Extrading retweeted
الست الفرنسية المزهولة دي .. عضوة في حزب اليسار الفرنسي وناشطة من اكبر دعاة فتح الهجرة لشعوب شمال افريقيا وبتدافع عنهم كل مايعملوا مصيبة وكارثة وتتهم الأخرين اللي عايزين يوقفوا طوفان الهمج دول بالعنصرية ...
هي عندها محل شيك صغير في مكان راقي وقفت في محلها تحاول تنقذه من النهب والسرقة وعشان هي ناشطة مشهورة
بتقولهم انا فلانة الناشطة بتاعة اللجوء انا معاكوا انا دافعت عنكم وعن دخولكم و بقائكم في فرنسا زقوها وقعوها وداسوا عليها بأقدامهم كويس انها لسه حية
ونهبوا المحل كله الست بقت علي الحديدة خسرت كل كفاح وشغل عمرها في لحظة ..
واقفة مذهولة مش مصدقة ان اللصوص الغوغاء المجرمين البرابرة دول هما اللي طول عمرها بتدافع عنهم وعن انهم يغزوا بلدها ويخربوه
عشان شهرتها في الدفاع عن لاجئين شمال افريقيا كل الميديا الفرنسية امبارح جايبة قصتها وصورتها قدام محلها المنهوب شقي عمرها كله وهي مش مصدقة مدي غبائها اللي وصل فرنسا للخراب والدمار ده كله 😢
Extrading retweeted
Just some words of emotional support for your SK Hynix, $SNDK, $MU fanclub.
$DELL CEO on memory, suggested: "It's a 4 year problem because there's just not enough supply"
Factories take about 4 years to build, meaning new production would emerge 2029 or 2030.
and that semiconductor shortages could worsen in 2027 stating:
"I do think it gets worse next year in terms of demand and supply".
I think it's strange we're getting unprecedented forward visibility with memory... as $GOOGL and richest hyperscalers seek LTAs into 2035.
And a major sector crash as well over durability/financing concerns.
$GOOGL is now discussing deals to extend their memory LTAs through 2035 from KeyBanc channel checks.
Samsung already has 70% of their capacity allocated through 2031, and this move would suggest a more durable cycle for $MU, $SKHY, and the others.
They also expect HBM prices up 80% to 100% in the Q1 2027.
Having all your capacity potentially allocated out into 2035 via LTAs would ease a lot of memory durability fears.
Extrading retweeted
Jensen reveals he first met the Founder of TSMC by cold calling their number off Yellow Pages
"I heard about this company in Taiwan, TSMC and UMC at the time. They were of course quite poor companies themselves, and they were barely making it."
"TSMC was probably about $300 million, something like that. It's a tiny little company. I heard about Morris Chang. And so I tried to call their US office."
"Back in the old days there were yellow pages. I looked it up, TSMC. There's a number there. I called their number and it's a recording. I left my name and my number and said a few things about NVIDIA: we're a new startup company building computer graphics chips. I called several times. Nobody called me back."
"And so I finally wrote a letter to Morris Chang. Didn't hear anything for a long time, and then one day it was 6 o'clock or something, this person called and said 'this is Morris Chang from TSMC.'"
"And meanwhile, everybody outside my office was playing video games and they're all laughing and screaming. 'Hey shut up! It's Morris!' I was so happy that the founder would actually call us back, because hey, maybe we won't have to die."
"And so we talked and he invited me to Taiwan. I came back to Taiwan, and that was the start of everything. Not just that partnership, but just the whole industry."
Extrading retweeted
If you're a memory investor worried about China's CXMT flooding the market & taking share from $MU & $SKHY think about this for a sec.
The US govt blocked Chinese EVs over national security. You really think it's going to let China take over the memory inside American AI? Buy!!
Extrading retweeted
The memory trade became so consensus even the Holy Father knows the base case.
Intelligence must draw on memory. The mind must not simply compile data — as an algorithm now does more quickly than we can — but recall lived experiences, which contain depths of meaning that only the human soul can recognize, and connect more deeply with the goodness of these experiences. Everything that is good begins “with the Spirit” and thus possesses a breath, a dynamism and a sense of communion that, once experienced, are engraved upon the heart. vatican.va/content/leo-xiv/e…
Extrading retweeted
$MU | DA Davidson 𝗺𝗮𝗶𝗻𝘁𝗮𝗶𝗻𝘀 𝗕𝘂𝘆 on 𝗠𝗶𝗰𝗿𝗼𝗻 𝗧𝗲𝗰𝗵𝗻𝗼𝗹𝗼𝗴𝘆, raises PT to $𝟯,𝟬𝟬𝟬 from $𝟮,𝟭𝟬𝟬 😳😳
Analyst sees investors early in understanding Micron’s value, leading them to assign a far higher multiple.
Extrading retweeted
$MU Morgan Stanley projects Micron will make $281B in revenue in fiscal 2027.
F4Q26 (actual): $54.2B
F1Q27E: $61.5B
F2Q27E: $67.0B
F3Q27E: $72.5B
F4Q27E: $80.0B
Extrading retweeted
We cut our RAM in half for the Tesla AI5 chip (now 72GB of LP5) and 1/3 for AI6 (now 144GB of LP6).
This was the only way to get enough volume for Optimus production and greatly reduces cost.
As it turns out, we think this will have a negligible effect on Optimus performance, as memory bandwidth is a bigger limiting factor than total memory storage (bandwidth was held constant).
Extrading retweeted
$MU You know what's the craziest thing about Micron right now?
Micron is about to out-earn every single Mag 6 name next quarter in operating income.
Respect the CASH
Extrading retweeted
The world is short memory.
Micron reports Q4 tomorrow after the close.
Street: $51B revenue / $31.50 EPS / 86% GM
Guide: $50B ± $1B / $31 ± $1 / ~86% GM
That’s on top of a Q3 that already did $41.5B, 84.9% GM, and $25.11 EPS.
Then look further out.
Consensus still has FY28 EPS close to $180, with margins modeled north of 80%. FY27 is already $160. The stock is $1070. That’s roughly 7x forward earnings.
AVGO margins are in the 70% range.
The market is pricing MU like 1990s NAND: violent cycle, peak margins, then a cliff.
The bull case is that this isn’t a 2-year product cycle. AI memory is still infrastructure. HBM + data-center DRAM tightness is being modeled into 2028, not 2026. Long-term customer agreements are locking in volume and putting a floor under pricing.
So the setup is simple:
Tomorrow is about whether Q4 and the FY27 guide keep the print-and-raise cadence alive.
The real debate is whether $180 in FY28 at 80%+ margins is a peak to fade or still early.
IMHO Memory names need to be rerate to at least a market multiple.
If this cycle lasts into the end of the decade, MU gets rerated. At 6x, the market is betting is betting on the past. $MU
Extrading retweeted
The presumption that the Fed raising short-term rates reduces inflation is predicated on the belief that higher rates reduce demand and investment.
But what if higher rates don’t reduce demand and investment because the demand for intelligence and energy is unaffected by higher rates because winning the race for super intelligence has a near infinite ROI and the demand for compute will remain incalculable.
Why won’t higher rates at this unique moment in history therefore lead to more inflation as interest costs are embedded in everything?
And the problem is compounded as the more the Fed raises rates, the more inflation we will have and the more the Fed will need to raise rates further and so on.
But what if the old models don’t apply to the current paradigm and the Fed is wrong?
I think the Fed might have just made a mistake. Am I right or am I wrong?
Extrading retweeted
Every prior form of labor was bound by biology. Train a person for years, you get one person.
Now train an AI model once, you get infinite instances of it, running in parallel, forever, for the cost of electricity. The Sun. Humanity has never had a workforce that scales by copy-paste. Now it does, digitally.
But a copy-pasted mind still has to hold a thought. And holding a thought is physical. It happens on memory. Real silicon. Made by three companies who've already said the shortage doesn't ease until 2028.
The workforce can be virtually infinite. Copy-paste.
The place it thinks cannot. Fab output is finite.
Infinite demand. Finite supply. You know what to do.
A lot of people are still wondering why trillions of dollars are flowing into AI.
Why is $TSLA building Robotaxi when people can drive just fine?
Why do we need more $NVDA GPUs and $MU $SKHY memory. Why @elonmusk claims money won't matter in some distant future?
It's all economics.
It's because every one of us, including YOU, wants the cheapest and most infinite form of labor. A labor that lets us live in more prosperity than we do today. Labor that doesn't break your bank, doesn't fatigue, scales without biological limit.
The $400/hour legal fees. The $1000 outlet installation. The $10,000 medical bill. These can break a wallet or bankrupt a household.
Ridiculous. We want reasonable costs. Reasonable spending for everyday basics. We want cheap and infinite labor.
Look at what technology has always done: it brings the cost of things down over time. Remember when renting a single DVD cost $10? Now you get millions of hours of entertainment for that same price on $NFLX. Free on YouTube $GOOGL. No driving, no waiting for a disc to ship.
My annual fuel cost used to run over $2000. Now that I drive a $TSLA, it's about $300.
International calls once cost $10 an hour just to hear a loved one's voice. You had to walk to a store and buy a calling card. Now you dial them instantly, for free.
See the pattern?
AI and Robots such @Tesla_Optimus are the next step. Low cost labor that produces so much that having a lot of money stops mattering for a family. Everyone gets what they need to live in abundance, because AI made it possible.
AI is INFINITE & AUTOMATED & CHEAP labor.
People will work much less but will receive much more in terms of good and services.
Every prior form of labor was bound by biology. Train a person for years, you get one person. Train a model once, you get infinite instances of it, running in parallel, forever, for the cost of electricity. The Sun. Humanity has never had a workforce that scales by copy-paste. Now it does, digitally. Physically, we're getting there.
But a copy-pasted mind still has to hold a thought. And holding a thought is physical. It happens on memory. Real silicon. Made by three companies who've already said the shortage doesn't ease until 2028.
The workforce can be virtually infinite. Copy-paste.
The place it thinks cannot. Fab output is finite.
Infinite demand. Finite supply. You know what to do.
$MU $SKHY $NVDA $TSM $TSLA $SPCX $LITE
Extrading retweeted
$TSM $NVDA $MU $SKHY $SNDK
$AAPL is building enterprise AI servers on its own M8 Ultra chips, 2 chip and 4 chip versions, to sell to AI developers, businesses and governments by 2029.
Two things come to mind.
First: Apple believes it can secure memory allocation, which means signing LTAs with the memory trio.
Second: Apple believes the AI infra buildout lasts decades, not a few years. You don't target a 2029 ramp for something you expect to fade.
If all of this holds up (still worth validating), it's a signal we can't ignore.
$NVDA $MU $SKHY Goodness. Apple is eyeing a return to enterprise servers, this time built for AI inference and sold to outside customers
Apple may sell a 2029-class M8 Ultra AI server to enterprises and may use NVIDIA NVLink Fusion to stitch chips together
Unlike Apple’s internal servers, this system would be sold to outside customers, a possible return to enterprise servers after Xserve was discontinued in 2011. Target: companies that want to run AI models on their own hardware.
The effort reportedly started about a year ago with backing from John Ternus (then hardware chief, now CEO).
Scaling Apple’s existing chip-to-chip links for this box could hit cost and speed limits, so Apple is looking at NVIDIA interconnect. NVLink Fusion lets NVIDIA sell networking into systems that don’t use NVIDIA GPUs.
Related AI-hardware color
-Mac lineup: record $10.4B last quarter.
-Some enterprises already rack Mac Studios for local inference.
-OpenAI has bought “tens of thousands” of Mac mini and Mac Studio systems for RL and agent training.
-Anthropic has rented Mac minis via AWS.
-Separate track: Apple AI server chips codenamed Baltra (Broadcom), aimed at ~2027 for Private Cloud Compute / Apple Intelligence.
One more giant lining up for HBM, DRAM & NAND. Bullish.
Extrading retweeted
Inference becomes 70 to 90% of all AI compute by 2030, and KV cache is its single biggest memory consumer.
$MU has solutions for all levels of KV cache to enable Contextual Intelligence: HBM > DDR5 > disaggregated DRAM > NVMe > data lakes. Every tier is theirs.
Extrading retweeted
You never raise rates in the teeth of an energy price shock. An energy shock is already contractionary — hiking into it is a double tap.
Extrading retweeted
🚨 WOW! President Trump reveals he told Fed Chair Kevin Warsh to vote WITH the Board to raise interest rates because 47 knows his single vote opposing wouldn't accomplish anything
"We should be paying the LOWEST rates anywhere in the world!"
"He's got a very tough Board...I talked to Kevin, and I said, you might as well vote with the Board because it's not gonna matter. The Board is HOSTILE and POLITICAL!"
The Fed voted today to raise rates 25BPS, to 3.75%-4.00%
Trump has even vowed to CUT OFF TRADE with nations who we have a trade deficit with, if the Fed doesn't do the right thing.
The Fed should stop playing games and being petrified of "growth causing inflation." CUT THE RATES!