Ras @RasInvests

@Dr_Crossroads Agreed! Are you planning to add to your $CRWV position today? 🤗

$IREN Here is what the co-CEO had to say about the most sensitive topic shared on the weekend, on slowing down the pace of frontier AI lab and what it means for the AI infrastructure companies such as $IREN $NBIS $MU $AAOI et al.

Daniel Roberts @danroberts0101

2 days ago

A few people asked over the weekend what the calls to slow the pace of frontier AI mean for the buildout. Some thoughts: Even if models never improved from here, just rolling out what they can already do would take more compute than the world can build for years. The debate

174 421 3K 804K 503

$IREN he said it again 😂

Kalshi Finance @Kalshi_Finance

4 days ago

BREAKING: Trump says AI data centers are the oil of the next 50 years

167 295 4K 2.1M 345

@Han_Akamatsu @grok is this data accurate? 🤔

$IREN I found this very useful.

Daniel Roberts @danroberts0101

5 days ago

Goldman's CommTech conference, San Francisco. 48 hours of @IREN_Ltd investor meetings, and the mood has shifted meaningfully. Key takeaways:

165 281 2K 437K 609

@HyperTechInvest I agree. Based on your analysis, is it a good approach from $IREN ? Why now?

@OInvests That is a high-level confidence. Its my biggest holdings as well! Lets make some money😄

$IREN $NBIS $CRWV this is music to our ears 😍

Wall St Engine @wallstengine

5 days ago

MICROSOFT PLANS 3X DATA CENTER EXPANSION $MSFT plans to expand its global data center capacity from roughly 12 GW today to more than 38 GW by 2032, adding about 26 GW of capacity. AI-specific compute is expected to grow from around 2 GW today to roughly one-third of total

63 149 1K 456K 192

$IREN some good news on their 800 MW Bundey project in Australia. Bwt: wasn't this planned for 2028 or did I miss something here? 😳

IREN @IREN_Ltd

6 days ago

IREN is partnering with NVIDIA to accelerate deployment of large-scale AI factories by combining NVIDIA’s DSX AI factory architecture with IREN’s expertise across power, land, data centers, GPU deployment and infrastructure operations.

69 299 3K 379K 163

@Han_Akamatsu Mr Akamatsu, the pleasure is truly ours. I wish I knew you a bit earlier. Your HMA gudie is very useful so does your daily Gamma analysis. I am learning big time and very greatful for all you do 🙏

@OInvests Not yet 😕. Good luck for you🤞

$IREN great news!

IREN @IREN_Ltd

a week ago

IREN's 2GW Sweetwater Hub (Sweetwater 1 and Sweetwater 2) has been included in ERCOT's Batch Zero process as Base Load. Sweetwater 1 (1,400MW) and Sweetwater 2 (600MW) form part of our >5GW global data center development portfolio. At Sweetwater 1, construction of 300MW of data

82 271 2K 211K 93

@moninvestor Ah what a signal! Regardless, I will skip selling my $CRWV position as its a hedge against my $IREN and $NBIS positions😄

@jiahanjimliu @MarkosAAIG @Mark_AAIG Just followed them, thanks for the recommendation 🤗

@TheProfInvestor There are a lot of great guys including yourself helping us with the analysis. But my guy @nanotitan28 is worth a follow. He has relatively small followers but I would be happy if you support his work, Prof. Thanks!

@bitcoinbutcher1 Congrats bro and enjoy your new life!

@retail_mourinho I hope so! I am ready to fly to the moon and beyond ☺️

@EchoAnalysis Thanks for all you do 🙏 Honestly you are one of the first names I look everyday for chart analysis. Keep on doing the great job 🤗

My post about the missing 2027 bridge got way more attention than I expected. Seems I wasn’t the only one thinking $IREN had all the right ingredients in the earnings… but somehow forgot to connect them. 😂 So here is the second slide I think was missing. And this one may actually be more important for the long-term thesis: $25–30 BILLION FY27 CAPEX When I first saw that number my reaction was: WTF IREN is a ~$15B company today (ok, after friday selloff it is more like 13,5B) and plans to deploy almost TWICE its entire market cap in capex within ONE fiscal year. But before we even talk about financing: Can we appreciate how insane that transformation is? IREN was a Bitcoin miner. And within a ridiculously short period of time they secured a >5GW announced data center pipeline, additional multi-GW development opportunities and are now preparing to deploy capital on the scale of the largest AI infrastructure companies in the world. CRWV now guides to $35–39B capex for 2026. NBIS guides to $20–25B. IREN: $25–30B in FY27. Just think about that for a second. A former Bitcoin miner is now discussing $30B annual capex like this is a normal Tuesday. 😂 I can understand why a lot of people are scary to invest in $IREN! But obviously the next question is: Who the hell pays for it? More debt? Another massive ATM? More dilution????? This is where the earnings call changed my view quite a bit. IREN says approximately $14B of the FY27 funding requirement is already covered by existing cash + committed GPU financing + customer prepayments. We actually know part of that quite precisely: $7.6B cash at June 30. Of that, $1.7B is restricted cash, mostly reserved for Microsoft GPU capex. The remaining portion of the $14B consists of committed GPU financing and prepayments. Important: I would NOT try to break that remaining ~$6.4B down further because some financing/prepayment cash can already sit on the balance sheet. Otherwise we risk double counting. Then IREN targets another: ~$8B of additional GPU financing + customer prepayments. And here we should also be honest: We DON’T know the split. We don’t know if that becomes $6B financing + $2B prepayments. Or $4B + $4B. Or something completely different. IREN only told us that late-stage customer discussions cover a “significant portion” of 2027 capacity. We don’t know exactly how many MW are already effectively spoken for. So I’m not going to invent numbers. Important detail though: The ~$8B target itself is specifically GPU financing + prepayments. Data center financing, operating cash flow and corporate sources come AFTER that to fund the remaining gap. So at the midpoint: $27.5B capex minus $14B already available / committed minus $8B targeted GPU financing + prepayments = ~$5.5B still needs to be funded. Now it gets interesting. IREN originally put a $6B ATM in place. (yeah, every IREN-shareholder knows about that!) Around $2.5B has already been used. (glad, they didn't use it in the last weeks!) So roughly $3.5B ATM capacity is still sitting there. Personally, I assume they WILL use some of it through this buildout. Maybe a lot of it. I’d rather model dilution than pretend it doesn’t exist. But just for illustration: If IREN eventually used the entire remaining ~$3.5B ATM during late 26 and till mid 27, a midpoint $5.5B residual funding gap would fall to roughly: ~$2B. And that is BEFORE considering meaningful data center financing, future operating cash flow or additional corporate debt! Do you understand? That is a very different picture than, "IREN needs $30B, shareholders are screwed". Now here is the easiest way I found to understand why management thinks this can scale. Imagine one deployment costs $100. Dan explained that GPUs are roughly two thirds of the total cost. So very simplified: ~$67 GPUs / compute ~$33 data center / infrastructure. Recent customers are prepaying roughly 45–55% of GPU capex. Use 50%. That’s ~$33.5 coming from the customer. Then lenders can finance around 90% of GPU capex. 90% of $67: ~$60. So theoretically you can have roughly: $93.5 of funding for a $100 deployment… before even using the underlying data center as collateral. And this is not some fantasy spreadsheet. IREN already says that in some structures customer prepayments + GPU financing provide MORE funding than the associated GPU capex, with the excess helping fund data center capex. And THIS is where I think the comparison with CRWV and NBIS becomes really interesting. Because people post their capex numbers next to each other on X as if all three companies are buying exactly the same thing. They’re not. Think about real estate. CRWV is closer to: “I lease a purpose-built house and put insanely expensive machines inside.” That’s obviously simplified, but CoreWeave has a huge amount of leased data center infrastructure. Its latest filing shows $35.5B of estimated future lease payments for leases not yet commenced alone. NBIS increasingly runs a mix, which I actually really like: Own infrastructure. Lease infrastructure. Asset-backed financing. And now asset-light partnerships where third parties can finance and own the infrastructure while Nebius operates its cloud platform on top. IREN increasingly looks more like: "I secure the land, secure the power, build the substation, build the house, own the house, install the cooling, buy the machines and then rent the compute." Of course that requires more capital upfront. But there is one VERY important difference: IREN OWNS THE HOUSE. And this could become a real structural advantage for many years to come! The land remains. The secured power remains. The substations remain. The buildings remain. The cooling infrastructure remains. The data center remains. These are hard assets. And Dan basically confirmed that the entire data center portfolio is currently unencumbered, including Horizon 1–4. Meaning those assets can potentially become another pool of collateral. That is the part I’m now most interested in. Because if IREN can finance those hard assets at attractive rates and decent LTVs, the flywheel becomes even more powerful: Customer prepays part of GPU capex ↓ GPU lender finances most of the GPUs ↓ IREN builds and owns the data center ↓ The data center itself gets financed seperately ↓ IREN recycles its equity into the next site ↓ Repeat And repeat. And repeat. That is a fundamentally different model from constantly issuing equity for every new GW. And it potentially gives IREN something CRWV doesn’t have to the same degree: a growing balance sheet full of hard infrastructure assets underneath the cloud business. NBIS has more of this than CRWV and I really like their mixed approach too. But IREN is probably the most vertically integrated bet of the three. That increases capex. It increases execution risk. But IF those assets can be efficiently financed… it also becomes an advantage. And that “IF” is important. I don’t know yet whether IREN will get attractive data center financing. Neither do you. The key numbers we need next are: LTV. Interest rate. Term. And ultimately: How much actual IREN equity is required per deployed MW AFTER customer prepayments, GPU financing and data center financing? That could become one of the most important metrics of the entire $IREN thesis. Because if this flywheel really works the way it currently starts to look… FY27 could be the last year where dilution is still one of the defining funding concerns. Not because $IREN suddenly becomes massively FCF positive. They may actually stay deeply FCF negative for years if they keep reinvesting aggressively into new capacity. The key difference is this: Negative FCF does NOT have to mean shareholder-funded growth. If customer prepayments + GPU financing + data center financing + growing operating cash flow can fund most of the buildout, IREN can keep investing tens of billions without equity remaining the primary source of capital. That’s a VERY different setup. But after the huge FY27 buildout, operating ARR and cash generation should be dramatically higher. At some point the business itself starts financing more and more of the next expansion. That is what changed for me after this earnings. The $25–30B number initially scared me. Now I almost see it the opposite way. The crazy thing isn’t that IREN NEEDS $30B. The crazy thing is that a former Bitcoin miner has apparently built enough demand, financing access, secured power and physical infrastructure that management believes it can DEPLOY $25–30B in a single year. And if the hard assets really become the second financing layer Dan is describing… then the $30B capex number may end up being one of the most bullish numbers from the entire earnings. Once again: Dan… why the hell wasn’t THIS a slide in the presentation? 😂 Ah. One last thing. Since IREN Bears are still losing their minds over the Warriors marketing deal… The reported price is >$50M per year. Sounds like a lot. Until you put it next to $25–30 BILLION of FY27 capex. That’s roughly 0.17–0.20%. Or put differently: For every $100 IREN plans to deploy in FY27, roughly 17–20 CENTS go to the Warriors partnership. The infrastructure spend is around 500–600x larger. You can absolutely debate whether the deal is worth $50M/year. I did too. But at the scale IREN is trying to become, I’m honestly not going to lose sleep over 0.2% of annual capex being spent on building a global brand. Especially when your target customers are sitting right in the Bay Area. 😄

@Bonek2801 Excellent write-up 🤗