If this turns out to be true, it may also be worth thinking about what it could mean for memory stocks going forward.

英伟达上周又历史新高了 我看B站除了好多恶搞视频 不得不说，现在Ai做视频，基本看不出来痕迹 太夸张了

🚨🚨 HBM density Alert HBM4 will have to go back to 12Hi instead of 8Hi

Memory is being modeled as a $1.6 trillion market by 2028, and almost all of the profit still sits with three names. Bloomberg and Bernstein consensus put combined supplier revenue at about $1.0 trillion in calendar 2026, $1.5 trillion in 2027, and $1.6 trillion in 2028. Operating profit climbs faster: about $0.8 trillion, then $1.2 trillion, then $1.3 trillion. The stack stays concentrated. Samsung Memory, SK Hynix and Micron take the large majority of both revenue and profit in each year. SanDisk, Kioxia, CXMT and YMTC are the smaller slices. The gap between the top three and the rest stays wide through 2028. Profit growing faster than sales is the tell. Consensus is pricing sustained pricing power and high utilization, not a fast return to oversupply. The jump is sharpest from 2026 to 2027. After that the gains slow. These numbers were out before Micron’s fiscal 2026 print. That print later showed record quarter and full-year sales on AI demand, with management saying tightness runs into 2027 and 2028. $SKHY $MU $SNDK

I am very uncomfortable about people trying to ascribe religious force or a surrender of human judgment to AI models, and think it is a real safety issue.

JPMorgan sees memory supply remaining tight through 2028, with customer order discussions already extending into 2031. Its estimates point to: • HBM bit demand growth: +63% • Non-HBM server DRAM: +37% • CY27 HBM blended ASP: +54% YoY JPM also believes Micron’s >35% revenue coverage under long-term supply agreements through 2030 could prove conservative, noting Asian memory peers have LTAs covering 50%+ of capacity.

When there's alot of a external noise. Best to tune it out and focus on your own execution. I have faith that Lip-Bu is making the best decisions for long term shareholders. If you don't then sell the stock and move on. Thank you for your attention to this matter.

$MU $SKHY Now this is a really interesting development and it’s bullish for memory. HBM stops being a part you can swap and becomes something designed into the chip years ahead, with Nvidia's controller built into the stack itself. Once you're designed in, you're in for the whole generation, which means longer contracts, more value per stack and a lot less of the old commodity cycle. $MU, SK Hynix and Samsung just got a stickier product, and $NVDA gets more compute out of every chip. Both layers win.

This is my thesis: no one uses AI. I repeat, absolutely no one. We live in a bubble. Even among my friends who pay for it, when I ask them to open ChatGPT and show me their queries, it’s the same handful of basic things. Most don’t even know they can upload a photo and ask questions about it. Connecting Gmail so an agent can read and send emails blows their minds. An agent opening a browser and checking them into a flight? They’ve never even heard of it. The massive challenge right now is adoption, and then getting people who already signed up to actually use what they’re paying for. Most have absolutely zero clue what’s possible. Imagine the compute shortage when everyone starts using AI like the top 1% of users do today.

DRAM and NAND memory shortages will deepen in 2027 to supply deficits of 5.9% and 4.6% respectively, according to Goldman Sachs. Unusually tight manufacturing capacity and surging enterprise model usage are expected to prolong severe market tightness through 2028. This supply deficit strengthens structural pricing power and operating margins for key memory producers such as Kioxia. Global token consumption is projected by Goldman Sachs to expand 24-fold to 120 quadrillion tokens monthly by 2030, driven by autonomous AI agents. In response, Kioxia management is already negotiating long-term supply agreements to lock in 50% of its 2028 output. Slower adoption of consumer AI workflows remains the key downside risk to monitor for memory demand. #AI #MarketStrategy Report date: 2026-10-01 · GS

I keep hearing that OAI's first generation chip, Jalapeño, is having yield issues.... +There are also rumors that they're pulling in the second generation because of it.

Trump gives out Tom Cotton’s cell number over the daylight savings bill.

Here’s what annoys me about how insanely cheap all the memory names from $MU and $SNDK to $DRAM, $SKHY, etc. are: Has nobody freaking figured out that these companies, who are AT THE HEART of the AI buildout and are printing billions in cold hard profits every quarter, will be able to go on an unprecedented buying spree and buy all the critical companies, resources, etc. in the supply chain to build integrated AI systems much closer to NVIDIA than legacy commodity memory? Market, turn on your brain please. You think these CEOs are retarded or got here by accident? These are people who have multi-year plans, my friends.

Agents are expensive because they need a lot more inference But in fact, most of the cost is NOT inference. It's keeping an agent alive AI agents need A LOT OF CPUs and A LOT OF memory to get things done!!!

Didn't Ben Franklin you can count on life, death, taxes, and Europeans realizing America was right after throwing temper tantrum hissy fits for years on end