Musk’s xAI Holds Talks for Another Massive Fundraising Round

To catch up in the AI race, Elon Musk’s xAI is going to have to burn some proverbial rubber — and lots and lots of literal cash.

On Tuesday, Bloomberg reported that Musk’s artificial intelligence startup is in talks to raise $4.3 billion in an equity investment — on top of the $5 billion it had already been trying to borrow from debt investors. According to a separate report from Bloomberg, the fundraising comes as the company continues to burn through a remarkable amount of cash every month.

Burn, Baby, Burn

Training and operating AI models is expensive. Like, very expensive. According to one estimate from tech pundit Ed Zitron, OpenAI’s total costs may exceed $25 billion this year alone. Carlyle Group CEO Harvey Schwartz recently predicted in a shareholder letter that the industry as a whole will shell out some $1.8 trillion by 2030 to build out AI infrastructure.

xAI is no different than its peers (and, in fact, may be burning through cash even faster than its peers in an attempt to catch up). Unlike OpenAI and industry frontrunners, the Grok chatbot operator has hardly established a revenue stream to offset its rapid burn rate:

• xAI has been burning through $1 billion per month to build its advanced AI models and expects to spend $13 billion this year alone. The company has raised $14 billion in equity fundraising since its launch in 2023, though as of March 31, it had only $4 billion left on its balance sheet, according to materials seen by Bloomberg.  
• On the flip side, the company expects just $500 million in revenue this year. Comparatively, OpenAI recently claimed to have hit $10 billion in annual recurring revenue, driven in part by 3 million paying business users.

X Marks The Spot: Still, xAI does expect the tables to turn eventually. That’s based partly on the hope that its recent merger with fellow Musk venture X (formerly known as Twitter) could prove to be a fruitful partnership. The thinking is that access to X’s content could offer a cheap (or free) alternative to paying for content elsewhere. Because nothing says “quality training data” like an endless stream of memes, jabs and cat videos.