Not Tesla, Elon Musk is “burning” 1 billion USD/month on this game

Elon Musk, who is aiming to make xAI a leader in artificial intelligence (AI), is facing a huge financial challenge: his startup xAI is burning through $1 billion a month while revenue remains very limited. According to Bloomberg, xAI is trying to raise $9.3 billion in debt and equity to cover costs, but the financial pressure is still huge.

According to people briefed on xAI’s finances, the company is burning through $1 billion a month, and is expected to “burn” a total of $13 billion by 2025, or more than $1 billion a month without stopping. This is mainly due to the huge cost of building AI infrastructure such as the Memphis data center (with 100,000 Nvidia H100 GPUs) and buying specialized chips to train AI models like Grok. Meanwhile, its 2025 revenue is expected to be just $500 million, rising to over $2 billion by 2026, far behind OpenAI (projected to be $12.7 billion in 2025).

To compensate, xAI is raising $9.3 billion, including $4.3 billion in equity and $5 billion in debt, backed by Morgan Stanley. However, more than half of this amount will be spent within the next 3 months, showing a terrible “burn rate”. The company also expects to receive a $650 million refund from a chipmaker, but this is just a “drop in the ocean” compared to its expenses. Since its founding in 2023, xAI has raised $14 billion in equity, but only $4 billion remained at the beginning of 2025, expected to be almost exhausted by the end of Q2 2025. To sustain operations, xAI plans to raise an additional $6.4 billion by 2026.

Musk responded on X, calling the Bloomberg article “nonsense,” but did not provide a detailed rebuttal. This suggests that he remains confident in the long-term strategy despite initial investor reservations about the financing terms. After xAI adjusted its terms to be more friendly, the equity round was completed, attracting major investors such as Andreessen Horowitz, Sequoia Capital, and VY Capital.

xAI’s astronomical costs are no exception. According to the Carlyle Group, the AI ​​industry will need more than $1.8 trillion by 2030 to build infrastructure, from data centers to specialized chips. “Companies building AI models will have to take on debt and burn through a lot of money to compete for leadership,” said Jordan Chalfin of CreditSights. Compared to its competitors, xAI is lagging in revenue. OpenAI is expected to generate $12.7 billion in revenue by 2025, while Anthropic also has a clearer revenue path thanks to its AI commercialization deals.

The main reason for the high costs is AI infrastructure. For example, xAI’s “Colossus” supercomputer in Memphis, with 100,000 Nvidia GPUs, is one of the largest systems in the world, but it is extremely expensive to operate and expand (it is expected to double in size). In addition, xAI has to compete with OpenAI and Google for scarce AI chips, driving up costs. According to Forbes, xAI has spent at least $12 billion in 2024 on Grok and Colossus, indicating a huge scale of investment.