Microsoft has entered into a remarkable new collaboration to further expand its infrastructure capacity to meet the increasing global demand for artificial intelligence-supported services. In this context, the company made a major agreement worth 9.7 billion dollars with the Australia-based digital infrastructure company IREN, which will last for five years. This partnership is based on the next generation data processing facility to be established in Childress, Texas. Microsoft plans to significantly increase the capacity of Azure artificial intelligence services with the processing power it will obtain from this facility.
The basis of the agreement is high-performance systems containing Nvidia’s new GB300 graphics processors. These systems are planned to be gradually integrated into the data center by 2026. On the other hand, this data center is expected to reach a high electrical capacity of 750 megawatts when completed. This will allow multiple AI models to be trained and run simultaneously at the center. This center, which will directly contribute to Microsoft’s system, will also be equipped with remarkable solutions in terms of energy efficiency.
Microsoft prepares Azure infrastructure for more AI load
IREN was known as a company operating in the field of cryptocurrency mining for a long time. However, the company quickly restructured its infrastructure, taking into account the increasing artificial intelligence needs in the last few years. IREN, which started using its high-capacity GPUs for artificial intelligence workloads, started to gain serious profits from this strategic transformation. CEO Daniel Roberts stated that the agreement signed with Microsoft covers only 10 percent of the company’s total capacity. Despite this, it is estimated that this agreement could provide IREN with approximately $1.94 billion in annual revenue.
On Microsoft’s side, this agreement is seen as a natural consequence of its plans to make the Azure infrastructure ready for more AI loads. The first production cluster with Nvidia GB300 NVL72 systems, deployed last month, was a clear example of this. It is stated that these systems are optimized for causal modelling, multi-modal generative artificial intelligence and agent-based AI solutions. With the introduction of the new infrastructure, Microsoft is expected to provide more stable and faster results in artificial intelligence services. In addition to all this, the company does not deviate from its goal of expanding its data centers on a global scale.
Not only the agreement with IREN, but also ongoing negotiations with other data suppliers shape Microsoft’s plans. An agreement was recently reached with Nscale for the supply of approximately 200 thousand Nvidia GB300 GPUs. This hardware will be distributed across four different data centers, three in Europe and one in the US. Thus, instead of focusing on only one region, the company aims to guarantee system performance and uninterruptibility by switching to a multi-centered structure. This approach stands out as a factor that directly affects user experience.
IREN’s additional $5.8 billion purchase with Dell to support its hardware infrastructure is also a part of these developments. Under this order, hundreds of thousands of GPUs and custom AI hardware will be included in the data center. Although the Microsoft agreement stands out, it is known that IREN is also negotiating with other technology companies. This shows that the company’s position in the market is not temporary, but has turned into a permanent infrastructure provider. The diversification of business partners also reveals how determinedly this transformation is being pursued.
Microsoft’s artificial intelligence investments are no longer just at the software level; It has gained a scale that goes directly down to hardware and energy resources. This growth brings with it higher computing power, more complex models and faster results. In this period when user demands for artificial intelligence services are increasing exponentially every month, such infrastructure moves have become a necessity, not a choice, for technology giants. However, the sustainability of investments at this level is directly related to the effective use of energy resources. Therefore, companies need to manage not only hardware but also energy intelligently.