EveryMegawatt

[EveryMegawatt @Everymegawatt]

Keep up to date with filings on all your favorite neoclouds at https://t.co/GK06z7Hm6c

$CRWV $NBIS

[IREN @IREN_Ltd]

Horizon 1: delivered.

IREN has delivered Horizon 1 to Microsoft and achieved NVIDIA Exemplar Cloud status on NVIDIA GB300 NVL72.

Read more: https://t.co/ItVXSXK1V7

[EveryMegawatt @Everymegawatt]

The Innovator’s Dilemma and $IREN

In 1997, Clayton Christensen explained why great companies fail: they listen to their best customers, protect their best margins, and rationally cede the low end, right up until the low end eats them.

$IREN is running the playbook in real time.
The setup. Hyperscalers built the most profitable infrastructure business in history: 30%+ margins on managed services, enterprise lock-in, egress fees. Raw GPU compute is their worst business , capital-heavy, commoditized, margin-dilutive. So they underinvest in it. Not stupidity. Good management. That’s the dilemma.

The entrant: A Bitcoin miner. The most dismissed business model in tech, doing the most stigmatized thing possible with electricity. But while hyperscalers optimized site selection for fiber and tax breaks, IREN was hoarding the asset nobody valued: gigawatts of cheap, energized, grid-connected power. Then the binding constraint of the entire industry flipped from capital to megawatts and incumbency didn’t transfer. The interconnection queue is now 4-7 years. IREN’s queue position was already stamped.

The beachhead: Microsoft, protecting its own capex model, rents rather than builds. $9.7B over 5 years. Customers prepay GPU capex. The incumbent’s rational move funds the entrant’s buildout: 3MW → 480MW of self-built AI cloud in ~24 months. 1.2GW targeted for 2027.

The climb: This is the part Christensen said always comes next. Bare metal → managed cloud → platform. The customer list already shows it: Perplexity, Together, Fireworks, Fal, Hume, Figure AI. The AI natives, the enterprise giants of the 2030s, are forming their infrastructure defaults outside the hyperscaler value network.

The kicker: The market still prices $IREN like a landlord, real estate multiples on powered shells. But landlords don’t have AI natives on managed cloud. Landlords don’t book 85% of a $4B+ ARR target under contract on 4-year terms. Landlords don’t get customers to prepay almost half the capex.

That’s not a tenant roster. That’s a platform being born, funded by the incumbent’s own rent checks.

Christensen’s entrants always look small right up until they look inevitable. AWS was a bookstore’s side project. IREN was a Bitcoin miner.

The dilemma isn’t coming for the hyperscalers. It’s already inside the building and owns the substation.

[EveryMegawatt @Everymegawatt]

@_clarktang Yup.

[EveryMegawatt @Everymegawatt]

The Innovator’s Dilemma and $IREN

In 1997, Clayton Christensen explained why great companies fail: they listen to their best customers, protect their best margins, and rationally cede the low end, right up until the low end eats them.

$IREN is running the playbook in real time.
The setup. Hyperscalers built the most profitable infrastructure business in history: 30%+ margins on managed services, enterprise lock-in, egress fees. Raw GPU compute is their worst business , capital-heavy, commoditized, margin-dilutive. So they underinvest in it. Not stupidity. Good management. That’s the dilemma.

The entrant: A Bitcoin miner. The most dismissed business model in tech, doing the most stigmatized thing possible with electricity. But while hyperscalers optimized site selection for fiber and tax breaks, IREN was hoarding the asset nobody valued: gigawatts of cheap, energized, grid-connected power. Then the binding constraint of the entire industry flipped from capital to megawatts and incumbency didn’t transfer. The interconnection queue is now 4-7 years. IREN’s queue position was already stamped.

The beachhead: Microsoft, protecting its own capex model, rents rather than builds. $9.7B over 5 years. Customers prepay GPU capex. The incumbent’s rational move funds the entrant’s buildout: 3MW → 480MW of self-built AI cloud in ~24 months. 1.2GW targeted for 2027.

The climb: This is the part Christensen said always comes next. Bare metal → managed cloud → platform. The customer list already shows it: Perplexity, Together, Fireworks, Fal, Hume, Figure AI. The AI natives, the enterprise giants of the 2030s, are forming their infrastructure defaults outside the hyperscaler value network.

The kicker: The market still prices $IREN like a landlord, real estate multiples on powered shells. But landlords don’t have AI natives on managed cloud. Landlords don’t book 85% of a $4B+ ARR target under contract on 4-year terms. Landlords don’t get customers to prepay almost half the capex.

That’s not a tenant roster. That’s a platform being born, funded by the incumbent’s own rent checks.

Christensen’s entrants always look small right up until they look inevitable. AWS was a bookstore’s side project. IREN was a Bitcoin miner.

The dilemma isn’t coming for the hyperscalers. It’s already inside the building and owns the substation.

[EveryMegawatt @Everymegawatt]

Wall Street last week: "GPUs depreciate too fast."

Jensen on Monday: "No actually, it's bankable infrastructure like electricity. Sign here for $500B."

Wall Street: "Say less, King."

Imagine looking at this and staying bearish on $IREN. 💀

[EveryMegawatt @Everymegawatt]

But Rubins will be useless in 2 years or something....

[Rittenhouse Research @RHouseResearch]

CoreWeave $CRWV: We recently signed an A-100 contract that extends into 2029. This SKU was introduced in 2020.

[EveryMegawatt @Everymegawatt]

@MiningDataAI I hate the ATM, but also fine with the ATM. They gotta fund the buildout until revenue can start paying and the flywheel starts. It is what it is.

[EveryMegawatt @Everymegawatt]

Good news, Papa Jensen just solved $IREN financing issues.

They’re going to buy back all the ATM shares at a discount!!!

😂

[Prakash @8teAPi]

Incredible. Jensen is completing the circle.

- Bankers don’t like GPUs as collateral because the depreciation is unpredictable
- It’s unpredictable because a new GPU can obsolete an old one
- Jensen knows his own roadmap
- so he’s offering depreciation insurance to the banks
- the depreciation insurance (up to 25%) helps the banks get marginal deals over the line

Speculation
- Nvidia will also advise the banks on “reference designs” for datacenters that will make them fungible
- Having them be fungible means that the debt can repackaged into Asset Backed Securities, Collateralized Loan Obligations and Collateralized Debt Obligation (ABS, CLOs and CDOs from 2008 haha)
- This allows tranching to get investment grade ratings on the debt so that it can be resold to pension funds and insurance firms
- It also allows the banks to trade idiosyncratic project specific credit risk for sector wide credit risk

So Jensen is trying to get his customers the same cost of financing as real estate rather than venture equity.

This is going to move the data center game out of the VCs and into the big leagues.

[EveryMegawatt @Everymegawatt]

@bitcoinbutcher1 You rang?

[EveryMegawatt @Everymegawatt]

@DollarCostAvg @grok sum of the parts on $cifr and compare to today’s close.

[EveryMegawatt @Everymegawatt]

@a2ycapitalinc @MatthewDorey3 @ilzmcfly

[EveryMegawatt @Everymegawatt]

@BitcoinAIGuy @DollarCostAvg

[EveryMegawatt @Everymegawatt]

@FrostByte123456 Listen to it, you'll buy more.

[EveryMegawatt @Everymegawatt]

@RockwaterEQ @rftylerpage Very

[EveryMegawatt @Everymegawatt]

$CIFR

1. Saw ERCOT fud.
2. Saw EPS miss.
3. Saw stock price tank.
4. Listened to @rftylerpage earnings call.
5. Bought the dip.

More bullish than ever.

[EveryMegawatt @Everymegawatt]

Workin on somethin...

https://t.co/GHdUagwnXD

[EveryMegawatt @Everymegawatt]

@rftylerpage No worries; great call today.
https://t.co/mQSfkStMPj

[EveryMegawatt @Everymegawatt]

$CIFR

1. Saw ERCOT fud.
2. Saw EPS miss.
3. Saw stock price tank.
4. Listened to @rftylerpage earnings call.
5. Bought the dip.

More bullish than ever.

[EveryMegawatt @Everymegawatt]

@patrick_baille @rftylerpage Twitter: ERCOT new proposal is terrible for CIFR!!!

Tyler: This is probably going to create more demand for our existing batch zero and we’re not too concerned for the rest of our portfolio.

[EveryMegawatt @Everymegawatt]

@MktMavPro @CDavidFloyd This might be one of the dumbest 15% dips we’ve seen yet. Nothing from the earnings call made me want to do anything but buy more $CIFR.