Why IREN Limited Stock Is Falling After a $6 Billion Share-Sale Plan and Huge Nvidia GPU Order

March 6, 2026
Why IREN Limited Stock Is Falling After a $6 Billion Share-Sale Plan and Huge Nvidia GPU Order

NEW YORK, March 6, 2026, 08:12 EST

IREN stock dropped roughly 8.5% in premarket trading Friday, after news broke of a hefty stock sale alongside a major AI hardware purchase. The data center firm revealed it’s acquiring more than 50,000 Nvidia B300 GPUs, pushing total capacity to 150,000 units. Alongside that, IREN announced a new at-the-market offering targeting up to $6 billion. 1

IREN’s latest step comes just as it works to pivot from bitcoin mining’s wild swings and secure hard-to-get AI chips. Notably, it’s happening alongside a $9.7 billion, five-year cloud deal with Microsoft. Back in November, Microsoft CFO Amy Hood warned that the company’s AI capacity shortage could persist “at least mid-2026.” 2

IREN plans to roll out the new chips in stages over the latter half of 2026 at its Mackenzie site in British Columbia and in Childress, Texas. With the expanded fleet, the company is projecting annualized run-rate revenue of more than $3.7 billion—the yearly sales rate IREN expects this level of capacity could bring in.

IREN is looking at roughly $3.5 billion in extra capital spending to back the expansion, the company said. It’s set to rely on funding secured in the last eight months—customer prepayments, convertible notes, GPU leasing deals, and GPU financing all in the mix. Payment comes after shipment, according to IREN, which takes some pressure off working capital in the short term. 3

IREN disclosed in another SEC filing that its fresh at-the-market (ATM) program is taking the place of its prior $1 billion facility, which the company has now completely tapped out. Under that earlier arrangement, IREN sold 66.7 million ordinary shares to bring in $1 billion. Now, with the new program, the company has clearance to raise as much as $6 billion, working with banks such as Goldman Sachs, Jefferies, JPMorgan and Citi.

“Scaling to 150,000 GPUs positions IREN among the largest AI cloud infrastructure providers globally,” said co-founder and co-CEO Daniel Roberts. Roberts added that locking in hardware early could shave down time-to-compute and boost execution certainty, especially in today’s supply-constrained market. 4

The new order ups the ante. Back on February 5, IREN announced a target of 140,000 GPUs and $3.4 billion in annualized run-rate revenue by the close of 2026, despite quarterly revenue dropping to $184.7 million from $240.3 million the previous quarter, with a net loss of $155.4 million. At that time, Roberts described the market as the “strongest demand environment to date.” 5

IREN faces plenty of company. Back in January, Reuters noted Bitfarms brought in advisers to explore shifting toward AI data centers. Riot Platforms and MARA Holdings, according to a different Reuters piece from December, have also started steering their mining operations to support AI computing, a move that’s picking up as crypto markets get shakier. 6

The upside hinges on IREN’s ability to deliver. That $3.7 billion revenue target? It’s internal, not locked in by contracts. H.C. Wainwright notes the firm still has to demonstrate it can line up competitive financing for hardware—especially for compute that hasn’t been contracted yet. Short-term dilution risk is also up. 7