The generative artificial intelligence-powered coding startup Magic AI Inc. is reportedly hoping to raise more than $200 million in a new funding round that will bring its value to $1.5 billion.

The talks are ongoing just months after the company closed on a $23 million Series A round of funding. Reuters cited three unnamed sources today as saying the venture capital firm Jane Street is in talks to lead the new round, which would triple Magic AI’s valuation, even though the company does not yet earn any revenue nor have any product currently available.

Magic AI was previously valued at $500 million in the wake of its last funding round in February. In total, the company has raised over $140 million since being founded in 2022, with backers including Alphabet Inc.’s CapitalG and Nat Friedman and Daniel Gross’s NFDG Ventures.

The funding talks highlight the enormous potential investors see in the promise of generative AI coding technologies. Enterprises spend enormous sums of money on hiring software developers, and many struggle to find enough of them, so AI tools that can either generate code or assist developers are extremely enticing.

Generative AI sits at the heart of existing products such as GitHub Inc.’s Copilot and OpenAI’s ChatGPT, which are both capable of suggesting fixes for lines of code written by developers. But though these tools can aid in code creation, startups such as Magic AI extend that promise by going whole hog and crafting the codebase for entire applications automatically.

Magic AI’s tool, which is still under wraps, is said to enable software engineers to use their natural language to describe what kind of application or function they’re trying to create. Magic AI will then proceed to write all of the code needed to create that app. The startup’s creators define their tool as “software that builds software,” enabling developers to work with AI to find code, consider its use, reuse it and collaborate on code changes.

In other words, Magic AI says its tool can be thought of as a “colleague inside the computer,” acting as a smart agent that can perform all of the grunt work involved in creating and editing code.

Magic AI is one of a host of startups pursuing the dream of AI-generated code. Encouraged by the success of GitHub’s Copilot, investors have thrown millions of dollars at those companies in recent months.

One of the most recent funding rounds came in April, when the AI coding assistant startup Augment Inc. closed on a hefty $227 million Series B round that lifted its valuation to $977 million. Meanwhile, Cognition AI Inc., creator of Devin, closed on a $175 million funding round led by Founders Fund in March, bringing its value to $2 billion.

There are many other well-funded rivals in the market too, with the likes of Amazon Web Services Inc. and Google LLC offering their CodeWhisperer and Gemini Code Assist tools that compete with Microsoft Corp.-owned GitHub’s Copilot, plus various startups that have not yet secured megabucks funding rounds, such as Tabnine Ltd., Codegen Inc., Refact, Laredo Labs Inc. and TabbyML Inc.

Others, such as the French AI coding startup Poolside AI, are also looking at raising millions from investors.

Investors see the success of GitHub’s Copilot as proof of just how big the generative AI coding industry will likely become. Over the last year, GitHub increased its revenue by 40%, with its AI coding service driving the bulk of that growth with more than 1.3 million paid subscribers.

“The success of Microsoft has validated the commercial market for AI code assistants, leading everyone to believe there is clear market demand and a customer willingness to pay for the right product,” Brian Dudley, a partner at the venture capital firm Adams Street Partners, told Reuters.

However, building a generative AI coding assistant comes at a heavy cost. Startups must secure enormous datasets for training their underlying large language models. They also need to access significant, energy-intensive computing resources to carry out the training.

According to Reuters, Magic AI will use the funding from its next round to improve its coding assistant models, which are said to be capable of supporting long-context windows, meaning they can process more data in one query.

The company has previously said that its models’ ability to understand and process longer prompts and a large amount of context is thanks to the innovative design of its LLMs, which go beyond the traditional transformer model that sits at the heart of models like OpenAI’s ChatGPT.

Image: SiliconANGLE/Microsoft Designer