IREN delays Microsoft data center deliveries amid $15 billion spending push

Timeline shifts for Microsoft

IREN, an Australia-based firm specializing in AI data centers, announced a delay in delivering some infrastructure to Microsoft. The company clarified that these delays are not resulting in any financial penalties for the firm. Instead, the shift was prompted by specific requests from the software giant to make changes to the network design.

The company also adjusted its internal goals for reaching a specific revenue milestone. IREN originally aimed to hit a $4 billion annualized run-rate revenue target by the end of the year, but it has now pushed that goal to January. The company noted that the timeline for this target could potentially extend until the end of February.

These adjustments come as Microsoft reportedly tightens its oversight of projects managed by neocloud providers. These are smaller, specialized cloud companies that provide the massive computing power needed for AI. Microsoft is reportedly evaluating whether to continue with certain providers as it seeks to manage rising costs and project risks.

Massive infrastructure spending

Despite the delivery delays, IREN is moving forward with a significant spending plan for its hardware and facilities. The company estimates it will spend about $15 billion in capital expenditures over the remainder of its fiscal year. This spending is intended to support the expansion of its AI cloud business and the acquisition of necessary computing equipment.

The company's financial filings show a sharp increase in these commitments over a short period. At the end of September, IREN reported capital commitments of $18.95 billion, which was a notable jump from the $13.81 billion reported three months prior. This reflects the rapid pace at which the company is scaling its physical footprint to meet AI demand.

A major portion of this growth is tied to specialized hardware for artificial intelligence. The amount of GPU hardware on IREN's balance sheet nearly tripled in the last three months, rising from $2.04 billion to $5.82 billion. This move underscores the company's strategy to secure the physical components required to run large-scale AI models.

Growth in AI cloud revenue

IREN reported strong performance in its AI cloud division during the first quarter. Revenue from this specific segment grew from $71 million in the previous three months to $200 million in the most recent quarter. This growth contributed to a total quarterly revenue of $228 million, which exceeded the average analyst estimate of $221.28 million.

The company's revenue is supported by a diverse set of contracts with major technology players. These include agreements with Nvidia and Microsoft, along with 13 other partners. IREN noted that roughly two-thirds of its revenue backlog is tied to investment-grade counterparties, which provides a level of stability for its long-term planning.

Executives shared during a post-earnings call that the company ended September with $1 billion in annualized run-rate revenue. They expect roughly one-quarter of that amount to be recognized during the December quarter. This trajectory shows the company is scaling its operations even as it navigates the complexities of large-scale construction and hardware procurement.

Market pressure and regulation

The news comes at a time of significant volatility for the broader AI sector. IREN's stock experienced a 17% decline during a wider selloff that impacted chipmakers and other neocloud companies like CoreWeave. This selloff was driven by fears that a victory for Democrats in the US midterms could lead to stricter regulations on the technology sector.

The pressure on the industry is compounded by the high costs and logistical hurdles of building out data centers. As companies like Microsoft look to manage these risks, they are becoming more selective about which third-party providers they use. This shift could impact how smaller firms like IREN, Nebius, and Nscale manage their relationships with the world's largest tech companies.

IREN is currently building 200 megawatts of data-center capacity for Microsoft in the state of Texas. While the company remains committed to significant investment in computing equipment, the evolving regulatory and economic landscape creates a complex environment for growth. The company must balance rapid infrastructure expansion with the changing demands of its high-profile clients.