Alex A.C.

I'm seeing a new kind of Memory Panicans, but much worse: the former Memory Bulls that have sold their Memory names recently and now are saying: "ey, the cycle! downturn is near!"🤦‍♀️ Preaching against the ones we're holding. You're pathetic and know nothing🤷 $MU SK hynix $SNDK

My friend went on a date to the Guggenheim with a financebro in his 30s. He made her buy her own ticket, then used a fake student ID to get a discount on his. Then he told her she should get one too.

For Nvidia, it's bearish when it comes to margins (its own production op margins >> openAI ones), though it diversifies with that operation, plus its margines won't remain so high for ever I assume... It's bullish for AI in general (more demand, etc.) BUT if openAI goes for Open Models, that can be bearish Nebius.... 😬

[ 노무라 반도체 테스팅 핵심 분석 ] 테스팅은 제조 지원 기능에서 전략적 초크포인트로 전환 중이며, HBM-패키징이 누렸던 가격 결정력을 테스팅 벤더들이 다음으로 가져갈 가능성이 높다는 논지. 1. 테스팅의 위상 변화 과거엔 불량품 걸러내는 마지막 단계였지만, 이제는 가치 보호 수단이 됨. 칩렛, HBM 12단, 실리콘 포토닉스 등으로 패키지 하나가 수백만 달러짜리가 되면서, 늦게 발견된 불량 하나가 파괴하는 가치가 과거와 비교 불가. 2. 테스팅 비중 급증 (노무라 추정) 최종 테스트 시간: Hopper 대비 Blackwell 4배 에서 Rubin 7배 테스팅 비용 비중: GPU 원가의 1.9% (Hopper) - 2.5% (Blackwell) - 3.3% (Rubin) 비율은 작아 보여도 시스템 가격 자체가 급등 중이라 절대 금액은 훨씬 큼. 3. 테스팅의 전진 배치 생산 끝에 한 번 검사 - 웨이퍼 프로빙부터 KGD 검증, 하이브리드 본딩, 번인, 시스템 레벨 테스트까지 공정 전체에 걸친 다중 품질 게이트로 진화. 테스트 삽입 횟수가 늘수록 장비-프로브카드-소켓 수요가 증가. 4. 패키징과의 연동 CoWoS 캐파만 늘려봐야 테스팅 캐파가 못 따라오면 병목 해소 안 됨. 복잡도가 올라갈수록 테스팅은 선택이 아닌 필수. 5. CPO = 새로운 테스팅 시장 광 엔진과 ASIC 간 신호 무결성 검증은 이전 세대에 없던 완전히 새로운 테스트 카테고리에서 테스팅 업체의 추가 성장 동력.

Nomura just initiated coverage on CXMT, assigning it a price target of RMB 116 and arguing that it deserves a 20x P/E multiple—twice Micron’s valuation. That implies a 13.4x increase from the IPO price and a market capitalization of roughly RMB 7.76 trillion... This is insane. Seriously. RMB 7 trillion??

🚨 Apple just solved the biggest problem in AI.

Andrea Levy (PP) acaba de soltar una burrada monumental en La Sexta. Dice que en los últimos 15 años la temperatura de la Tierra ha aumentado 1.5º, cuando en realidad ha aumentado 0,3º. Ya basta de implantar el terror climático para justificar vuestra inutilidad.

Here is very well explained the mechanism of the recent 2-weeks HUGE drop in the Semis Industry.... and Photonics, and Big Tech, and..... well, you got the idea 😅 Let's say, the huge drop in the "AI names" 📉📉 🎇 And I add: the trigger was on purpose (the MMs, Big Boys, Robber Barons... call it whatever you want); too easy, too tempting. Too much money flowing in a very predictable way.

Nomura just initiated coverage on CXMT, assigning it a price target of RMB 116 and arguing that it deserves a 20x P/E multiple—twice Micron’s valuation. That implies a 13.4x increase from the IPO price and a market capitalization of roughly RMB 7.76 trillion... This is insane. Seriously. RMB 7 trillion??

CXMT will start trading in 3 hours It will be the craziest IPO I have ever seen

JUST IN: Berkshire Hathaway now holds enough cash to acquire 95% of the S&P 500

Insanity. Jensen Huang expects the semiconductor industry to grow 10x over the next decade. There is simply not enough compute. Foundries like $TSM are not able to step up production quickly enough, $ASML can't deliver EUV lithography systems fast enough, the Memory Big 3 (SK hynix $SKHY, Samsung $005930.KS and Micron $MU) are not producing enough memory, you name it. The demand is there. Everything is just being constrained by supply. And with inference & agentic AI, and physical AI at some point, that demand will not go anywhere.