Shkreli Argues DRAM Not True Commodity Due To High Capex Barriers

its often a lot easier to impute what the market is assuming instead of forecasting. humans are much better at tasks like this. so, for $MU:

F2027: consensus F2028: slightly below F2029-terminal: well below stock price: current price returns on cash: meager

it would take a large, new technical player to break ranks and undercut. they would need huge capacity. typically a new player will match price with the crowd rather than cause a price war which only hurts themselves and others

@MartinShkreli I think we're going to see Jevon's paradox rear its head in reference to memory

@MartinShkreli If Burry is itching to short something, why not the quantum fraud, instead he’s shorting the money printer. Even J Powell is envious of $MU

Shkreli explains it's often easier to back out what the current $MU price implies the market expects, rather than forecasting earnings yourself.

Per the model: F2027 assumptions match consensus, F2028 is slightly below, and F2029-terminal is well below consensus. This means near-term results are fully priced in, but longer-term growth and profits are discounted significantly. It also points to meager implied returns on cash at today's valuation.

@MartinShkreli Is $MU at least a 5-10x from here?

@MartinShkreli So youre telling me i can get $700-$800 in earnings for 1000$ share price over the next 8 years? And this is a bear case…. It Assumes AI doesn’t gobble up 5 to 10T in enterprise knowledge labor over the next decade and we stop at a $1T 10GW cluster

@MartinShkreli My strategy analysis !

as follows📈📈 👇 👇

@MartinShkreli yeah imputing market assumptions is way easier than guessing next quarter's earnings. makes sense. still need a good way to track all those assumptions though

@MartinShkreli Details of my stock holdings are as follows

⬇️

@MartinShkreli Chinese and even possibly indian memory manufacturers will flood the markets. We are not talking about some specialized, sophisticated chip here, memory is much easier than processors ir gpus.

@MartinShkreli @grok Explain this in more detail.

@MartinShkreli Even in commodity industries it takes few years to construct new plants and operate at full capacity

@JoelHadden @MartinShkreli how? thru dividends?

@MartinShkreli Agreed. It makes no sense to undercut in an environment where supply exceeds demand, assuming,of course, that the hypothetical new player can't out-supply the excess demand.

@MartinShkreli Are you an investor in the near term, or avoiding due to downtrending a couple years out?

@MartinShkreli FYI:

He’s highlighting that the current price reflects optimism on the boom years but deep skepticism afterward — and he follows up by arguing the business has higher barriers than a true commodity, which could challenge that pessimistic terminal view.

@MartinShkreli What all about all the increased memory demand once all cars and semis are electric & autonomous? Also how much dram will be needed when Humaniods scale and we have data centers in space? Do your numbers reflect this?

@MartinShkreli The problem is that forecasts are good Ceteris paribus, research team for micron is not forecasting the CPI is getting hotter and interest rate may rip and the stock may fall for completely non company reasons

@MartinShkreli I think individual cos in the space are difficult to understand due to the complexity of supply chain, and geopolitics

I bought the SMH, up 180pc so far