Microsoft and IREN's $9.7 Billion Artificial Intelligence Agreement

IREN, which transitioned from Bitcoin mining to the artificial intelligence sector, drew attention with a $9.7 billion GPU cloud service agreement signed with Microsoft. In a statement made on Monday, IREN announced that a five-year AI infrastructure agreement had been reached, showing that this move has created significant excitement among investors.

As part of this agreement, Microsoft will provide IREN with access to Nvidia GB300 GPU hardware. The total value of the agreement was shared as $9.7 billion, and it was also noted that Microsoft made a 20% upfront payment. IREN also announced that it signed a separate agreement with Dell Technologies worth $5.8 billion for the purchase of GPUs and related equipment. It was stated that these GPUs will be gradually deployed at the 750 MW Childress campus in Texas by 2026.

IREN expressed that it will cover the investments with existing cash reserves, customer prepayments, and additional funding sources. Daniel Roberts, co-founder and co-CEO of IREN, stated that the agreement showcases the power and scalability of the company’s “vertically integrated AI cloud platform.” Jonathan Tinter, Microsoft’s head of business development, emphasized that IREN’s expertise, spanning from data centers to GPU infrastructure, has made the company a strategic partner in the artificial intelligence field.

Following the agreement, IREN shares saw a nearly 30% increase in pre-market trading, drawing attention. The current share price is trading at $74.51, and the company’s shares have gained 489% in value since the beginning of the year. According to Bernstein analysts, IREN’s expansion into the AI cloud could generate $500 million in annual revenue by early 2026. Analysts have revised their price target for the shares to $75, a level that was tested immediately after the news.

It is becoming a new trend for companies that come from crypto mining to shift towards the AI sector, leveraging their existing energy infrastructure and data centers. Although the fact that ASIC devices are not suitable for AI operations is a disadvantage, these companies gain a significant advantage in GPU hosting services due to their robust infrastructures.