Sabitlenmiş Tweet
IREN - The Unexpected Hyperscaler
In business, SCALE is best way to make big profits. If there is any team I trust to scale, it's $IREN. At Childress, they printed row after row of data centers, becoming the most profitable BTC miner. Now, with more funding than ever, they are ready to scale their AI Cloud at breakneck speed.
Oracle presented their gross profit for IaaS, the numbers look promising. IREN could achieve even lower costs with their DCs.
If IREN nails their AI DC building efficiency, scales up to secure GPU/server discounts, and power costs remain stable, they will have by far the lowest cost of operations in the market.
As they scale, they will get incremental increases in their revenue stream, secured by years long contracts. This will be used to fund further expansion.
Scaling to 3GW and beyond will need a lot of capital. Fortunately, they have the most CRACKED financing team in the industry. And there is a huge appetite to fund the buildout.
Challenges
1. Maintaining high AI Cloud uptime as they scale
They have 25 years HPC expertise through the acquisition of PodTech in 2020. Since then, they have evolved the team further, launching their AI cloud in February 2024. Undoubtedly learning a lot along the way.
The most common reason for cloud downtime is third party failure. IREN is fully vertically integrated: owning land, power, buildings, and servers.
2. Building fast and efficiently
Here we can look to the buildout of Childress. While most investors were twiddling their toes, IREN went from 100MW (June 2024) to 650MW (June 2025) in operating capacity. Making them the least dilutive, most profitable, and fastest building bitcoin miner. Some mining companies still suffer PTSD from this event.
3. Competition
This is a lucrative industry, hyperscalers and neoclouds are spending their hearts out. Even software companies like Poolside are jumping in. There will be competition that puts downwards pressure on GPU-hour pricing. But let it be clear, building data centers is not easy! That's why others outsource it. And long term, having the lowest expenses is a great advantage. If bare metal GPU prices fall, IREN will be able to out-earn others.
Demand and supply is currently highly imbalanced, so I predict good times for all in the next few years.
AI Cloud
As they scale, they will learn what big customers want in a cloud, and evolve theirs accordingly. Building up talent and expertise as they go along. They are taking a slow and steady approach to the software stack, waiting to see how things evolve.
I think the software part of the business can surprise in the future. If I may speculate, IREN could acquire a promising software startup that scales up on their Cloud. But that makes sense at a bigger market cap, no rush now.
Closing
Some might object, saying this is a commodity, and a race to the bottom. I'd say this commodity is INTELLIGENCE ON DEMAND, which has a different non-physical demand structure. And, in commodity markets, scale is king.
That's my summarized investment thesis going forward. Don't panic. Scale = Good. IREN has more cool things about it, I'll link the main X resources below if you want to learn more.
@FransBakker9812 satellite image progress at IREN sites.
@litigious_dulce AI Cloud deep dives, horse.
@bitcoinbutcher1 Valuation, X spaces, cooking tips.
@TheKamaHsutra Financial calculations, bedtime reading.
@Umbisam Financing insights.
@jiahanjimliu Technical & software insights.
@pepe_maltese Industry insights, frog.
Many good people and animals in this community. This is also a very fun industry! Many cool useful AI tools are coming, and IREN will power them. 😋
Objections welcome, always looking to learn more.
English
