อย่าง Jane Street (enterprise) มี data center เป็นของตัวเอง (สร้างใช้กันแค่ตัวเองไม่ได้แบ่งใครด้วย) ทำโมเดลเองไม่ใช่แค่เฉพาะเทรด, R&D ตัวเองอื่นๆ
เทียบสเกลเหมือน
- คนใช้ หลอดรักโลก vs ไพรเวทเจ็ท
- ระดับ consumer vs enterprise
ผู้ใช้ทั่วไปเลยเป็น %น้อยมากของ impact ที่เกิดขึ้น
Very interesting statement today: $MU CEO predicts a multi-decade memory demand cycle driven by humanoid robots.
"Humanoid robots, he says, will require roughly ten times more memory than today’s Level 2+ autonomous vehicles."
"And that demand wave is set to begin before the decade is out."
Something as well as was "Over time, we expect the value of on-device AI combined with pent-up unit replacement demand to drive memory demand growth"
Which is also another trend (Apple Intelligence is currently dog, but I'm sure we'll see innovations with localized/edge AI).
Feels like all the industry leaders from $TSM Chairman, $TSLA Elon Musk, to $MU CEO see humanoids as the next major trend so physical AI is probably next.
I wonder if the world is going to have enough memory. Or if we'll see enough breakthroughs to shrink memory usage.
Leopold Aschenbrenner laid it out in 2024 and nobody listened
But it's not too late to make money and here's why:
• GPT-4 training cluster: $500M, 10 megawatts
• 2024 cluster: billions, 100 megawatts
• 2026 cluster: tens of billions, 1 gigawatt
• 2030 cluster: $1 trillion, 100 gigawatts, 20% of all US electricity
He said energy was the constraint the entire time
Here's his full thesis:
• GPUs were the trade of 2023-24
• Energy and power infrastructure is the trade now
• AGI arrives by 2027 and markets haven't priced it yet
• Electricity was the trade all along
SK Hynix: The king of high bandwidth memory.
• Gross profit +233%
• Net income +253%
• Revenue +192%
And $NVDA Rubin Ultra (2027) is about to push memory demand up even more.
The DRAM ETF launched back in April and is already up +112% since.
The top 3 memory gaints are all in there:
• $HXSCL (SK Hynix) - 28%
• $MU (Micron) - 25%
• $SSNLF (Samsung) - 20%
I love breaking down complex tech into visual bits and pieces.
My latest research piece on photonics covers:
• How it all works
• The 5 layers of the value chain
• Core businesses per layer
• A business quality tier list
And much more. I hope you find it useful!
Just a hunch a why Aschenbrenner might like $NBIS:
Estimated revenue growth:
• 2025: $0.5B
• 2030: $33B
Quite surprised he didn't get in earlier to be honest.