The Count of Monty Bitcoin

Bitcoin, miners and AI/HPC Proud Mining Mafia member

[Rosanna Prestia, MBA @RosannaInvests]

$IREN heading into next week. 👀

The setup: the stock took a headline "miss," bled for four sessions, then reversed 19% off the low with volume expanding on the up days. That's what it looks like when the sellers who wanted out are out. Volume confirming direction is the one technical signal I trust. ✅

The business underneath didn't change: $1B of ARR operating today, $4B contracted for year end, three sites commissioning between now and December, new contracts signing above $20M per MW, and ~$19B of funding secured with only ~$3B of it equity. Dilution went from mandatory to optional. ✅

And the valuation math still holds. MC per dollar of next-year capex: $NBIS roughly $2.50, $CRWV roughly $1.25, $IREN roughly $.60. Same buildout, the cheapest ticket in the sector by a mile. ✅

Most contracted, cheapest per buildout dollar, and a chart that just told you where the buyers live. ✅

Bonus: Cramer prefers $CRWV over $IREN 😎

Long $IREN

[The Count of Monty Bitcoin @CtMontyCrypto]

@f @peterkirn Check out Koala... lots of fun with an ipad: sample, edit and sequence... easy-peasy and sounds great!

[Perle @veritebeaute]

Quand Albert Einstein donnait une conférence dans les nombreuses universités des États-Unis, la question récurrente que lui faisaient les étudiants était :

Vous, Monsieur Einstein... Croyez-vous en Dieu ?

Ce à quoi il répondait toujours :
- Je crois au Dieu de Spinoza.
Seuls ceux qui avait lu Spinoza comprenaient ...

Spinoza avait passé sa vie a étudier les livres saints et la philosophie, un jour il écrivit :

Je ne sais pas si Dieu a réellement parlé mais s'il le faisait, voici ce que je crois qu'il dirait au croyant :

Arrête de prier et de te frapper à la poitrine !
Ce que je veux que tu fasses, c'est que tu sortes dans le monde pour profiter de ta vie.
Je veux que tu t'amuses, que tu chantes, que tu t'instruises... que tu profites de tout ce que j'ai fait pour toi.

Arrête d'aller dans ces temples sombres et froids que tu as construit toi-même et dont tu dis que c'est ma maison !
Ma maison est dans les montagnes, dans les bois, les rivières, les lacs.
C'est là où je vis avec toi et que j'exprime mon amour pour toi.

Arrête de m'accuser de ta vie misérable,
Je ne t'ai jamais dit qu'il y avait quelque chose de mal en toi, que tu étais un pécheur, que ta sexualité ou ta joie étaient une mauvaise chose !

Alors ne me blâme pas pour tout ce qu'ils t'ont dit de croire.
Arrête de ressasser des lectures sacrées qui n'ont rien à voir avec moi.
Si tu ne peux pas me lire à l'aube, dans un paysage, dans le regard de ton ami, de ta femme, de ton homme, dans les yeux de ton fils...Tu ne me trouveras pas dans un livre !

Arrête de te faire peur.
Je ne te juge pas, je ne te critique pas, je ne rentre pas en colère et je ne punis pas.
Je suis pur amour... je t'ai rempli de passions, de limitations, de plaisirs, de sentiments, de besoins, d'incohérences...et je t'ai donné le libre arbitre...

Comment puis-je te blâmer si tu réponds à quelque chose que j'ai mis en toi ?
Comment puis-je te punir d'être ce que tu es, si je suis celui qui t'ai fait ?
Tu penses réellement que je pourrais créer un endroit pour brûler tous mes enfants qui se comportent mal, pour le reste de l'éternité ?
Quel genre de Dieu peut faire ça ?
Si j'étais ainsi, je ne mériterais pas d'être respecté.
Si je voulais juste être vénéré, je n'aurais peuplé la terre que de chiens. ..
Respecte tes semblables et ne fais pas ce que tu ne veux pas pour toi.
Tout ce que je te demande, c'est que tu fasses attention à ta vie, que ton libre arbitre soit ton guide.

Toi et la nature vous constituez une seule entité ....alors ne crois pas que tu as un pouvoir sur elle.
Tu fais partie d'elle.
Prends-soin d’elle et elle prendra soin de toi. J'y ai mis et rendu accessible tout ce qu'il y a de bien pour toi et j'ai rendu difficile d'accès ce qui ne l'est pas.
Ne mets pas ton génie à y chercher ce qui est mauvais pour cet équilibre.
A toi de garder intact cet équilibre.
La nature elle, sait très bien le garder, juste ne la trouble pas !
Je t'ai rendu absolument libre.
Tu es absolument libre de créer dans ta vie un paradis ou un enfer.
Je ne peux pas te dire s'il y a quelque chose après cette vie, mais je peux te donner un conseil,
Arrête de croire en moi de cette façon,
Croire, c'est supposer, deviner, imaginer.
Je ne veux pas que tu crois en moi, je veux que tu me sentes en toi.
Que tu me sentes en toi quand tu t'occupes de tes moutons, quand tu abordes ta petite fille, quand tu caresses ton chien, quand tu te baignes dans la rivière....
Exprime ta joie et habitue-toi à prendre juste ce dont tu as besoin !

La seule chose sûre, c'est que tu es là, que tu es vivant, que ce monde est plein de merveilles...et que dans toutes ces merveilles tu es capable de savoir exactement ce dont tu as vraiment besoin.

Ne me cherche pas en dehors,
Tu ne me trouveras pas....
Je suis là... La nature,
Le cosmos... C'est moi.

Extrait de Baruch Spinozza

[The Count of Monty Bitcoin @CtMontyCrypto]

@bitcoinbutcher1 @StockSavvyShay This is insane value territory. I took profits on a chunk of BTC after the rally and poured it into IREN and IREG on Friday.

[Stern Drew @SternDrewCrypto]

🚨 THE MOST DANGEROUS SENTENCE IN MARKETS JUST CAME FROM JAPAN

BoJ’s Yuto didn’t speculate. He told you the plan.

“How convenient it would be if they could delay the collapse of the current economic order until it coincides with a global crisis. AI. Quantum. Energy shock. Pick one. Our ‘allies’ are already hunting for the perfect scapegoat.”

Read that again.

Japan was about to end its era of free money and initiate the catastrophic carry trade unwind. The BoJ and Finance Minister threatened of bold actions and Yuto even apologized for the measure prepared.

Then boom…. U.S. Treasury Secretary directly intervened and took over BoJ operations. He stopped Japan from dumping their $1.4 Trillion Treasury holdings and instead asked the FED to lend Japan U.S. dollars to dump it and try to save the yen.

IT ALL FAILED. Yen is falling back to pre-intervention levels.

Rates are surging while the yen is falling.

The whole point of intervening in the yen was to ease inflationary pressure and stop rates from surging.

Japan’s 2 year government bond auction today saw almost no demand. Last week, foreign institutions dumped short-dated JGBs at a record pace.

Both sides know the current order is cracking. They’re stretching it just long enough to attach the collapse to a bigger story… AI disruption, energy shock, quantum breakthrough, anything except the debt and the yield spiral itself.

That’s why every major central bank on earth is stacking gold like the building is on fire.

This was exactly warned by the famous City of London banker @LordBelgrave at the start of the year.

The delay is the strategy. The scapegoat is already being written.

Yuto just said it out loud.

[StockAnalystPro @StockAnalystPro]

https://t.co/eOXLDEFhhh

[Long_Vol_(₿) @planA_thru_Z]

The chart below compares enterprise valuations to the weighted average ARR for power that has been monetized for AI compute. I had to make assumptions regarding PUE and MWs associated with some contracts given incomplete info, but it’s clear $IREN is in deep value territory.👇

[The Count of Monty Bitcoin @CtMontyCrypto]

@growthrapidly STRONG BUY ! Added $IREN heavily throughout the day Friday (my first purchase since it was $9.50 in May25) increasing by holdings by over 20%, and when it hit $35 I added some leverage $IREG as a short term play to capitalize on what should be a hard bounce next week.

[Rohit @rohitboolc]

One business demolished on purpose. One pre-leased. Guess which made the headlines. https://t.co/tXflvB1MKQ

[The Count of Monty Bitcoin @CtMontyCrypto]

@felixprehn Dollar please!

[The Count of Monty Bitcoin @CtMontyCrypto]

@moninvestor Besides the hope for new deals and other surprises, I'll be most interested in the amount and depth of the guidance we get. Since they stopped providing monthly updates last year Dan has been tight lipped in ECs and it has not served them well. I'm hoping to see a change.

[xenomecha @xenomecha]

Introducing my Fibonacci Masterclass.

Everything I know about the Fibonacci tool for charting stocks, in one video.

Arguably the most elaborate Fibonacci Retracement video on the internet. Don't miss it.

Available in 4K 60fps.

Timestamps:
00:00 – Introduction
00:50 – Numbers
03:07 – Retracement vs Extension
04:01 – Linear vs Log
06:03 – Creating a Template
13:00 – Drawing Fibs Correctly
23:35 – Profit-taking Using Extensions
29:29 – Confluence Stacking
37:00 – Closing

If this helps you, please share it with a friend and follow for more.

[MindArch @mindarchx]

In 1986 a guy got kicked out of every casino in Vegas for counting cards. So he flew to Hong Kong with $180,000 and started betting on horses instead. He walked away with almost $900 million.

It's Bill Benter. He figured horse racing was just another counting problem. Same math, more moving parts.

He and a partner showed up with $180k and a computer. Benter spent years teaching that computer to guess one thing, the real chance each horse had to win. If his number was better than the odds the bookies gave, he bet. If not, he skipped it.

That's the whole trick. Expected value.

EV = p · b − (1 − p)

Only bet when your win chance p, at odds b, is worth more than your chance of losing.

This recording was never meant to be some hidden gem. Nobody expected Professor Tsitsiklis to hand the whole foundation away in 45 minutes, but that's exactly what happens on the board. Students in that room pay over $80,000 a year to sit through it. It's free right here. It's free right here.

Every quant, every professional bettor, every hedge fund analyst started with this exact hour. Benter just watched it and actually did the homework.

Almost nobody knows this lecture even exists. Watch it before it gets taken down.

The answer is in this video.

[Music Live @MusicLive__]

#ClassicsMTV

[franklee6924x @franklee6924T]

The Life-or-Death Red Line of the Neo-Cloud Industry — The Most Important Thing $IREN Is Doing

The neo-cloud industry is one of the most promising sectors today, while at the same time imposing extremely demanding requirements on the financial quality of the companies involved. To remain competitive in the coming boom of the AI industry, companies must directly confront the unavoidable problem of heavy capital investment. Faced with this choice, emerging companies and established companies with strong resources are confronted with essentially the same dilemma. This is also why the industry presents investors with such a significant opportunity. AI's appetite for compute is almost unlimited. There are very few industries in which companies can remain unconcerned about demand over such a long period of time. The real challenge is to continuously provide high-quality compute while simultaneously making the company's financial condition increasingly healthy.

The balance between capital expenditure and revenue is a red line. Although there are many ways to assess the quality and development of the AI infrastructure industry, this red line is a matter of survival and is therefore critically important. More specifically, it is the question of establishing a positive return loop between capital investment and cash flow. It can be said that whoever gains an early advantage on this front is the one most likely to truly own the future.

At present, CoreWeave is performing the worst when it comes to this red line. Its enormous business is built on an expensive debt foundation. Although its backlog continues to rise and it is making active efforts in other areas, its ability to withstand risk is extremely fragile. So far, there has been no clear sign of a fundamental reversal.

Oracle originally had a very healthy business. However, because it has expanded too aggressively, its free cash flow has been negative by more than $20 billion over the past four quarters, and this shortfall could potentially widen to $42 billion in the next fiscal year. Although it has healthy cash-flow-generating businesses to support it, the direction of development is dangerous, resulting in a deterioration in its credit rating. Its rating is now only one notch above junk status.

Nebius appears to be in a much more comfortable position on this issue. The market believes that it has limited dilution and strong execution. In reality, however, its difficulties are only just beginning. It has effectively sold off most of the compute capacity it can control at low prices. Its revenue visibility over the next two years is essentially locked in, but the uncertainty surrounding its expenditures has increased substantially.

First, this comes from its positioning as a full-stack technology company. In order to remain competitive, it has recently acquired two software technology companies. Both are positioned to be unable to contribute meaningful revenue for a considerable period of time, while continuing to require substantial funding to develop. It is also reportedly considering the acquisition of an Israeli unicorn, which would represent another major expenditure. These companies are certainly promising, imaginative, and positioned at the technological frontier, but they are all still at stages where they require sustained financial support.

Even more concerning is that the compute supply underlying NBIS's contracted commitments is itself subject to a much greater degree of uncertainty: opposition from residents and communities, increasingly stringent regulation, the execution capabilities of third parties, an increasingly complex supply chain, the maturity of new technologies, and so on. Uncertainty surrounding these factors is inevitable. The market is currently significantly underestimating these multiple sources of uncertainty. The execution discount they create is multiplicative, and ultimately, a grand and enormous set of future expectations can be continuously fragmented and dismantled by all kinds of seemingly minor execution problems.

This week's Nebius convertible bond issuance served as a wake-up call for the market. Within less than a year, the company has issued convertible bonds three times, totaling $9.7 billion, all based on expectations of extremely strong execution. The coupon has increased with each issuance. In the latest convertible issuance, $2 billion carries a maturity coupon of as much as 4.5%, while also carrying an additional 125% maturity accreted principal. This indicates that the bond market's confidence in the company's ability to execute successfully is continuing to decline. Across the $9.7 billion of convertible bonds, maturity coupons range from 0.5% to 4.5%, with the trend clearly moving upward.

An even bigger issue is that all three rounds, covering four separate notes, include maturity accreted principal provisions, without exception. The only difference is the percentage, ranging from 110% to 125%. In other words, every NBIS convertible bond contains a form of tail liability under which, if the bonds are not converted at maturity, the company must repay an additional 10%–25% in cash.

Overall, Nebius has better financing terms than CoreWeave. CoreWeave's borrowing costs can reach as high as 9%. However, compared with the top student in the class, Nebius still has a very large gap. Nebius's biggest risk is that it has created a chain-reaction architecture that demands an extremely high level of execution, leaving virtually no room for error. If any key element encounters a problem, it could trigger a chain reaction throughout the entire structure.

From this perspective, Nebius may actually be more dangerous than CoreWeave. CoreWeave's allocation of resources is relatively concentrated, and it has also made corresponding financial preparations, such as securing relatively broad access to capital-market financing. At worst, its share price performs poorly and dilution becomes severe. In my personal view, Nebius's risk-management preparations are not yet sufficiently secure.

The top student is $IREN.

Over the past year or so, IREN has issued a total of $6.05 billion in convertible bonds. There have been no exchange agreements and no high-interest compensation provisions. Approximately 96% of its convertible bonds carry coupons largely within the 0.25%–1.00% range. Only about $233 million of the convertible debt carries a 3.5% coupon, and even that resulted from IREN's proactive decision to use equity financing to repurchase higher-cost legacy debt. This was an active compensation measure, demonstrating that IREN has considerable control over the cost of its convertible financing.

At the same time, the bond market's confidence in IREN's execution certainty is extremely high. Several tranches of its convertibles even carry a 0% coupon and do not require maturity accreted principal compensation. In effect, the money is being borrowed for almost nothing, with investors primarily receiving the right to convert into equity. Compared with NBIS, ORCL, and CRWV, the nature of the risk is simply not on the same level.

Over the past year or so, IREN's extensive use of ATM offerings has attracted emotionally charged criticism, with some arguing that these transactions placed the company's survival and debt reduction ahead of long-term shareholder value. But this is precisely one of IREN's most important initiatives for building a high-quality credit system and establishing a high-growth flywheel between capital investment and cash generation.

When evaluating the trade-off between financing methods and dilution, the key issue is the quality and structure of the assets corresponding to those financing methods.

The differences among Nebius, Oracle, CoreWeave, and IREN can ultimately be traced back to how they convert capital into sustainable assets, and whether those assets can continue to support lower capital costs in the next round of financing.

Nebius is positioned as a full-stack technology company, which is an extremely competitive field with a very high risk of commoditization. It wants to strengthen its capabilities in infrastructure and heavy assets, but in practice it has discovered that this window of opportunity has already closed, forcing it to turn toward a lighter-asset horizontal alliance model. The assets it is building cannot themselves form a collateralizable, compounding financial structure. Using GPUs as collateral is far from sufficient, and the credit trend associated with GPU collateral is gradually deteriorating.

Oracle relies on pure credit financing to support enormous capital expenditures. It has healthy cash-flow-generating businesses, but these capital expenditures are highly concentrated around a single customer, OpenAI, which accounts for nearly half of its remaining performance obligations. Its collateral is entirely dependent on the market's pricing of the company's historical creditworthiness. When the substitutability of the underlying assets is constrained by customer concentration, credit spreads will continue to widen.

Oracle's asset structure is therefore becoming increasingly "customer-defined": its capacity is customized, its cash flows are tied to a specific customer, and its bargaining power is one-sided. It is not building collateralizable assets; it is building long-term dependence on a single customer. That dependence continuously dilutes its credit quality.

CoreWeave's situation is even more awkward. Forget about establishing a growth flywheel—if its 9% borrowing costs are not fundamentally improved, even the sustainability of its growth becomes questionable. The probability of such a company delivering high and attractive long-term returns is very low.

Finally, let's focus on IREN and examine what IREN believes is most important—and what it has done about it.

IREN's capital structure has been built in the opposite direction. In its early stages, before it had collateralizable assets, it repeatedly used ATM equity offerings as a bridging tool, obtaining the capital required for expansion at the lowest-friction cost.

The marginal cost of ATM financing is dilution alone. There is no interest, no conversion provision, and no fixed repayment obligation. It is therefore a form of "pure equity" transitional financing.

The key point is that IREN used this capital not to purchase short-term capacity, but to acquire fully owned physical assets such as power, land, and liquid-cooled data centers. These assets possess natural financial characteristics: they can enter the ABF market and be recognized by rating agencies as collateralizable infrastructure assets, thereby gaining access to investment-grade ratings and extremely low capital costs.

Once ABF is in place, the balance sheet gains operating, cash-flow-generating physical assets. Those assets can then support the next round of refinancing at a lower cost.

In this process, ATM is the bridge; ABF is the flywheel.

The ultimate purpose of issuing ATM shares is to completely close off the path toward future dilution. Over the past several years, IREN has used continuous ATM financing to build a large base of fully owned physical assets, giving IREN strategic options for its commercialization path.

With the delivery of Horizon 1, the nature of these assets is undergoing a qualitative transformation. Relatively simple physical assets are evolving into highly valuable compute infrastructure assets. The driving force behind this entire process is IREN, together with a group of the most important players in the AI industry.

IREN owns the entire asset base. It has already earned its place in the ABF flywheel. The next step is simply to make it bigger, stronger, and more tangible.

IREN is currently the only company in the neo-cloud industry to have successfully issued ABF financing. On June 1, 2026, IREN officially announced the completion of $3.65 billion in investment-grade GPU financing to support the delivery of its AI cloud contract with Microsoft.

In its public materials, the company repeatedly emphasized two points: this was the first transaction of its kind in the U.S. private placement market, and it was the highest-rated investment-grade GPU financing publicly disclosed at the time.

Structurally, the financing consists of $2.1 billion of U.S. private-placement fixed-rate notes and $1.55 billion of delayed-draw term loans. The former carries an interest rate of SOFR + 2.13%, while the latter carries SOFR + 2.25%. Interest-rate hedges have been used to lock in the cost, keeping the overall debt cost at 6.00%.

When the $1.94 billion upfront payment provided by Microsoft under the five-year contract is taken into account, the all-in financing cost falls further to 3.31%. The financing and upfront payment together cover approximately 96% of the $5.81 billion of GPU capital expenditures required under the contract.

On the ratings side, Fitch assigned an A rating, while DBRS assigned A (low). There are other GPU-backed loans or ABS transactions in the market from other cloud providers, but none simultaneously carry both an investment-grade rating and the distinction of being the first transaction of its kind in the private-placement market. This demonstrates the scarcity value of IREN's financing structure.

From a strategic perspective, the value of this transaction goes far beyond the financing amount itself.

First, it represents a structural breakthrough in the cost of capital. An all-in cost of 3.31% is far below the industry's generally observed 8%–12% range for GPU financing. This means IREN has successfully converted customer credit into a balance-sheet advantage, and that advantage is supported by the quality of the contract and IREN's owned data-center assets rather than by a leasing model.

Second, the signaling value is extremely high. For potential large customers in the future, IREN has now received the backing of major rating agencies as well as major banks such as Goldman Sachs and JPMorgan. This demonstrates that IREN has the ability to raise inexpensive capital and deliver on time, significantly reducing due-diligence friction for future large contracts.

One point should be emphasized, however: the credit foundation of this structure exists because it is tied to the Microsoft contract. The anchoring effect of Microsoft's AAA-rated contractual cash flows provides the credit support. As a growth company, IREN is not yet capable of independently obtaining this level of rating.

This is therefore not an improvement in IREN's standalone credit quality, but rather a single-customer-driven leverage amplifier.

Nevertheless, it opens the door to a very powerful approach and demonstrates that IREN deeply understands both the importance of establishing high-grade credit and the mechanics required to do so. In the future, this structure can certainly be replicated across contracts with other customers.

The immediate priority is to execute the Microsoft contract exceptionally well. Once that happens, the quality of the next contract and the underlying asset structure will have a foundation capable of satisfying rating-agency requirements again. Once there is a first transaction, there can be a second, then a third. At that point, this extremely important flywheel will have been established.

It requires patience—and flawless execution.

The value of this approach is enormous because it can evolve into a variety of powerful financial instruments with strong credit characteristics, such as selling future compute capacity with the underlying compute infrastructure serving as the "physical anchor."

Driving this structured-finance capability is IREN's management team, with its deep financial background, together with a growing number of senior professionals specializing in financial structuring.

The arrival of talent from KKR will play an important role in helping IREN transform physical assets into financial instruments that institutional capital can purchase. Through securitization and layered financing structures, these assets can obtain high ratings and gain access, under the right conditions, to long-term institutional capital.

At the same time, as IREN's integration with NVIDIA's business continues to deepen—from flagship AI factories built around the DSX architecture, to Exemplar Cloud qualification, to Microsoft's acceptance and prepayment structure—the assets become not only collateralizable, but also sustainable.

Its credit quality continues to improve, ultimately creating a high-quality growth flywheel capable of supporting the massive build-out of AI infrastructure.

This is the most important thing IREN is doing right now.

[Music over time🎻 @Peal8888]

Still in Love Under the Harvest Moon

The first time I hit play on the video, I only meant to get through it once. Then AJ Lee’s voice rose—clear, unhurried, almost floating—and everything slowed down. Warm yellow light filled the small room. A mandolin rang out soft and bright. The Morrison brothers’ harmonies settled around her like an old quilt. I sat completely still. By the time the last note faded, my eyes were wet. Now every time I return to it, that same quiet feeling comes back—not dramatic, just familiar and steady.

Neil Young wrote “Harvest Moon” in 1992 for Pegi, the woman he had already loved for nearly two decades. On the Neil Young Archives he said it plainly: “Harvest Moon is a song I wrote for Pegi, my wife of many years, who gave me two beautiful children and helped bring up my first child Zeke. She was a dancer and floated around when she was happy.” They met in 1974 at a little place called Alex’s on Skyline Boulevard, near his California ranch. He used to order food just to watch her walk across the floor—the same image that opens “Unknown Legend.” She had lived in a tent, kept a dog, ridden a motorcycle through the desert. Together they raised children with special needs and co-founded the Bridge School. The song isn’t about the first spark. It’s about what remains after years of real life: a quiet invitation to step closer, to dance again under the harvest moon.

Young’s original 1992 recording has a beauty all its own. His voice is raspy and warm, slow and intimate, like a private conversation. His acoustic guitar sits beside Ben Keith’s sweet pedal steel, the soft pulse of pump organ and banjo-guitar, and those luminous harmonies from Linda Ronstadt. No heavy drums, no flashy solos—just wide-open space and gentle restraint. Young once called Harvest Moon “the quietest record I’ve ever made,” and the title track embodies that perfectly. The moonlight in the song isn’t dramatic. It’s just bright enough to see the face of the person who’s been beside you for a long time. The moon appears in more than twenty of Young’s songs. He once said that before organized religion there was the moon, and he’s followed it in his own way ever since.

Thirty years later, AJ Lee and The Brothers Comatose sat down with acoustic instruments in a warm room and played the song again. They didn’t try to reinvent the melody or the structure. They simply let it breathe through their own hands and voices. Lee’s clear, floating vocal brings a different kind of lightness than Young’s weathered tone. Mandolin takes the place of pedal steel. The Morrison brothers’ harmonies feel warmer and closer, carrying a gentle bluegrass/string-band feel. The cover keeps the same restraint, but adds the intimacy of a small room, yellow light, and easy smiles at the end. When Lee sings “I’m still in love with you… on this harvest moon,” the line still carries the weight of the diner, the ranch, the children, the long road, and the dancer who floated when she was happy—only now it arrives in a newer, younger voice that somehow feels just as deep.

After the video ends I usually sit quiet for a minute. The mandolin still rings in my head, along with those relaxed smiles. I don’t need any more explanation. I just know I’ll keep returning to both versions—Young’s original with its wide, quiet depth, and this cover with its warm, close-up glow—whenever I need a few minutes under the same soft harvest light. The kind that asks for nothing, only reminds you that some loves keep shining long after the louder years have passed.

[franklee6924x @franklee6924T]

HORIZON 1 — The Game-Changing Weight on the Scale Between Speed and Slowness

This week marks a major milestone for $IREN. Horizon 1 was officially delivered nine months after the contract was signed, while also receiving dual validation from Microsoft and NVIDIA. For $IREN, this represents the transition from 0 to 1—the most difficult stage, with the greatest uncertainty and the largest number of components requiring validation. The workload involved in modularization and standardization is also at its highest at this stage. Very soon, Horizon 2–4 will follow in succession, continuing to refine and complete the work of moving from 0 to 1. Once IREN reaches Horizon 5–6, however, it will take on a special significance and could have a powerful impact across the entire AI industry.

The reason is that, as a standalone 50MW data center, the technical difficulty, complexity, and moat effect are relatively limited. But as the total scale continues to increase—especially when moving toward GW-scale AI factories—the technical difficulty and complexity do not simply increase as 1+1; in many areas, they rise exponentially. This is why IREN needs Dell and Lenovo to participate in the design and construction: to overcome technical challenges together through extensive practical experience. This will be highly significant for the next-stage SW1 DSX flagship AI factory. At present, the market is paying very little attention to the technical capabilities required to build AI factories, but very soon this will become a key factor determining the winners and losers. In particular, the technological leap that could result from combining modular liquid-cooling systems with systems such as the 750-mile fiber network deployed across the Childress campus is something that is extremely worth watching—and something the market has so far paid far too little attention to.

In the design and construction of AI factories, IREN is not fighting alone; it is working collaboratively, partnering with multiple top-tier technology companies. NVIDIA, Dell, and Lenovo are working together with IREN to overcome these challenges. IREN is doing this work on physical hardware assets that it owns outright, and the long-term benefits could be enormous. @jimjiahualiu argued in a recent excellent analysis that NVIDIA is taking on the most difficult work. I do not completely agree. From an engineering perspective, IREN is responsible for the systemic engineering integration, and the difficulty of actually executing this work is extremely high. The two sides have different areas of emphasis when it comes to overcoming technical challenges within their respective roles. IREN COO KENT addressed this issue in an interview in June. IREN has taken control of all operational links through vertical integration, while strengthening cooperation wherever specialized expertise is required. In engineering, IREN plays the role of overall coordinator and operator. Through this model, cooperation with leading mainstream technology companies will ultimately produce a flagship technology platform that balances applicability, efficiency, and room for future development.

$IREN possesses multiple unique advantages, which is why it has become the sole underlying physical-asset partner for this undertaking. Given that it has now been three months since this collaboration was announced, and NVIDIA has not said that it will work with a second new cloud company to build another DSX-based AI factory, it is essentially safe to conclude that there will not be a second one. Otherwise, the meaning of the word “flagship” itself would have to be rewritten. A flagship is, by definition, independent and unique.

The biggest development in the AI industry recently has been NVIDIA’s proactive transformation. It is beginning to explicitly position itself as an infrastructure provider rather than simply the chip manufacturer it was in the past. I wonder if everyone has noticed that during every major discussion at IREN’s RAISE SUMMIT, NVIDIA participated alongside IREN, and each time it emphasized its new positioning. This new positioning has also been repeatedly confirmed in Jensen Huang’s recent speeches. He has said that the AI era is shifting from the software industry toward heavy industry, and that the traditional asset-light model is no longer sufficient. AI operations are constrained by the laws of physics. Every Token is the physical output of resources such as electricity, chips, and data centers. Compute demand is growing exponentially while supply remains structurally constrained. Therefore, companies must build powerful, scalable, asset-heavy foundations and invest enormous amounts of capital to construct a new type of industrial facility—the “AI factory”—otherwise they will be unable to participate in the next round of competition.

Put simply: if you want to establish a lasting position in the AI industry of the future, your assets must become increasingly heavy.

IREN is already moving along this path. It is one of the most central players in enabling NVIDIA’s broader transformation, and how IREN shapes itself through this process will become increasingly visible over the next 12 months.

Over the past ten months, IREN has devoted its full attention to TIME TO COMPUTER, while NBIS has focused on TIME TO REVENUE. Supported by its sales data, NBIS’s strategy has been recognized by the market, and both its share price and market capitalization have risen substantially, demonstrating a significant advantage in speed. By comparison, IREN’s share price has largely remained range-bound, making it appear significantly slower.

For IREN, there are three major things that need to be accomplished in TIME TO COMPUTER.

The first is the continued growth of secured and connected power. IREN has executed exceptionally well on this front, with nearly 3GW of secured power added within just six months.

The second is the quality and speed of liquid-cooled data-center construction. Horizon 1’s delivery, together with its dual validation from Microsoft and NVIDIA, has now confirmed this milestone.

The third is IREN’s ability to turn capital investment into a flywheel. This is where IREN’s advantages are strongest. First, IREN’s core team has particularly strong financial expertise. Second, vertical integration gives its assets a high degree of financialization potential. Third, its deep partnerships with leading AI companies provide credibility. Fourth, IREN is the only AI-factory designer and builder that independently controls the entire EPC process from beginning to end. Fifth, IREN continues to make its own financial moves, including hiring two senior experts from KKR. Sixth, NVIDIA itself stepped forward to bring together six major financial institutions to finance infrastructure construction.

These are the equally important three pillars for advancing IREN’s TIME TO COMPUTER. Over the past ten months, NBIS’s TIME TO REVENUE clearly had the upper hand and attracted sufficient market attention. But this week’s launch of Horizon 1 is a heavyweight counterbalance that could shift the balance between these two stage-based models. Although NBIS continues to surge in terms of share-price performance, while IREN has even experienced a “SELL ON THE NEWS” reaction, the real dividing line begins from this point forward.

$IREN believes that TIME TO REVENUE is not the priority. Especially when compute prices continue to rise and market enthusiasm remains high, selling expectations amounts to selling your assets too cheaply. The focus should instead be on executing TIME TO COMPUTER properly. Ultimately, the winner will be the company capable of continuously and reliably delivering high-quality compute capacity. This is a complex systems-engineering undertaking, and it requires comprehensive preparation across many dimensions.

There is currently a view that the window created by the scarcity of power will last only one or two years, and that IREN’s advantage will soon be challenged. This view is far too simplistic. It is not like that at all—not even close.

TIME TO REVENUE is much easier. CoreWeave is a prime example in this respect, but it has built an extremely risky corporate operating structure. Nebius is the second example: it has continually met the market’s short-term demands and has executed accordingly. Most importantly, NVIDIA also needs executors like these to ensure the strength of its GPU ecosystem. But ultimately, the fundamental factor that truly ensures continuously growing REVENUE is still TIME TO COMPUTER. From this point forward, the companies that genuinely execute well on the three core elements of TIME TO COMPUTER will gradually begin to demonstrate their power.

IREN experienced a similar situation during its Bitcoin-mining phase. Compared with that period, getting reliable compute capacity online rapidly is far more difficult and requires much greater patience and much more work. Even though GPU depreciation is extending, and H100s can still command good prices and enjoy longer service lives, IREN will nevertheless patiently wait for GPUs with the optimal price-to-performance ratio rather than allowing the short- and medium-term demands of TIME TO REVENUE to constrain its decisions.

I also came to understand the reason why NBIS’s VINELAND project was halted. The entire process consisted of a series of forced compromises and reactive adjustments, inadvertently creating a drama that carried an element of dishonesty. Across the nine batches of the project, the first two were approved based on relatively safe and environmentally acceptable power arrangements, allowing construction to begin. Later, because the environmental requirements associated with gas turbines could not gain community acceptance, the project was forced to switch to BE fuel-cell technology. However, this technology required LNG storage tanks to provide redundant safety, and these changes triggered renewed anxiety and questions within the community.

Even worse, the project contractor, DATAONE, jumped the gun and began construction without obtaining the necessary approvals, effectively starting work in secret. This provoked anger among local residents and further intensified the conflict, ultimately leading municipal regulators to forcefully halt the project while awaiting a new ruling.

Whatever the final outcome, even if the project eventually manages to complete the intended construction through a difficult and stumbling process, similar conflicts will continue to arise, because these projects genuinely have significant impacts on many aspects of residents’ daily lives.

Therefore, TIME TO COMPUTER is absolutely not simple. If there are long-term and persistent concerns around laws and regulations that conflict with residents’ daily lives, these risks accumulate over time and can become enormous. So, despite NBIS’s impressive financial results, it still has not convinced me to lower my risk assessment of its business model. In this respect, its uncertainty risk is actually becoming increasingly significant.

Among the three key elements required to achieve TIME TO COMPUTER, IREN has just cracked the biggest source of uncertainty. After Horizon 1, the pace of construction will increasingly accelerate. We will also see financial support join the process and help increase the speed of this flywheel. After some more time, IREN’s ability to supply compute capacity will increase significantly, ultimately enabling it to generate high-quality REVENUE. Over the medium to long term, IREN has the potential to surpass both NBIS and CRWV.

[Coffee House Stocks @CoffeeStocksGuy]

https://t.co/beuBz9iH1G

[The Count of Monty Bitcoin @CtMontyCrypto]

@LawrenceLepard Hi my name is Monty and I'm a golf-aholic...
If you ever come back to the hometown shoot me a DM and we'll play Lesie Park.