MetallRain

just an average human

[MetallRain @MetallRain_]

@strictrope @bubbleboi What positions are you holding rn?

[MetallRain @MetallRain_]

@bubbleboi @AleksandarGenov Translation: yes.

[MetallRain @MetallRain_]

@DeusXVI @menhguin @bubbleboi CMXT's capacity is fully booked for the next ~2 years just with local chinese costumers. Also, see below (although this may just be US building leverage in preparatio to the next fall's Xi-Trump meeting)

[Jukan @jukan05]

Will China still allow InP exports to the U.S.? I’m very bearish on AXTI. Now I’m starting to wonder whether U.S. companies will even be able to use CXMT DRAM.

Not financial advice. Do your own research.

[MetallRain @MetallRain_]

@kyuu47427 How are you playing the logic trade? Top 1-3 names?

[MetallRain @MetallRain_]

@menhguin @bubbleboi "people" as flash memory enterprise customers, not retail traders buying memory stocks

[MetallRain @MetallRain_]

@bubbleboi @bubbleboi what happened to the people who can't get enough flash?

[MetallRain @MetallRain_]

@bubbleboi https://t.co/9SyXIbO5Cv

[bubble boi @bubbleboi]

Fair value of Sandisk is 600B market cap. Meaning it’s 5k by end of year. People can’t get enough flash.

[MetallRain @MetallRain_]

@Vulpescap @plainyogurt21 I highly recommend the book "More money than God"

[MetallRain @MetallRain_]

@a_a_free You mean $RVMD ?

[MetallRain @MetallRain_]

@zephyr_z9 Are you working on your NAND piece? 🙏🏻

[Jukan @jukan05]

If you push this argument to its logical extreme, you arrive at the following conclusion:

GPUs are getting three times faster every two years, while storage is failing to keep pace.

“The criterion for selecting storage media has expanded from $/TB to $/GPU-hour.”

Traditional high-capacity storage has primarily been evaluated in terms of $/TB—in other words, how cheaply it can store large volumes of data.

But in AI training infrastructure, where GPU availability, power, and latency are critical constraints, that metric alone is no longer sufficient. If data arrives too slowly from storage and leaves GPUs idle, the cost of wasted GPU time may exceed any savings achieved on storage.

In simplified terms:

Incremental cost of flash < GPU idle time eliminated by flash × cost per GPU-hour

For workloads where this condition holds, using flash is economically rational despite its higher cost per terabyte. GPU-adjacent storage tiers should therefore be evaluated not only by $/TB, but also by the total system cost per effective GPU-hour.

In other words, up to a certain ceiling, it could make economic sense to spend several times more on flash than we do today.

[MetallRain @MetallRain_]

@mzuhair123 Thanks! I heard some prestigious voices are now bullish flash over dram, specially after Kimi release.

[MetallRain @MetallRain_]

@bubbleboi Translation:

"I’m trying to pull off a genius, outside-the-box breakthrough in flash memory that humans in the field wouldn’t normally consider"

- DogBoiGO

[MetallRain @MetallRain_]

@biotech_jack Do you have the note?

[MetallRain @MetallRain_]

@LaminarRed @Maximus_Holla He didn't disclose th size of his position

[MetallRain @MetallRain_]

@labubu_trader @floaturboat1 What I struggle to catch is where will the premium of US ADR over the european GDR €HY9H settle.

[Jukan @jukan05]

SK HYNIX CEO: NEXT YEAR IS EXPECTED TO BE THE WORST YEAR IN THE INDUSTRY’S HISTORY FROM A SUPPLY PERSPECTIVE — RTRS

SK HYNIX CEO: DESPITE AGGRESSIVE CAPACITY EXPANSION, MEMORY DEMAND WILL CONTINUE TO EXCEED THE COMPANY’S PRODUCTION CAPACITY OVER THE NEXT DECADE — RTRS

[MetallRain @MetallRain_]

@blondesnmoney Worth buying HY9H for it to fill the gap up on Monday morning?

[MetallRain @MetallRain_]

@seedy19tron @janniksin @DjokerNole 🙌🏻