Jim Chanos

@DingDongFB Lol, and I’m pretty sure the famous “clam strips” weren’t real clams.

HoJo’s, obv.

Really…?! So 6-8x the cost of a 1GW data center w/GPU’s. So why not raise prices 3-4x? It would still leave plenty of “profit” for his customers if that statement is accurate. $NVDA

Wait, so their Earthbound team couldn’t figure out latency issues…?! Lol, ok. $SPCX @EdLudlow

OpenAI is considering drastically lowering the prices it charges users as it seeks to win customers from its rival Anthropic https://t.co/OZUHS0SjoW via @WSJ (2) Price CUTS anywhere in the AI ecosystem when demand for “compute” is supposedly infinite, seems…problematic.

I wouldn’t worry, it’s only $22.7T of SpaceX’s claimed $28.5T TAM.

So SpaceX will require the entire US savings rate ($1T) by 2030…?! Seems reasonable.

So not only have the $CRWV execs sold almost a quarter of their holdings since going public, but Magnetar has cut its holdings in half, as well…?! Lol, ok. https://t.co/RbgtchE8Wa

Another Bitcoin miner-turned-data-center company ($APLD) announces a couple of transformational deals this year and 2027-28 estimates skyrocket. Oh, wait… https://t.co/d2joEvIJWS

So is $GOOG compute constrained as the deal w/SpaceX implied, but not so constrained that they can sell $35B of their chips to their biggest competitor (Anthropic)…?!

If only we had setup a special agency to investigate these claims.

This isn’t even remotely true. $AMZN went public in 1997 at a $450M valuation, or 3x revenues. $GOOGL went public in 2004 at a $23B valuation and at 7x revenues. $META had a $104B valuation in 2012 at 20x revenues(and immediately sold off almost 50%). SpaceX dwarfs these numbers.

This nine-month contract has more easy outs than a kid’s t-ball game. $GOOG $SPCX https://t.co/A5BdFe47Nm (2) Because if these economics are valid, then Nvidia will simply rent their GPU’s for a year, then sell them. Or better yet, simply raise prices to reflect these rental returns. Which is why you should never pay more for a neocloud (like xAI/SpaceX) than you should for $NVDA.

You can’t make this up…On their epic price target hike for $TSLA right before the SpaceX IPO, JP Morgan actually cut its 2027 estimates for Tesla. https://t.co/cxZNUjMeVR

If true, literally the “Golden Age Of Fraud”.

Yikes

$GOOG did have $126B of cash and marketable securities at March 31st. https://t.co/kPHPfBdIJD

As is the size of SoftBank’s balance sheet since 2000. https://t.co/LQDKKB1H7q

Reg FD and MNPI is for the little people. (Or of course, this is all made up) $TSLA

Who needs data centers in space…?! https://t.co/xDRjurLZHV

@matthew_sigel (2) And BTW, the legacy data center operators, who have been doing this for years, have property-level NOI margins of roughly 60%, not 85%. So we’ll see if $IREN can deliver on that guidance. Which given their track record, I wouldn’t count on.

Cantor’s report hiking its price target for $IREN today actually underscores the problem with its business model. They have average capital employed in FY 2027 of over $10B, with Adjusted EBITDA reaching $2B. Using 10-yr life(!)for its GPU’s/DC’s gets you to a 10% ROIC. 16% WACC. https://t.co/nAFd9AWeSW

@JohnTinsman So 120 quadrillion tokens per month in 2030 will be roughly $3T annually at today’s prices, if my math is right. Or roughly equal to current US after-tax corporate profits.

@JohnTinsman It’s cancellable within 90 days, i.e., a spot market pricing deal.

@C_S_Skeptic @BrianRoemmele (2) And $AMZN had a market cap of $440M at its IPO in 1997, or $900M in today’s dollars. The IPO’s being priced in 2026 are fully-formed.

Jeez, I see that a lot of you attended the Hopium Brunch this morning: “GPU’s will last 15 years!” “Asteroid mining is a real business model!”, etc.

So is Starlink March ARR actually below that of December….?! $SPCX https://t.co/1NiOwixrUj

SpaceX Adjusted EBITDA Table https://t.co/x3dJYj1RrI (2) The Details https://t.co/vTGLRWY4qV (3) More Details https://t.co/xop7YZ0ZaF

That’s a shame…

Congratulations to all my @YaleSOM students graduating today! #BoolaBoola