This is AMAZING NEWS!! The Department of Education will start tying federal student loan eligibility to graduate earning outcomes. Under the rule: > undergrad programs must out-earn high school graduates and > graduate programs must out-earn undergraduate programs This is going dismantle many BS predatory college programs. A would be win for the future of American students!

$IREN Unlike the Goldman Sachs conference, where Dan Roberts declined to speculate on the potential ARR embedded in IREN’s 2027 roadmap, Kent Draper was much more explicit in the @McnallieM interview. He basically handed us the inputs and told us to pull out an envelope: Kent Quote: “We haven't specifically guided to a 2027 ARR target, but all the building blocks are there. We mentioned where we've been contracting today: $23 million a megawatt. In terms of 2027, we have 500 megawatts of IT load expected to come online. We are seeing upward pricing pressure... so it's relatively easy to do some back-of-the-envelope calculations and work out where this could go." For anyone without an envelope handy, let me do the math for you: 500MW-IT of incremental 2027 capacity × $23M of ARR per MW-IT = $11.5B of incremental ARR. The key word is incremental. That ~$11.5B is the annualized revenue potential from the 500MW of new IT capacity expected to come online in 2027. It would sit on top of the $4B+ ARR associated with the 2026 roadmap. Kent also explicitly said pricing is seeing upward pressure, so $23M/MW-IT may not even be the ceiling. No, IREN has not formally guided to $11.5B of incremental 2027 ARR. But this is about as close as management can get without actually putting the number on a slide. And I think Kent’s confidence has less to do with squeezing another couple million dollars of ARR out of each MW in pricing than it does with what I believe is the much bigger embedded sandbag in the 2027 roadmap: PUE. Wait, what? Yes, PUE! One of the most important changes in IREN’s latest roadmap is that management is now talking in IT MW, not just gross facility MW. On the prior earnings call, the 2027 roadmap was roughly 730MW gross. The latest framing is ~500MW-IT. Using the disclosed site mix and reasonable PUE assumptions, the bridge looks roughly like this: NVIDIA 60MW Air: 60MW / 1.37 PUE = ~44MW-IT Childress 190MW Air: 190MW / 1.37 = ~139MW-IT Canal Flats 30MW LC: 30MW / 1.50 = 20MW-IT Horizon 5–6 LC: 150MW / 1.50 = 100MW-IT Sweetwater 1: 300MW / 1.50 = 200MW-IT That gets you to: 730MW gross → ~503MW-IT Which is almost exactly the 500MW-IT management is now guiding to. Here’s the catch: @danroberts0101 acknowledged on the earnings call that IREN has been deliberately conservative with the headline PUE assumptions it uses for these deployments, including the 1.5 PUE currently applied to liquid-cooled capacity. And that matters. I find it very hard to believe Sweetwater 1 (or frankly Horizon 1-4), or any future NVIDIA DSX AI Factory, operates anywhere near a 1.50 PUE. IREN hasn’t disclosed what the actual operating PUE could be, so I’m not going to pretend we know the answer. But for context, NVIDIA’s own Omniverse DSX Blueprint documentation uses a 1.20 PUE base case in its example dashboard. That is a huge difference. Because if gross power stays fixed but actual PUE comes in materially below 1.50, the amount of revenue-generating IT MW goes up. And not by a little. Take Sweetwater 1: 300MW gross / 1.50 PUE = 200MW-IT (guidance) 300MW gross / 1.20 PUE = 250MW-IT That’s an extra 50MW-IT from the exact same 300MW gross power. At today’s contracted pricing of roughly $23M ARR per MW-IT, that incremental 50MW-IT would represent: 50 × $23M = $1.15B of additional ARR Same site. Same gross megawatts. 50 more revenue-producing IT megawatts. Now apply the same math to Horizon 5-6 and Canal Flats. That’s why I think PUE will be one of the most underappreciated upside levers in IREN’s 2027 roadmap. And it doesn’t stop with simply outperforming the conservative design PUE assumptions. IREN has other levers it will pull to unlock even more revenue-generating IT capacity from the same underlying power footprint: 1.Liquid-cooling Prince George and Mackenzie. Those sites are currently constrained by air-cooled infrastructure, which limits how much of the gross power can actually be converted into IT load. Moving more capacity to liquid cooling will allow IREN to push closer to the sites’ gross power ceilings. 2. NVIDIA DSX power-management tools, including Max LPS. These tools will potentially allow them to oversubscribe megawatts because GPU clusters don't constantly pull their max power 100% of the time, simultaneously. And on to my broader point. The 500MW-IT 2027 roadmap may look like a fixed number, but it isn't! If PUE comes in better than the conservative assumptions, existing sites are pushed closer to their gross power limits, and DSX enables better utilization of that power, IREN could end up extracting far more IT MW from essentially the same gross megawatt footprint. At ~$25M of ARR per MW-IT, every additional 10MW-IT is worth roughly $250M of ARR. Yuge! 🤯 So when Kent says the building blocks are already there to form a 2027 ARR target, I think there may be more embedded in those blocks than the headline 500MW-IT suggests. (Great interview Bryce and Anthony!) (holy shit this went longer than I expected, I hope this post makes @franklee6924T proud)

Finally, Bitcoin has been on the move after holding the $60k support zone for almost a year. That’s how long a typical Bitcoin winter lasts, so I’m sensing that a new 4-year cycle bull market is underway. Note that the Z-score of BTC/gold has turned positive after being -100%. In the past that has generally been confirmation of a bottom. What does all of the above suggest? We are in a new secular regime of a higher cost of capital, which suggests that governments will respond with that oldest trick in the book: financial repression.

1/4 I have been trying to explain for a long time why $IREN is playing a different game. Credit to @mcF_dan for the inspiration for this thread. When I first met $IREN co-CEOs Dan and Will in 2019, that was obvious. They didn’t know everything. What I could see was how fast they could learn. That is the rare part. Not a perfect map on day one. A learning rate fast enough to outrun the problems and still land exactly where they needed to be. They built from first principles: secure the scarce physical layer first, keep as much optionality as possible, learn from people who had already been through it, then go further, way further than the starting playbook. That is why the strategy can look slow if you only watch deal-by-deal headlines. In a tightening market, waiting can mean better counterparties and better economics. Don’t get lost in one contract. Watch the whole stack in the age of AI.

Before you say it’s not “fair” that you’re poor, answer these questions: 1. Do you save half of your income? For my entire life, even when I was mowing grass for a living, I saved half of my after-tax income. How? I didn’t have my own place. I didn’t have a TV. I didn’t have a cell phone. I drove a 1983 Honda Accord with 225,000 miles on it—a car I bought from a friend for $2,000 when I was 19. I didn’t have a credit card until 2001, when I was 29. I didn’t have my own place to live until I was making more than $1 million a year. I bought my first stock—Coca-Cola—in 1992, when I was 21. I bought my second stock—Amazon—in 1997, when I was 26. Whenever I had excess liquidity in my checking account, I bought gold coins—almost every year—and I’ve never sold a single one. I’m not a slave. I don’t save their worthless money. 2. Do you work every day of the week, at least 12 hours a day? For the first decade of my career, from 1996 through 2007, I worked seven days a week, from 7 a.m. to 7 p.m. That didn’t include the reading I did when I got home. I started taking weekends off after I had my first child. Even today, I still work seven days a week. I write every day. This year, I’ve written a daily e-letter five days a week. I write two investment newsletters each month, and I’ve written two best-selling books. I run a business with more than 100,000 customers and roughly 30 employees. I also recently had my fourth child and am building a new house in Winter Park, Florida. I cook dinner for my family every night. That’s my hobby. 3. Did you get an education? In America, anyone can become almost anything they want professionally in about seven years. If you don’t have valuable skills, go acquire them. I got my first job in financial research by agreeing to work for free. I didn’t have a degree in finance or accounting, but learning the necessary skills wasn’t hard. I used the library. I read every letter Buffett ever wrote. I read Rothbard, Hayek, von Mises, Bastiat, and thousands of SEC filings. Use the library. Find a mentor. Today, it’s even easier because of the internet. There’s no question that if you save half your income, work hard, and educate yourself, you can become financially independent. Is it hard? Of course. But it isn’t as hard as being poor and ignorant. Here’s a big hint: No one can tell you how to succeed. And if you believe what they tell you, you will end up with only what they give you—nothing more. You have to find your own path. If you do what everyone else does, you’re only going to get what everyone else gets. I’m sure you can imagine what my friends and family said when I told them, after I got out of college, “I’m going to build the most respected financial advisory firm in the world over the next decade.” I’m sure they thought I was nuts. Finally, if you have a good income but never seem to build wealth, learn to use credit the right way. I constantly use low-cost credit to fund assets that appreciate—farmland, securities, my business, and real estate. Poor people use credit to buy things. Rich people use credit to buy assets. Remember what Ayn Rand said about the poor: “Don’t be one of them.”

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

March 2020. 👇 Six years later, the vision is very much alive. @MichaelDell 👊

As someone who’s been through the peaks and doldrums of the Bitcoin market for 10 years, this seems like a great time to take a position/expand or DCA. A bunch of metrics based investors will tell you that because vol has declined, bitcoin may not have eccentric and overheated bull markets going forward. I disagree. In each cycle since 2020 I’ve felt that new highs would undershoot expectations simply because retail expected it. Now that expectations are tempered this is the first time I’ve felt we have the potential to wildly overshoot any expectation. Also, the fiscal debt situation is finally reaching a point of no return that bitcoiners have hypothesized for many years. The Fourth Turning will almost certainly include Warsh/Bessent at war with the long end of the curve. Not many understand the implications of such a thing. Lock in, fren. Game time is almost here. Will delete if wrong. bitcoin:native