What happened
Yahoo Finance reports that Daniel Roberts, CEO of IREN, told Bloomberg Tech, "If you wanted to start today and build a gigawatt AI factory, you are looking 2030 before you get the first compute online." The article states that Nvidia revenue rose to $215.9 billion in fiscal 2026 and that Micron Technology and SK hynix face constrained HBM inventory, sold out months in advance, per the same coverage. Yahoo Finance also reports utilities from Virginia to Texas warning that grid demand is accelerating faster than transmission upgrades can follow.
Technical details
The Yahoo article describes IREN as operating pre-powered data center campuses with secured transmission connectivity, which the piece frames as enabling earlier deployment compared with sites still awaiting utility approvals that can take 18-24 months for site assessment, according to the Quick Read summary in the article. The reporting ties the constraints together: compute component supply pressure, HBM scarcity, and local grid/utility interconnection timelines.
Editorial analysis
Industry observers note that electricity supply and utility interconnection processes are frequent limiting factors for large-scale data center projects, especially when facilities require multiple hundreds of megawatts of dedicated capacity. Projects that appear supply-ready at the component level can still be delayed by permitting, transmission upgrades, and transformer or substation construction timelines.
Context and significance
For practitioners and infrastructure planners, the story underscores that chip availability is only one axis of readiness for AI scale. Observed patterns in the sector show that site selection, grid studies, and utility agreements often dominate lead times and capital scheduling for multi-hundred-megawatt deployments.
What to watch
Track utility interconnection queue times in key regions, announcements of transformer/substation buildouts, and corporate disclosures about pre-provisioned power at campus sites. Also monitor HBM supply announcements from memory vendors and fiscal reporting from major GPU suppliers for continued demand signals.
Key Points
- 1Power and utility interconnection timelines often become the critical bottleneck for AI-scale data centers, delaying compute deployments.
- 2Component shortages remain real: Nvidia revenue surged to $215.9 billion in fiscal 2026 while HBM inventories are reported sold out months ahead.
- 3Pre-powered campuses with secured transmission can shorten deployment lead time relative to greenfield sites still in utility approval queues.
Scoring Rationale
The reported constraint shifts the operational conversation from chips to power and grid readiness, a major practical issue for practitioners planning AI-scale deployments. The story affects deployment timelines and capital planning for data center projects.
Sources
Public references used for this report.
Practice interview problems based on real data
1,625 SQL & Python problems across 15 industry datasets — the exact type of data you work with.
Try 250 free problems
