop two "supercomputers" on the Google Cloud Platform. These supercomputers, named Magic-G4 and Magic-G5, will leverage Nvidia's cutting-edge hardware. The Magic-G4 will use Nvidia H100 GPUs, while the Magic-G5 is set to integrate Nvidia’s next-generation Blackwell chips, expected to be available next year. GPUs, known for their capability to handle numerous parallel computations, are essential for training and running generative AI models.
Magic plans to scale the G5 cluster to "tens of thousands" of GPUs over time, with the two clusters combined targeting a computational power of 160 exaflops—where one exaflop equals one quintillion operations per second.
"We are excited to partner with Google and Nvidia to build our next-gen AI supercomputer on Google Cloud," said Eric Steinberger, Magic’s co-founder and CEO. "Nvidia's [Blackwell] system will greatly improve inference and training efficiency for our models, and Google Cloud offers us the fastest timeline to scale, and a rich ecosystem of cloud services."
Founded in 2022 by Steinberger and Sebastian De Ro, Magic has been developing AI-driven tools designed to assist software engineers in writing, reviewing, debugging, and planning code changes. These tools function like an automated pair programmer, continuously learning and adapting to the context of various coding projects. While several platforms offer similar capabilities, Magic distinguishes itself through its models' ultra-long context windows, known as "Long-term Memory Network" (LTM).
Context windows refer to the input data that AI models consider before generating output. Magic's latest model, LTM-2-mini, boasts a 100 million-token context window, which is the largest of any commercial model currently available. This allows the model to handle vast amounts of data, such as 10 million lines of code or 750 novels, significantly outpacing competitors like Google’s Gemini models, which have a context window of 2 million tokens.
Thanks to its extended context window, Magic’s LTM-2-mini has autonomously completed complex tasks such as implementing a password strength meter for an open-source project and creating a calculator using a custom UI framework. The company is now training an even larger version of the model.
Despite its small team of around two dozen people and the absence of revenue, Magic is targeting a market that could be worth $27.17 billion by 2032, according to Polaris Research. Investors see great potential in the company’s vision and are willing to bet on its success.
Magic’s ambitions go beyond just automating software development. The company envisions a path to Artificial General Intelligence (AGI)—AI that can solve problems more reliably than humans. To advance this goal, Magic has recently hired Ben Chess, a former lead on OpenAI's supercomputing team, and plans to expand its teams in cybersecurity, engineering, research, and systems engineering.
As AI-powered assistive coding tools gain traction, with GitHub reporting high adoption rates among developers, Magic is well-positioned to make significant strides in the industry and potentially shape the future of AI in software development.
