Satellite analysis highlights possible data center delays tied to Microsoft’s AI infrastructure plans.
Nvidia-backed IREN said Microsoft-requested networking equipment changes delayed its Childress, Texas AI data center project, pushing a key revenue target from year-end to January or February.
IREN said a Microsoft-requested change to networking equipment delayed its large AI data center project in Childress, Texas. The Nvidia-backed cloud provider is building the four-phase project under a $9.7 billion contract with Microsoft.
IREN now expects to reach $4 billion in operating annual recurring revenue by January or February, after previously targeting year-end. The company said the requested changes were intended to help support the latest AI models.
IREN co-CEO Daniel Roberts said the first building was delivered in August and described the data centers as “practically done.” He said the company is waiting for remaining equipment to arrive so it can finish commissioning.
Roberts also said IREN and Microsoft have a “great relationship,” framing the delay as an equipment and commissioning issue rather than a breakdown in demand.
The delay adds to scrutiny of neoclouds, which rent AI compute to businesses and have faced power shortages and construction setbacks. Microsoft recently said it spent $5 billion less than expected on AI infrastructure because of delays from third-party cloud providers, though it did not name providers.
Business Insider reported that satellite images analyzed by research firm Aterio pointed to IREN and Nebius. For investors and customers, the key takeaway is that AI infrastructure demand may remain strong while delivery timelines stay vulnerable to equipment, power, financing, and construction constraints.
IREN shares plunged Thursday and later rebounded in after-hours trading following its earnings report. The company reported $228 million in quarterly revenue and a net loss of $263 million.
Roberts said IREN is slowing expansion plans because of tougher financing conditions, though customer demand remains strong. Neoclouds like IREN are important to Nvidia because they expand Nvidia’s customer base beyond major cloud providers such as Google, Amazon, and Microsoft.
Satellite analysis highlights possible data center delays tied to Microsoft’s AI infrastructure plans.

A new filing shows OpenAI-related arrangements made up nearly 11% of Microsoft’s quarterly revenue.

Anthropic’s S-1 shows fast revenue growth, usage-based demand, and huge compute commitments.
Mustafa Suleyman said Microsoft’s MAI-Base-2 beat an OpenAI base model on certain pre-training metrics using less compute.