Spatterings

content among other things

[Spatterings @_Spatterings]

Top

[Tushar Jain @tushar_jain]

We are watching the greatest trade of all time unfolding. Saylor will sell unlimited amounts of equity and debt to buy BTC.

Size and conviction are what makes trades legendary. Saylor's profits from this will far exceed Soros' and Paulson's legendary trades. The greats just keep adding to their position until they can't anymore.

The end result is that Saylor keeps issuing equity and debt until MSTR trades at a discount to NAV. But NAV at that point might be hundreds of billions or even a trillion.

[Spatterings @_Spatterings]

Technicians get called astrologists all the time by fundamentalists, but wait until you open the model of someone trying to be variant in the middle of a bubble

[Wall St Engine @wallstengine]

Google & Microsoft were both in talks to rent capacity at Nscale’s West Virginia Monarch campus before the 460MW lease went to Anthropic for $45B

$MSFT walked away during a data center portfolio review, while $GOOGL backed off after reassessing compute needs & capex

- Semafor

[malinvestment.jpeg @malinvested]

NVDA’s guide implies something like 30% GW growth FY28/FY27 vs 70-80% GW growth FY27/FY26

GW-driven end-markets with steep ‘27 capacity ramps gonna loosen

factor in double-ordering & pre-build of powered shells to maintain optionality (see SPCX, MSFT) => could get gnarly

[Spatterings @_Spatterings]

The better question is do we have 10trln to support this projection for the U.S.

[The Kobeissi Letter @KobeissiLetter]

We don't have enough power.

AI chip-driven electricity demand is projected to surge to a record ~315 gigawatts globally by 2033.

That would represent an increase of more than +1,100% since 2025.

The US is expected to account for ~64% of this new AI electricity demand, or ~200 GW.

Meanwhile, AI data centers are also putting huge strain on power infrastructure.

During model training, hundreds of thousands of GPUs can power up and down simultaneously, causing electricity usage to spike as much as 50% above design capacity.

These sharp power swings can accelerate wear on batteries, generators, cooling systems, and other critical infrastructure, increasing maintenance and replacement costs.

We need new energy infrastructure.

[Kalani o Māui @MauiBoyMacro]

“We ran a survey of 50 data center procurement leaders (~40% hyperscalers, ~60% colos / neoclouds / enterprise DCs) making purchase decisions for North American projects... Double ordering seems to be a phenomenon for ~50% or more of the respondent set.” 👇🏼

https://t.co/XgJUVW5blM

[Peter Berezin @PeterBerezinBCA]

$10 Trillion In Annual AI Revenue May Be Necessary To Monetize All The Capex Being Plowed Into Data Centers

Hyperscaler capex is expected to reach $1 trillion in 2027, most of which will be AI-related. Let us assume that a capex bust is avoided and capital spending remains at $1 trillion. Let us also assume a blended depreciation rate of 13%, which is roughly what the hyperscalers are currently assuming.

In steady state, the gross value of hyperscaler assets will then converge to 1/0.13=$7.7 trillion, with $1 trillion in annual depreciation expense. Using a straight-line depreciation approach, the net stock of hyperscaler assets will settle at about 0.5*7.7=$3.8 trillion.

The hyperscalers currently enjoy a pre-tax return on invested capital of 30%-50%. Just to steelman the argument, let us use the lower end of that range. In that case, they would need to generate 0.3*3.8=$1.2 trillion in annual EBIT, implying 1+1.2=$2.2 trillion in EBITDA. 

Analysts expect the EBITDA margins for the hyperscalers to rise to around 50% by the end of the decade. If they were to achieve this, they would need to generate 2.2/0.5=$4.3 trillion in annual revenue. That is $526 for every man, woman, and child on Earth. However, if EBITDA margins were to fall back to 30%, which is what they were in recent years, the required revenue would rise to 2.2/0.3=$7.2 trillion. 

Keep in mind that the foregoing calculation does not even include revenue from SpaceX, the neoclouds, or Chinese AI companies. If one were to include those companies and others, we are potentially talking about AI needing to generate $10 trillion in annual sales to justify all the capex being thrown at it.

For reference, global spending on food (including restaurants) is around $10 trillion. Health care is about the same amount. The entire global software market is only $1.4 trillion. 

Clients can read the rest of the report here:

https://t.co/u18LAEl6Hv

[Spatterings @_Spatterings]

@rennyzucker And they are doing a fairly major expansion due to the traffic/visitor count

[Spatterings @_Spatterings]

@rennyzucker Dallas and Houston are huge hubs

[Michael McDonough @M_McDonough]

Daily tanker traffic through the Strait of Hormuz vs. the U.S. 10-year Treasury yield:

[Spatterings @_Spatterings]

@ShanuMathew93 If the going rate of a GW is 50bln, total capex for even 2030 has to be on the order of 3trln+ for the backlog to clear. Pretty clear there is a lot of BS orders, just like the interconneciton queue. $CAT's reciprocating engine expansion alone could support 500 bln+ of capex

[Spatterings @_Spatterings]

The problem with 2030 numbers on AI names is that they are implicitly assuming another 4+ trln of capex up to that point

[Patrick Saner @patrick_saner]

Chart below for perspective.

"Virtually nothing matters more to markets [and the economy] at present than the AI buildout", as Bloomberg writes.

[Spatterings @_Spatterings]

Probably the most underrated earnings call comment and surprised it hasn’t gained more attention

[τop τick crypτo 📁 🤖🧠 @toptickcrypto]

Just remembered that on the MSFT earnings call they mentioned that all the executives were reading 1873, a book about the aftermath of the railroad bubble, lmao

[Spatterings @_Spatterings]

In 27 years I’ve drunk 50,000 beers / and they all wash against me like the sea into the pier

[Fireside Alpha @firesidealpha]

Dylan Patel says memory capacity is growing 20-30% per year while demand is doubling, so the shortage is not a short-term shortage, it'll last years.

Memory capacity is only growing 20, 30% a year for the next three years. And yet demand is doubling, is doubling."

"And so what's going to end up happening is memory prices are going to keep soaring. Users of memory who are less elastic or less capable of adapting to the elasticity of pricing will drop out of the market."

"Smartphones, laptops, because the costs are going to soar so much, they're going to drop out of the market."

"And that's going to all give way to AI. And what that means is the price is just going to have to soar and soar and soar until that happens, because capacity is not going up enough."

"And so ultimately our point there was memory is in a shortage, and this is not a short-term shortage. It's a shortage that's going to last years."

[Spatterings @_Spatterings]

@ContrarianCurse I mean that’s only if they are perfect, you can get natural diamonds for significantly less than that

[Mark Zuckerberg @finkd]

Releasing Muse Code in beta today. It's a terminal coding agent that takes on complete software engineering tasks across large repos: planning changes, writing code, validating the results. Powered by Muse Spark 1.2, a coding-focused model update.

[Silicon Data @Silicon_Data]

Our LLM Token index continues to decline while the debate over open vs closed model rages on. So we decided to take a closer look at how closed and open models respectively contributed. Interestingly, we are seeing some convergence btwn them.

The effective prices paid for proprietary models have decreased sharply over the last few weeks as OpenAI releases powerful frontier models at lower prices. Open models otoh have in fact seen their effective prices increase as more powerful near-frontier Chinese models like GLM 5.2 and Kimi K3 are released and served at higher prices.

Overall, this should be unsurprising. Econ 101: competition is up and prices are down. This is good for consumer and enterprise users of AI (agents) and promotes much wider and faster AI adoption.

[Spatterings @_Spatterings]

@jvb_xyz I will say bidding it to half the market cap of NRG is absurd

[Spatterings @_Spatterings]

@jvb_xyz Also have the same liabilities as other REPs during extreme weather events (ie look at NRG during Uri)