Jim Chanos

“Any year now…” $TSLA

In distress and/or after an adverse legal ruling, these long-term annuities may become shorter-term deposits…subject to a run and/or mandatory redemptions.

https://t.co/txJEMzCK5i

How does this keep happening in the @NFL….?!

Anybody recall when the $SPCX lockups come off?

Best description I have seen. #GoldenAge

20 GW on the ground in the next 18 months is $1T. Which they will then render obsolete by launching space-based data centers. Got it…?! $SPCX

“I’ll call your Anthropic, and raise you an OpenAI…”

Yes, of course if that “trend line holds” then AI capex will be $64T in 2032, and $128T in 2033, effectively doubling global GDP. Where’s Anthony Odlyzko when you need him…?! #GetTheNet

@LesCrane3 I’m familiar with “aspirational statements”. #CorporatePuffery $SPCX https://t.co/fC6cMsR6lW

@mehancock1 @brad_D6 Yeah, no. Both the actual and marginal pre-tax ROIC’s are abysmal here. They have ticked-up slightly in the last four years, but are still LSD to MSD, well below the related WACC’s. $DLR $EQIX https://t.co/M74Mx5kNUa

@brad_D6 Nope. For the legacy data centers if you replace a piece of equipment (the HVAC goes down) and can convince your internal/external auditors that you can raise rents/attract new tenants-it’s considered “growth” capex. Which is ridiculous if the HVAC has to be replaced in any case. @brad_D6 (2) For example, $DLR would have you believe that “recurring” capex for the LTM was less than $400M on a $50B capital base. Over 100-yr life…?! https://t.co/VaLZuY5Ylh

@brad_D6 To be even more fair: The capex estimates are increasing far faster than revenues or FFO. The FCF at the legacy data centers is deteriorating rapidly. $DLR $EQIX https://t.co/Whp5a7MK1Q

@brad_D6 To be fair: $DLR has roughly $50B of capital employed in its business.

The 10-yr just passed 5.3%. If legacy data center $DLR’s $18.6B in debt (ex-pfds) cost 5.3% their annual interest cost would be almost $1B, which would wipeout their LTM Operating Income. Refinancing risk is real for the legacy data centers. $EQIX $CSQR (2) Speak of the devil…$DLR https://t.co/usRi7He2bU

@TMFAssociates The concept of “corporate puffery” doing a lot of heavy lifting here. $SPCX

At the midpoint (75 launches) this is an 80% cut in forecasted 2027 Starship launches by Musk. Seems material. $SPCX

@JohnProperBTC @robinivski I believe that when they say they are signing $20M per MW deals that figure is a “grossed-up” metric that does not take into account pre-payments (which reduce revenues in the out years), penalty clauses, less than 100% utilization, etc. The effective rates will be much lower.

@JohnProperBTC @robinivski Please stop. $IREN is expected to have roughly $5B in Calendar 2027 revenues. On average contracted MW of 550. That’s less than $10M per MW. Next year. Despite the $20M “deals” they are signing. https://t.co/ggnph64PwK

Why not $845B…?! This is all stock promotion, given the terms of the contract for the out years. $SPCX https://t.co/CFBUwmnDIm

@JohnProperBTC @robinivski That’s literally what they are doing. $IREN https://t.co/V8NV8PYTcg

@robinivski They are (with the help of promotional managements) conflating episodic spot price contracts of indeterminate size with overall revenues per MW. Then extrapolating that out. $CRWV $IREN $NBIS

Neocloud Math: JP Morgan seems to believe that $5B of D&A in FY 2030 on $80 to $100B of capital employed (2.1 GW) is proper. $IREN

And how much capital will $IREN have to invest to bring in that $13B of Adjusted EBITDA in 2030…?!

It was Mike McCarthy after the 2014 NFCCG. #IYKYK @packers

Per my earlier posts…

I’m sorry, but how does this work, if they are paying $20-40M per MW ($100-200B at YE 2026 ARR compute costs), with an ARR of $65B at the end of July? (2) And if Anthropic can’t pay $20M per MW and be profitable, how are the neoclouds going to be able to charge that rate consistently to other customers…?! $CRWV $IREN $NBIS

$CAVA announcing a buyback at almost 100x EPS should look at $WING since its raised buyback plan announced last March. https://t.co/ffNcIcOu83

#GoldenAgeOfFraud https://t.co/rSIR3yHyD4

Again, this would represent roughly 25% of US GDP in 2030.