The AI computing power race: How xAI’s ambitions impact the market
Some opponents of global digitalization and cautious investors suggest that the boom in investment in AI infrastructure is close to its peak. Tech-heavy indices like NASDAQ 100 (including its derivative, NASDAQ 100 futures) continue to respond sharply to AI-related announcements, reflecting strong investor confidence in the sector’s ongoing growth. However, the actions of key players indicate the opposite — the computing power race is only gaining momentum.
The announcement of Elon Musk’s intention to attract an additional $12 billion in debt financing for his company xAI was a fresh confirmation of growing investment and adoption trends. The billionaire’s main goal is the large-scale purchase of accelerators, which are critical for the development of the Grok chatbot, thereby influencing Nvidia stock price.
Looking at the generous funding attracted by rival OpenAI to grow its computing clusters, the head of xAI doesn’t intend to lag behind.
Earlier, there were rumors that Musk planned to direct the resources of the sister companies, SpaceX and Tesla, to the development of an AI startup. However, the latest information points to a different, larger source: increased debt burden directly on xAI.
This new $12 billion is an addition to $10 billion that СEO raised just a few weeks ago. The previous amount was formed via a combination of loans and proceeds from the sale of shares of xAI Holding to investors.
The funding structure is also notable: Valor Equity Partners received creditors’ orders to raise the necessary capital. Valor’s founder, Antonio Gracias, is known for his close and long-term business ties with Elon Musk, which adds credibility to the information.
The loan funds raised are expected to be used to purchase a large batch of Nvidia accelerators, presumably the latest models. This equipment will then be leased to xAI to service the needs of the new data center. The main objective of the infrastructure enhancement is to ensure the development and training of the next generation of the Grok chatbot.
Not all lenders are willing to take unlimited risks. Some of the potential participants in the transaction seek to set risk-mitigation conditions, insisting on limiting the repayment period to three years and setting a hard limit on the amount of new borrowings.
Elon Musk himself revealed this week that 230,000 Nvidia accelerators are already involved in Grok training. 30,000 of them belong to the advanced GB200 generation. At the same time, the load associated with reasoning processes is transferred to external cloud providers. Musk also promised to launch a new supercluster of unprecedented power, consisting of 550,000 GB200 and GB300 accelerators.
According to some estimates, xAI’s total spending on computing power could reach a whopping $13 billion until the end of this year. Musk reacted to early rumors about plans to raise external capital for the startup by denying the claims, stating that he and his partners had “enough money.”
The current moves to raise $12 billion in debt, coupled with the previous $10 billion, clearly indicate that xAI’s ambitions require financing well beyond available capital, even for a billionaire like Musk.