IREN on Nov. 5 shifted its target for putting more than $4 billion of annualized AI cloud revenue capacity into operation from year-end 2026 to January 2027, potentially extending through February.
Its results release specifies Jan. 31 and says commissioning, testing and customer acceptance affect revenue. This run-rate metric reflects commissioned GPUs under contract, not GAAP revenue; recognized sales may be materially lower.
Microsoft delivery changes
Reuters reported that IREN said Microsoft had requested network-design changes affecting some infrastructure deliveries. The news agency also reported quarterly revenue of $228 million against a $221.28 million analyst average compiled by LSEG.
IREN also reported a net loss of $263 million for the quarter ended Sept. 30, compared with a $684 million loss in the previous quarter.
A contract predating the revised timetable
The projects sit within a previously announced relationship. In its Nov. 3, 2025, contract announcement filed with the Securities and Exchange Commission, IREN described a roughly $9.7 billion agreement to supply Microsoft with Nvidia GB300 computing capacity over five years.
The original deployment plan placed the GPUs at IREN’s Childress, Texas, campus in phases through 2026, in liquid-cooled data centers supporting a combined 200 megawatts of critical IT load.
That historical contract value covers multiple years. It is separate from both the revised operational run-rate target and revenue recognized in a single quarter.