IREN to add 50,000+ Nvidia B300 GPUs; expands at-the-market program to $6B
IREN said its IREN Nvidia B300 GPU order will expand processing capacity by roughly 50% after the company entered purchase agreements for more than 50,000 Nvidia B300 GPUs. Separately, IREN disclosed an expanded at-the-market equity program that allows sales of ordinary shares with an aggregate offering price of up to $6 billion, a move that can increase financial flexibility but may dilute existing shareholders.
What IREN announced
IREN, a data center operator focused on AI cloud infrastructure, said the additional Nvidia B300 GPUs would bring its total AI compute fleet to about 150,000 GPUs. The company expects to deploy the hardware in phases during the second half of 2026 across its air-cooled data centers in Mackenzie, British Columbia, and Childress, Texas.
The company also pointed to a revenue target: once fully deployed, the 150,000-GPU fleet is expected to support more than $3.7 billion in annualized AI cloud run-rate revenue by the end of 2026, which would place IREN among larger providers in the AI cloud infrastructure segment.
IREN Nvidia B300 GPU order and the build-out timeline
IREN said the new GPUs are expected to roll out in phases across existing sites in Canada and the U.S. in 2H 2026. The company’s disclosures emphasize the scale-up of compute capacity and the associated infrastructure needs (servers, networking, and related equipment) required to bring GPUs online for customers.
Why the IREN Nvidia B300 GPU order matters for AI cloud customers
For AI cloud buyers, GPU availability and delivery schedules can be as important as price. The B300 is part of Nvidia’s newer data-center GPU lineup aimed at large-scale AI training and inference workloads, and orders of this size can signal an operator’s intent to serve high-throughput enterprise and model-development demand.
Funding and the $6 billion at-the-market program
IREN also filed a prospectus supplement expanding its ATM program capacity to up to $6 billion. ATM programs allow companies to sell shares into the market over time, typically at prevailing prices, rather than all at once. IREN’s filing notes it previously sold shares under an earlier ATM program and updated the prospectus supplement to reflect the expanded capacity.
IREN has said it secured about $9.3 billion in funding over the past eight months through mechanisms including customer prepayments, convertible notes, and GPU leasing/financing arrangements, and it expects roughly $3.5 billion of additional capital expenditures associated with the new GPU deployments.
Context & Analysis
The GPU order underscores how AI infrastructure operators are racing to secure next-generation compute supply, while the expanded ATM program suggests IREN wants the option to raise equity opportunistically alongside other financing tools. Investors often react negatively in the short term to large ATM authorizations because they introduce potential dilution even if management ultimately sells less than the maximum amount.
Concluding Remarks
IREN’s plan combines a major Nvidia B300 GPU purchase aimed at lifting its fleet to roughly 150,000 GPUs with a broader capital strategy that includes a significantly expanded ATM equity offering capacity. The next key milestones will be the pace of GPU deliveries and deployment during 2H 2026, and any disclosures about actual share sales under the ATM program.
FAQs
Q : What is the IREN Nvidia B300 GPU order?
A : IREN Limited says it entered purchase agreements for more than 50,000 Nvidia B300 GPU GPUs, expanding its total AI compute fleet to around 150,000 GPUs.
Q : When will IREN deploy the new GPUs?
A : IREN expects phased deployment in the second half of 2026 across its Mackenzie Data Center and Childress Data Center facilities.
Q : What is an at-the-market (ATM) offering, and what did IREN file?
A : An ATM (At-The-Market) offering allows a company to sell shares gradually into the public market at prevailing prices rather than issuing them all at once.
IREN filed to expand its ATM capacity to up to $6 billion, giving it flexibility to raise capital over time.
Q : Why did the stock react negatively in pre-market trading?
A : Large ATM authorizations can create dilution concerns. Investors may worry that new shares could be issued later, increasing the company’s total share count and potentially reducing the value of existing shares.
Q : How much AI cloud revenue is IREN targeting with the expanded fleet?
A : IREN projects that its 150,000-GPU AI compute fleet could support more than $3.7 billion in annualized AI cloud run-rate revenue by the end of 2026.
Q : How does IREN plan to fund the expansion?
A : The company says it has secured about $9.3 billion in funding over the past eight months and expects around $3.5 billion in additional capital expenditure (capex) tied to the new GPU deployments.
Facts
Event
IREN announces 50,000+ Nvidia B300 GPU purchases and expands ATM offering capacityDate/Time
2026-03-05T02:30:00+05:00 (converted from Mar 4, 2026 16:30 ET press-release timestamp)Entities
IREN Limited (NASDAQ: IREN); Nvidia (NASDAQ: NVDA); Mackenzie, British Columbia; Childress, TexasFigures
50,000+ B300 GPUs; ~150,000 total GPUs; up to $6.0 bn ATM capacity; >$3.7 bn annualized AI cloud run-rate target; ~$9.3 bn funding cited; ~$3.5 bn additional capex expectedQuotes
(No direct quotes provided in the supplied material or cited releases within the sources used.)Sources
GlobeNewswire press release; SEC prospectus supplement; CoinDesk coverage