Generative AI coding assistant startup Magic AI seeks to raise $200 million at $1.5 billion valuation

Magic AI Inc., an artificial intelligence (AI) programming startup, is reportedly looking to raise more than $200 million in a new funding round, raising its valuation to $1.5 billion. The talks come just months after the company closed a $23 million Series A round of funding.

Reuters reported, citing three anonymous sources, that venture capital firm Jane Street is in talks to lead a new round of funding that could triple the company’s valuation, even though Magic AI has yet to turn a profit or launch a product.

Magic AI was valued at $500 million in its last funding round in February. The company has raised more than $140 million in total since its founding in 2022, with backing from Alphabet’s Capital G and Nat Friedman and Daniel Gross’s NFDG Ventures.

The funding talks highlight the enormous potential investors see in generative AI coding technology. Companies spend huge amounts of money hiring software developers, but many struggle to find enough of them, making AI tools that can generate code or assist developers very attractive.

Generative AI is at the core of existing products like GitHub’s Copilot and OpenAI’s ChatGPT, which can suggest fixes to the lines of code developers write. But while these tools can help you write code, startups like Magic AI are pushing the envelope by offering to create full-on, automated code bases for entire applications.

Magic AI’s tool, which is still under wraps, is said to allow software engineers to use natural language to describe the type of application or feature they’re trying to create. Magic AI then begins writing all the code needed to create that app. Developers at the startup define the tool as “software that builds software,” and it allows developers to work with an AI to find code, explore its uses, reuse it, and collaborate on code changes.

In other words, Magic AI says its tool can be thought of as a “colleague in your computer,” acting as a smart agent that can do all the heavy lifting that comes with writing and editing code.

Magic AI is one of many startups chasing the dream of AI-generated code. Spurred by the success of GitHub’s Copilot, investors have poured millions into these companies in recent months. One of the most recent funding rounds came in April, when AI coding assistant startup Augment Inc. closed a massive $227 million Series B round, raising its valuation to $977 million. Meanwhile, Devin’s developer Cognition AI Inc. closed a $175 million funding round led by Founders Fund in March, bringing its valuation to $2 billion.

The space also has many well-funded rivals, including Amazon Web Services Inc. and Google Inc., which offer CodeWhisperer and Gemini Code Assist tools that compete with Microsoft Corp.’s GitHub’s Copilot, as well as a variety of startups including Tabnine Ltd., Codegen Inc., Refact, Laredo Labs Inc. and TabbyML Inc. that have yet to secure big funding rounds.

Other companies, such as French AI coding startup Poolside AI, are also looking to raise millions of dollars from investors.

Investors see the success of GitHub’s Copilot as evidence of just how big the generative AI coding industry will become: Over the past year, GitHub’s revenue has grown 40%, with most of that growth coming from its AI coding service, which has more than 1.3 million paying subscribers.

“Microsoft’s success validates a commercial market for AI code assistants and leads everyone to believe there is a clear market demand and customers’ willingness to pay for the right product,” Brian Dudley, a partner at venture capital firm Adams Street Partners, told Reuters.

But building generative AI coding assistants is costly, requiring startups to secure huge datasets to train the large underlying language models, and access to energy-intensive computing resources to perform the training.

According to Reuters, Magic AI plans to use the proceeds from its next round of funding to improve its coding assistant model, which is said to be able to support long context windows, allowing it to process more data in a single query.

The company previously said its model’s ability to understand and process longer prompts and large amounts of context comes from its innovative LLM design, which goes beyond the traditional Transformer model at the core of models such as OpenAI’s ChatGPT.

Image: SiliconANGLE/Microsoft Designer