M. Eng. Electronic โก ใฐ๏ธ Financial data & analysis. Value investing. Growth? at a reasonable price ๐ฆ ๐ง๐ป
EU
Joined February 2023
- Tweets54.6K
- Following147
- Followers4.6K
- Likes97.3K
Pinned Tweet
A MONTH later....
- Micron $MU : from $724.66 ---> $1,141.75 [+57.5%] ๐
- SK hynix : from โฌ1,080 ---> โฌ1,580 [+46%] ๐
- Sandisk $SNDK : from $1,407.61 ---> 2,200.67 [+56%] ๐
Sorry not sorry for the former Memory Bulls turned... clowns, and a real sorry for the Memory Panicans ๐
Alex A.C. retweeted
The AI ROI debate is getting harder to dismiss.
Since late 2022, information, finance and professional/technical services produced **19.8% more real output with 1% fewer payroll jobs**.
Output per job is growing at a 5.1% annual trend, up from 2.3% before 2023. The rest of the private nonfarm economy? 1.4%.
Meanwhile, Anthropic reports over 6,000 customers spending at least $100K annually, versus over 600 a year earlier. Microsoft says automated tasks nearly doubled since July.
This is the shift weโve been writing about: **from answers to work.** Agents doing useful work on enterprise data, inside governed workflows.
Does this prove AI caused the acceleration? No. But โwhereโs the ROI?โ deserves a better answer than another chatbot joke.
Hard for me to look at these charts from the FT, especially the first one, and not see some evidence of โROI on AI.โ
Seems like a productivity surge post-Covid era overhiring that began to reverse before ChatGPT came out and then resumed its upward trend after ChatGPT.
Especially given the @a16z data on revenue per employee at their portfolio companies and the @EPBResearch data showing that the 3 industries that have most heavily adopted AI have gone from 2.2% productivity growth to 5% productivity growth.
And makes sense that it is showing up in the US first - America is 72% of Anthropic revenue vs. 50%ish at most application software companies.
Great stuff from @jburnmurdoch even if I disagree with his conclusion.
"Our ambition is to build an AI cloud, to become large and substantial enough to take on AWS and GCP" - $NBIS
๐ฑ๐คฉ๐ฅ๐ฅ๐ฅ
$NBIS: Don't call us a neocloud, we are an AI cloud. The term initially referred to a handful of providers, now it has lost all its meaning. It is borderline offensive, we should fade it into non-existence.
During a recent interview with Harper Carroll, Danila Shtan (CTO at Nebius) shared his thoughts on the word neocloud:
"I passionately hate this particular term [...] Putting everything under this one broad term, for companies that have a software platform and those that don't, is borderline offensive. I don't like the term neocloud. I think we should fade that into non-existence. Maybe we need something else"
The reason?
"The term neocloud was introduced to distinguish this new cohort of compute players from hyperscalers, the clouds. But when we started there were 5-7 companies in the field. Right now you see 500-ish companies. This term became so broad. Frankly, I have an issue with putting us or CoreWeave in the same category as the company that procures 2,000 GPUs and rents them to someone. These are different business models. These are different products. We refer to ourselves as an AI cloud"
Despite the terminology, Nebius's mission remains the same:
"Our ambition is to build an AI cloud, to become large and substantial enough to take on AWS and GCP"
Source: Harper Carroll AI (YouTube)
Alex A.C. retweeted
Visa and Mastercard see agentic commerce as a significant long-term opportunity, although current transaction volumes remain very small.
$V Visa CEO: "It's still very early. So we do have conviction that the opportunity is significant. We do have conviction that there's real incremental opportunities. But as we think about the next 4 quarters in '27, it all remains very small in the big scheme of the enormous size of our business."
$MA Mastercard CEO: "It is still early days, and the numbers are not yet significant in the grand scheme. But as the AI agentic assistants have emerged, agentic transactions are ramping up at impressive weekly growth rates. We're now seeing the early stages of consumer adoption."
Alex A.C. retweeted
Interesting visual shared by @friedberg on @theallinpod
Starting from $1 of Nvidia COGS, the money flows into $4.00 of Nvidia hardware revenue, scales to $8.6 per year in SpaceXAI compute revenue, grows to $17.8 per year for Anthropic, and ultimately unlocks $36โ178 per year in value for the end customer.
$SPCX $NVDA
Alex A.C. retweeted
ALERT๐จ๐จ๐จ: AMD's @AnushElangovan and his team have achieved up to 2x better performance per dollar than even B200 on Kimi K3 Agentic Inference! This is due to various software optimization efforts, including tuning AITER kernels and MonoKernel decode.
Along with MI355X having 2304GB of HBM per server, this allows it to avoid crossing slow 8x400Gbit/s scale-out networking between servers, which B200 needs to do due to having less than 1.5TB of HBM.
Great work to Anush, Peng Sun, Ramine, Emad, and the vLLM team for this amazing work!
"Our analysis suggests AI CPUs require ~1.5TB of memory today and potentially 2-3TB in coming generations, compared to ~0.5TB for traditional CPUs"
So up to x6 more DRAM Demand for AI CPUs vs Traditional CPUs ๐๐ฅ๐ฅ
$MU $SKHY
UBS: our recent checks increasingly suggest the ideal CPU:GPU ratio for AI infrastructure could settle closer to 3:1 and potentially 4:1. Against our expectation for ~44M accelerators in CY30, a 3-4x ratio would imply ~120M CPUs supporting AI infrastructure. Our analysis suggests AI CPUs require ~1.5TB of memory today and potentially 2-3TB in coming generations, compared to ~0.5TB for traditional CPUs.
$INTC $ARM $NVDA $AMD
"This unconstrained TAM would require 282EB of DRAM, 3.5x higher compared to current industry estimates of 77EB" ๐ฑ๐ฅ๐ฅ
Supply in the next years won't increase that amount not even in dreams. I don't think not even x2
teapawt.net/techfund1/status/21165โฆ
UBS' calculation on DRAM required for the "aspirational" agentic CPU build in 2030:
The setup:
~44MM accelerators
CPU-to-GPU ratio settling near 3:1, possibly 4:1
~120MM AI CPUs + ~30MM traditional CPUs
~1.5TB per AI CPU today, 2-3TB in coming generations
~0.5TB per traditional CPU
"This unconstrained TAM would require 282EB of DRAM, 3.5x higher compared to current industry estimates of 77EB"
Memory, not compute demand, remains the binding constraint in the AI buildout.
In order for the AI CPU market to take off, much more DRAM supply is needed.
Alex A.C. retweeted
Unitree shipments to increase to 30,000 units in 2027, up from 10,000-12,000 units in 2026.
Humanoids shipment volume picking up.
Don't be fooled guys, told you ๐ If you are selling your AI/Tech stocks now, you'll regret it A LOT in the months ahead....
teapawt.net/MgnStnlyDerivs/status/โฆ
The unwind in Hedge Fund equity exposures this year has been the biggest since at least 2017.
Active de-grossing (reduction in long + short positions) by global hedge funds has now hit over 8% of total gross exposure since June, according to the Morgan Stanley Prime Brokerage Content team.
This has taken gross leverage for US long/short funds down to the 10th percentile versus the last year.
Alex A.C. retweeted
The implicit assumption in the question is that higher rates help fix inflation in all contexts.
In reality, certain conditions/causes for inflation are not fixed by higher rates.
"Therefore, even if credit issues are being raised in the market, we are not experiencing any impact from them. Customer DEMAND IS ACTUALLY GROWING EVEN STRONGER than in the previous quarter" - $SNDK ๐
Fundamentals ACROSS industry stronger as ever. No fear guys
$MU $SKHY
์๋์คํฌ( $SNDK) FY27 1๋ถ๊ธฐ ์ด๋์ฝ Q&A ์ค
Q. [BofA) / ์์ ๋ชจํ]
๋ฐํ ์๋ฃ์์ ์ ๋์ ์ธ AI ๊ธฐ์
๋ค๊ณผ์ ํ์๊ฐ ์ถ๊ฐ์ ์ธ ์์ ์ฆ๊ฐ ๊ฐ๋ฅ์ฑ์ ์์ฌํ๋ค๊ณ ์ธ๊ธํ์ต๋๋ค.
์ฌ๊ธฐ์ ๋งํ๋ AI ๊ธฐ์
๋ค์ AI ์ฐ๊ตฌ์(AI Labs)์
๋๊น, ์๋๋ฉด ๋ค์คํด๋ผ์ฐ๋(Neocloud) ๊ธฐ์
๋ค์
๋๊น?
๋ํ ์ต๊ทผ ๋ค์คํด๋ผ์ฐ๋ ๊ธฐ์
๋ค์ ๋๋ฌ์ผ ์ ์ฉ์ํ ์ฐ๋ ค๊ฐ ์๋์คํฌ ์ฌ์
์๋ ์ด๋ค ํํ๋ก๋ ์ํฅ์ ๋ฏธ์น๊ณ ์์ต๋๊น?
A. [๋ฃจ์ด์ค ๋น์์ / CFO]
ํ์ฌ๊น์ง ๊ทธ๋ฌํ ์ํฅ์ ์ ํ ํ์ธ๋์ง ์์์ต๋๋ค.
์คํ๋ ค ๊ณ ๊ฐ๋ค๋ก๋ถํฐ ๋งค์ฐ ๊ฐ๋ ฅํ ์์๊ฐ ๋ํ๋๊ณ ์์ต๋๋ค.
์์ ๋ง์๋๋ ธ๋ฏ์ด, ๋น์ฌ์ ์ต๋ ๊ณ ๊ฐ๋ค์ ์ดํด๋ณด๋ฉด ์ฌ์ค์ ๋ชจ๋ ํฌ์์ ๊ฒฉ๋ฑ๊ธ(Investment Grade) ๊ธฐ์
์ด๋ฉฐ, ๋งค์ฐ ๊ฐ๋ ฅํ ํ๊ธํ๋ฆ์ ์ฐฝ์ถํ๊ณ ์์ต๋๋ค.
๋ฐ๋ผ์ ์์ฅ์์ ์ ์ฉ ๋ฌธ์ ๊ฐ ์ ๊ธฐ๋๊ณ ์๋๋ผ๋ ๋น์ฌ๋ ๊ทธ๋ก ์ธํ ์ํฅ์ ์ ํ ๊ฒฝํํ์ง ์๊ณ ์์ต๋๋ค.
๊ณ ๊ฐ ์์๋ ์คํ๋ ค ์ง๋ ๋ถ๊ธฐ๋ณด๋ค ๋์ฑ ๊ฐํด์ง๊ณ ์์ต๋๋ค.
$NBIS, $NVDA
Alex A.C. retweeted
Cloud growth is accelerating across Big Tech:
$AMZN CEO: "AWS grew 39.3% year-over-year, accelerating for the sixth straight quarter, our fastest growth in nearly 5 years. Backlog is $578 billion, growing triple digits year-over-year and well diversified."
$MSFT CFO: "Azure revenue was ahead of expectations and increased 48%, even with the 3P capacity delays noted earlier, driven by Foundry and improved CPU monetization."
$GOOGL CFO: "Cloud revenues were up 103% to $30.7 billion, driven by strong growth in GCP, with strength in AI, including both AI infrastructure and AI solutions, and core GCP"
