bayu

[Tyler Page @rftylerpage]

I think you are oversimplifying the concept of “spot” rates for leasing. Generally, tenants are very interested in powered shells available in the near-term and become less interested in ones available further into the future. This relative level of interest is reflected in what they are willing to pay in rent. Leases that are being signed and announced now are typically for sites available 12-18 months out, and they are being signed at the highest rates we have seen. But if you ask a hyperscaler about what they would commit to pay right now for a long-term lease for a site available in 2030+ (if they would even be interested now), they might offer to pay 7-8% YOC. At those levels, we would not commit to a lease and will wait if we have such a site available.

The binding commitment we received here is for a lease starting 10 years from now at a rate in-line with today’s elevated market rates for powered shells available in the near-term from a very large and successful leading AI lab. This is a big deal and evidence of a long-term sustained demand environment in my opinion.

I saw a few other questions along the lines of “Why sign now? Why not wait?” Presumably this logic is based on the idea that we should just wait 8.5 years until interest will be higher. My answer is that we have never seen interest at today’s elevated rates for a lease starting that far in the future and it provides an excellent projected IRR on our investment. Also, if we follow that advice, why should we ever sign a lease at all? Just keep waiting for infinitely higher lease rates. I don’t believe that is a winning strategy.

Lastly, keep in mind that headline lease rates don’t tell the full story in an inflationary environment with fast-changing design and engineering standards. You need to consider other elements that can control your costs and financing risks. We are focused on maximizing a return on our investment while managing risks. (Note that this is also why revenue comparisons with neoclouds are somewhat irrelevant—they have a completely different cost and risk structure)

[bayu @bayumifp]

@danroberts0101 Just how many more times do we have to lose our momentum, Dan?

[bayu @bayumifp]

@IREN_Ltd @kentpdraper Anyone can say anything; let's have an official announcement.

[bayu @bayumifp]

@Joseph_Invests Psychological operations seem very important, but Dan isn't good at that.

[bayu @bayumifp]

@Joseph_Invests I think Dan Robert is being far too cautious—to the point where it looks sluggish. The real issue is that everything still seems overly convoluted and overextended, requiring dilution; plus, that’s not really the kind of approach people take in today’s long-term market.

[Tyler Page @rftylerpage]

Very excited to announce this deal. We have been advocating for locations in west Texas for a long time. The latest incorrect thinking from traditional incumbents is that these data centers are a one-time thing in a historically frothy market and that they will not have use at the end of the leases as tenants seek to return to today's "Tier 1" locations. Well, we just debunked that myth by signing a lease that starts 10 years forward, and we are projecting a strong ROI on our upfront equity investment. Everyone needs to upgrade their terminal value assumptions. Texas data centers will trade at a premium to the rest of the market in the future and be the new "Tier 1". $CIFR

[bayu @bayumifp]

@danroberts0101 Do you know how to let go and relax a little? Just focus on what's right in front of you for now, @danroberts0101

[bayu @bayumifp]

@jiahanjimliu Let me tell you, you were wrong not to anticipate IREN hitting 100 this year. In the current market environment, if the company is solid enough to announce positive news repeatedly, it will easily reach that level. @jiahanjimliu

[bayu @bayumifp]

@danroberts0101 How much longer can you afford to watch other stocks keep hitting all-time highs while yours stays stagnant?

[bayu @bayumifp]

@FransBakker9812 I would like to see genuine protection of shareholders, at least once. @FransBakker9812

[bayu @bayumifp]

@StockSavvyShay Announcement deal $iren

[bayu @bayumifp]

@DustinHuntwn 100-120 eoy

[Chamath Palihapitiya @chamath]

Prediction: within 12 months, top three models will be open source.

Economic winners will be the American clouds that serve them:

Nebius
Iren
Baseten
Together
Fireworks

[Daniel Roberts @danroberts0101]

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 about how fast AI should be allowed to improve is a fair one to have. It's about future generations of models.

Existing demand is the part I think people are misreading. Anthropic CEO Dario Amodei said in May they'd planned for 10x growth and were running at an 80x pace in the first quarter, and that's why they've had trouble supplying compute. OpenAI president Greg Brockman said in July they'll be in a compute shortage no matter what, and are choosing which products to scale. Google says it's processing 7x the tokens it did a year ago. Hundreds of millions of people use these models today, and most of them use a small part of what the models can already do. And every time more compute comes online, usage steps up again: limits come off for people already using it, customers who were turned away get on, and new use cases show up that weren't in anyone's plan.

Now supply. The constraint is HBM, the memory that sits inside every major AI chip. Three companies make it and all three are sold out this year. A new memory plant takes years to build. TrendForce has HBM shipments growing 50-60% next year. NVIDIA, the biggest buyer of it, expects its revenue to grow about 70% next year and calls that outlook 'supply-constrained', noting its customers' forecasts point closer to 100%. On our own bottom-up work, the memory constraint lands in about the same place as NVIDIA's growth number.

Then the chips need a building with power connected, which takes longer again. Goldmans reckons only about half the US capacity scheduled over the next two years will actually be built on time.

The risk to demand continues to seem heavily weighted to the upside. The risk to supply continues to seem weighted toward less capacity getting built, not more.

[bayu @bayumifp]

@FransBakker9812 Wow 150$ come fast!

[bayu @bayumifp]

@danroberts0101 Will you have people keep buying other stocks in the same industry, or will you do something amazing right now?

[bayu @bayumifp]

@danroberts0101 In the current market, IREN is hovering unreliably within the $30–$60 range; it is no longer the same $IREN that traded at $3 last year. No one cares about the past.

[bayu @bayumifp]

@danroberts0101 A CEO is like a father figure; the most important thing is to protect everything, not just a part of it.