The whole sector has been slumping without news for quite some time now. We are still conceptually tied to BTC. Likewise, investors in the sector have a lot of BTC exposure one way or another, and when margin calls come from left and right, the volatility intensifies.
IREN’s ownership base is exceptionally weak, which has been a topic of discussion for a long time; I believe I’ve touched upon it here before.
Here is by far the lowest-risk performer in the sector, with a zillion contracts in their back pocket, minimum execution risk (colocator), and whose senior notes (just over 6% interest) were oversubscribed to 13B when the ask was 2B.
So, it is the securities market itself that is in no way in a stable state. Similarly, algos aren’t helping the situation at all. Of course, some of us see this as an opportunity ^^
Main points from the ER:
Financials: cra .. crummy. (this was known long in advance, but the additional dent from BTC was a surprise, of course). Fortunately, they are not important to me in terms of the investment case 
I’ll just paste Jim’s summary here (I’ll add that the SBC spike is related to the company’s reporting shift from AUS → US domestic and wasn’t actually a surprise).
2.3B in recurring annual revenue “under contract” - guidance of 3.4B by the end of 2026.
(1.9B MS + 0.4B CA now → 1.5B by the end of the year)
95% of MS GPU financing secured. GPUs 5.8B = 1.94B prepayments + 3.6B debt (<6% interest). That leaves 0.26B, and cash on hand is 2.8B. Still, the assumption is that some other financing will be used for that remainder. That sub-6% interest + prepayment makes the cost of financing ~3-3.5% when CRW et al are getting money at 9%. You can take this as an indicator of something if you want. Personally, I see it as a sign of a healthy competitive advantage.
Everything on schedule regarding GPU deliveries, a couple of notes:
- CA was supposed to be 0.5B at the turn of the year (previous guidance, maybe from summer -25), though it was mentioned even then that it depends on how GPUs are received from the supplier, but “a miss is a miss.”
- The usual phrasing in IREN’s case is that things are progressing “ahead of schedule”; this time the wording for the MS progress was “on schedule.”
The latter is more important, and I’ve heard through the grapevine that the small town of Childress is struggling to keep up with IREN, Lancium, etc.
Field intel: Foundation work for H3 is starting. H2 buildings are finished, and work inside H1 has been going on for about 1.5 months. In Jim Liu’s opinion, the construction schedule looks good; he confirmed that construction is timed according to expected GPU arrivals. Lots of different margins for testing etc. as well, so it’s unlikely to be much ahead of schedule. Let’s keep an eye on it.
Lastly, the OK new site which was already known, but the management’s announcement about it was stunning. Usually, Dan and Will sit on these while investor detectives boil as other players in the industry announce mere pipe dreams.
This brings the total capacity to 4.5GW (2028–>). MS deal 300MW (comparable gross MW). I see that some more bare metal offerings can be built between 2026-2028—the more air-cooled, the more in total. Some unusual collaboration setups with Nvidia/DELL, but it would be a miracle if a Google (TPU) type deal doesn’t come as a colo arrangement in the near future.
I haven’t even had time to watch the earnings call yet; I heard there are some nice nuggets and hints in there. I’ll supplement under EDIT below if any game-changing observations come up.
