rapacia

@GMHfund be cool • be kind • be minimal

[rapacia @gmhfund]

ปริศนาธรรมที่อยู่ในไพ่
ทุกคนจะสังเกตว่าผมชอบ ใส่สัญลักษณ์ ♣️ โพธิ์ดอกจิก ไว้ภายในชื่อตัวเอง
มันมีเรื่องเล่าบางอย่าง ที่แฝงอยู่ในลักษณะของไพ่ ซึ่งมันบ่งบอกถึงชีวิตมนุษย์ ได้อย่างดี

[TheBigBerbowski @TheBigBerbowski]

I can't stress enough the importance of having a stomach for investing, rather than focusing on intelligence alone.

[The Wheelie Investor @WheelieInvestor]

How it feels when you sold $SNDK at $800 so you don’t have to go through this brutal drawdown

[Jukan @jukan05]

Jensen Huang and the bottleneck bros.

[Serenity @aleabitoreddit]

$NVDA CEO, Jensen Huang in Tokyo today said:

"The reports are not true. Vera Rubin is already in production. Giant amounts of production incoming"

As a flat denial to the accusation that their AI servers would be delayed.

Jensen appears to be using the Umbrella term for Rubin Ultra/Kyber, then pointed to current gen ramp.

Regardless, lot of damage has been done already to Nvidia supply chains stock prices, after the multiple delay reports. Even if Nvidia refuted them by clarifying Kyber change was an architecture optimization, not impacting timelines.

It's a very dangerous and growing trend for folks to post overreaching broader claims from technical nuances to grow viewcount... Especially if they get Jensen has to respond.

In the end, I'm trusting $NVDA on timelines since they probably have the greatest visibility into their own supply chain.

[Ren @ren_stocks]

Me after loosing 30% of my port this past two weeks.

Anybody got more of them bottlenecks?

[Serenity @aleabitoreddit]

One of the dumbest thematic selloffs I’ve seen to to date off:

- $META compute news that’s not new
- CPO delay report 1, that got refuted by $NVDA
- CPO delay report 2, that got refuted by $NVDA

High confidence, institutions will end up going long on the same names they’re bearposting after retail capitulates.

[ByungJun Ahn @dubidubabap]

SemiAnalysis는 올해 연초부터 같은 패턴을 반복하고 있다.

1. Micron HBM4 배제론

연초에는 Micron이 NVIDIA Rubin HBM4에서 사실상 배제됐다는 식의 이야기가 돌았다.

시장은 “Micron은 차세대 NVIDIA HBM 경쟁에서 밀렸다”로 받아들였다.

그런데 이후 Micron은 Vera Rubin용 HBM4 출하를 발표했고, NVIDIA가 SK하이닉스·삼성전자·Micron을 모두 Vera Rubin HBM4 공급사로 인증했다는 보도까지 나왔다.

최소한 이건 너무 단정적으로 전달됐거나, 후속 현실과 맞지 않았다.

2. SOCAMM 수요 붕괴론

그다음은 Vera Rubin rack의 CPU-side SOCAMM 메모리 축소 이슈였다.

55TB에서 28TB로 줄었다는 이야기가 나오자 시장은 바로 “AI memory demand가 식는다”로 받아들였다.

Micron은 급락했다.

그런데 이 역시 단순한 수요 붕괴로 보기 어려웠다.

오히려 공급 제약 때문에 realistic configuration으로 조정됐다는 반론이 나왔다.

수요가 꺾인 게 아니라, 병목과 공급부족 때문에 구성이 조정됐을 가능성도 있었던 것이다.

3. CPO·800VDC 지연론

그다음이 CPO와 800VDC였다.

SemiAnalysis 리포트 이후 optical 관련주가 한꺼번에 박살났다.

AAOI, COHR, LITE, CIEN, MRVL 같은 이름들이 동시에 맞았다.

논리는 CPO 대량 양산이 2028~2029년으로 밀릴 수 있고, 800VDC ramp도 늦어질 수 있다는 것이었다.

기술적으로 어려운 건 맞다.

CPO는 yield, serviceability, optical engine failure, ASIC integration 문제가 있다.

800VDC도 전력 구조 전환, 부품 신뢰성, rack architecture 난이도가 있다.

하지만 “어렵다”와 “끝났다”는 다른 말이다.

CPO adoption이 느려질 수 있다는 말과 optical interconnect chain 전체가 무너진다는 말도 전혀 다르다.

CPO가 늦어진다고 AI 데이터센터의 optical 수요가 사라지는 게 아니다.

오히려 NPO, LPO, pluggable, ELS, laser source 같은 중간 경로가 더 오래 살아남을 수도 있다.

4. Kyber NVL144 지연론

그리고 이번엔 Kyber NVL144다.

SemiAnalysis는 NVIDIA Kyber NVL144 rack architecture가 78-layer PCB midplane 제조 난이도 때문에 2028년으로 12개월 이상 지연됐다고 주장했다.

여기에 NVL72x2 취소, NVL576 저물량 가능성, Rubin Ultra scale-up domain 제한, AMD·Google 기회론까지 붙었다.

즉 이번에도 단순 제조 리스크가 아니라 NVIDIA 차세대 rack-scale roadmap 전체를 흔드는 narrative가 만들어졌다.

문제는 아직 NVIDIA 공식 확인이 아닌데도 시장은 또 “NVIDIA roadmap 차질”로 바로 소비한다는 점이다.

5. 진짜 문제는 반복 패턴이다

올해 흐름을 보면 계속 같다.

강한 단정.
시장 충격.
후속 반론.

그리고 시간이 지나면 실제 산업 흐름은 훨씬 복잡했다.

Micron HBM4 배제론.
SOCAMM 수요 붕괴론.
CPO 지연론.
800VDC 지연론.
Kyber NVL144 지연론.

하나하나 기술적 논점은 있을 수 있다.

하지만 매번 결론이 너무 자극적이다.

복잡한 제조 리스크가 수요 붕괴처럼 소비되고,
타임라인 조정이 사이클 종료처럼 소비되고,
특정 architecture의 병목이 밸류체인 전체 악재처럼 소비된다.

이 정도면 SemiAnalysis를 단순한 기술 리서치로 보면 안 된다.

지속적으로 이런 식의 교묘한 시장 흔들기성 리포트를 내는 것처럼 보인다면, 리서치 기관의 본질인 “리서치 신뢰도”는 사실상 0에 가깝게 봐야 한다.

리서치 기관의 핵심 자산은 예측력이 아니라 신뢰다.

틀릴 수는 있다.

하지만 계속 강한 단정으로 시장을 흔들고, 나중에 보면 실제 산업 흐름과 맞지 않거나 훨씬 복잡했던 사례가 반복된다면 그건 단순 오판이 아니다.

리서치 품질의 문제이고,
표현 방식의 문제이고,
시장 영향력에 대한 책임의 문제다.

누군가는 이미 SEC에 제보했어도 이상하지 않은 사안이다.

핵심은 리포트가 틀렸냐 맞았냐가 아니다.

market-moving 리서치가 반복적으로 섹터 가격을 흔들었고,
그 과정에서 이해상충, 선행 배포, 포지션 노출, 파생상품 거래 여부가 불투명하다면,
규제기관이 들여다볼 영역이 된다.

특히 SemiAnalysis는 AI 반도체, 데이터센터, 네트워킹 모델을 팔고,
기관용 리서치와 컨설팅을 팔고,
이제는 메모리·포토닉스 ETF 파트너십까지 하는 유료 정보사업자다.

그런 주체가 리포트 하나로 섹터 밸류에이션을 흔들 수 있다면 시장은 당연히 물어야 한다.

사실관계는 충분히 검증됐는가?

기술 리스크가 과장돼 전달된 건 아닌가?
기관 고객 배포 구조와 시장 반응 사이에 이해상충은 없는가?
작성자나 관련 주체의 포지션, 파생상품 노출, 거래 패턴은 투명한가?

부정적 리서치는 필요하다.

NVIDIA roadmap에도 리스크는 있다.
CPO도 어렵다.
800VDC도 어렵다.
Kyber도 쉽지 않다.

하지만 어려운 것과 망한 것은 다르다.
지연 가능성과 수요 붕괴도 다르다.
제조 병목과 산업 thesis 훼손도 다르다.

SemiAnalysis 리포트는 기술 자료로 참고할 수 있다.

하지만 그대로 믿고 매매하면 안 된다.

반드시 NVIDIA, Micron, Broadcom, Marvell, GF, Jabil, Coherent, Lumentum, Sivers 같은 primary source와 대조해야 한다.

올해 SemiAnalysis가 보여준 건 기술 통찰만이 아니다.

시장 narrative가 어떻게 만들어지고,
그 narrative가 어떻게 주가를 때리고,
그 과정에서 nuance가 얼마나 쉽게 사라지는지도 보여줬다.

리서치가 시장을 움직이는 순간,

그건 단순한 의견이 아니다.

검증받아야 할 권력이 된다.

[dylan ツ @demian_ai]

> be semianalysis
> sunday night, markets closed, perfect drop window
> "MASSIVE DELAY: kyber nvl144 pushed to 2028"
> nvl72x2 backup cancelled, 4-die rubin ultra axed
> attach photos of a 78-layer lasagna only 3 companies on earth can bake
> log off
> be the market monday morning
> ibiden, kingboard, semco eat 10-18%
> timeline unanimously declares nvidia cooked
> be anon who reads roadmaps instead of headlines
> kyber was always a late 2027 architecture
> "delayed to 2028" = a thing not due for 18 months slipped a bit
> the 2026 money printer is rubin nvl72
> already ramping, untouched by the thread
> be the pattern
> micron hbm4 exclusion -> all three suppliers certified anyway
> socamm demand collapse -> was a supply adjustment
> cpo is dead -> opticals crash -> photonics etf launches weeks later
> strong claim sunday, sector heemed monday, nuance thursday
> every single time
> be nvidia
> your scifi rack slips
> "fine, we'll sell more of the racks hyperscalers are literally fighting over"
> most profitable cope in market history
> tldr for the group chat
> the midplane is a real monster, jefferies said it two weeks earlier and nobody cared
> the near-term impact is way smaller than the font size
> interconnect is the main character now
> mfw anon panic sold the pcb bottom because of a sunday night post

[Markos @MarkosAAIG]

Okay guys, FINALLY follow-up on the HBM stress test. because I cracked my head on this topic way harder then I ever tought to do. Put so much time in this engine and learned also so much. So must read here!

Besides all the supply constraints that i mapped in other parts which are already as tight as it can be I looked at demand now and how it develops.

So four levers drive the supply gap. I pushed each to its worst realistic setting, one at a time, and the shortage held every time… except one lever could flip it alone. Efficiency. How fast the industry learns to serve more AI per gigabyte of memory.

So we measured it. Properly.

First thing to understand… our base case already assumes brutal efficiency. Tokens are growing 4 to 7x a year (Google confirmed 7x at I/O), yet our memory demand line only grows ~40%. That wedge IS efficiency… roughly 30% per year, already priced in. MLA cutting KV-cache roughly 3 to 5x. FP8 to FP4 quantization. Sparse attention. All of it, already in this number.

To flip 2027–2028 into surplus, efficiency would need to run another ~25% per year ON TOP of that 30%.

The measured incremental rate? Around 5% a year. And decelerating… the big wins were one-time steps, mostly shipped in 2024–2025, and we are approaching hard floors. 4-bit precision. ~70KB per token on KV-cache.

Then the other escape hatch… “CXL pooling and flash tiers will fix it.” I ran that one too. Most memory tier below HBM serves the COLD side. CXL pools sit at hundreds of nanoseconds… good for cached prefixes and reuse. High-bandwidth flash is read-mostly, write endurance caps it at shared context. System RAM costs 10 to 50 milliseconds per fetch. But the hot decode set… the memory a GPU touches every single token… is pinned to HBM by latency. NVIDIA even built a separate GDDR7 chip for the part that can leave (Rubin CPX, prefill). The boundary is drawn in silicon. These tiers are additive, not substitutive. They grow what is servable… they do not shrink what needs HBM.

For the math people: balancing 2028 would need roughly three quarters of KV traffic off HBM. The physical ceiling is nowhere near that.

And here is the stat that settles it. No GPU generation has EVER shipped with less memory per system. 80GB… 141… 180… 288 and rising. Every byte the software saves gets spent immediately on longer context and bigger batches. Agent tasks now burn millions of tokens each. That is Jevons paradox in silicon.

Now the honest part, because there always is one. Nobody publishes memory-per-token over time, so the 5% is our inference… wide error bars. And I wrote down exactly what would change our mind: production KV compression clearing 5x+, a generation shipping lower memory per rack, or the flash tier proving it can hold the hot set. Low probability for all.

in my opinion the two escape routes everyone cites… efficiency, and moving memory off HBM… I measured. Neither runs fast enough.

$MU $005930.KS $000660.KS

[Chris Barber (in SF) @chrisbarber]

I asked people: "Which AI equity related newsletters do you like?"

The list:

Top voted:
- SemiAnalysis by @dylan522p @fabknowledge (+ more incl @SKundojjala @andrewlekashman) - 5 votes
- FUNDA @FundaAI - 5 votes
- Citrini Research by @Citrini (+ @zephyr_z9 @jukan05 and more) - 4 votes
- Irrational Analysis by @insane_analyst - 3 votes
- Vik's Newsletter by @vikramskr - 3 votes
- Fabricated Knowledge by @fabknowledge - 2 votes

1 vote each:
- Semi Doped by @semidoped
- Key Context by @firstadopter
- PhotonCap by @PhotonCap
- Nutty by @NuttyCLD
- ChinAI by @jjding99
- Interconnects by @natolambert
- Beyond The Hype by @enertuition
- TMT Breakout by @TMTBreakout
- Tech Taiwan
- Damnang by @damnang2
- Semiconductor Business Intelligence by @clausaasholm
- SemiVision by @semivision_tw
- Alex_Intel_ by @Alex_Intel_

Thanks to @TMTBreakout, @BenBajarin, @EricJhonsa, @TMTLongShort, @TBU12345678, @damnang2, @jaygoldberg, @Midnight_Captl, @illyquid, @mzuhair123, @bubbleboi, @insane_analyst, @lithos_graphein for the recs. (There were no self-nominations/self-votes.)

[Oma James🇺🇸 @ChelfcAbby]

He thought the Snowman was dead, big mistake 😲
This is a masterpiece 100/100

[KawzInvests @KawzInvests]

The entire copper vs. optics debate. In one chart.

[Manic @Manicdyor]

This trader told everyone to short and the price only went up 😂

[Trader Theory @tradertheory]

Demo account vs. live account

[Ehrmantraut Capital @EhrmantrautCap_]

If you think today was awful, imagine investing during the GFC in 2008:

[Ted @TedPillows]

Stocks are dumping.
Gold is dumping.
Silver is dumping.
Crypto is dumping.
Bonds are dumping.
Even Oil is dumping.

If everything is dumping, where the hell is money going?

[Sam Badawi @Sam_Badawi]

Culprit Research revealed that they are short $NVDA today.

Citron also revealed that they were short $SNDK in February.

When will people learn?

[Serenity @aleabitoreddit]

LOL chat no way... Can't believe $AXTI / InP substrates really a major part of the China-US Donald Trump's trade talks with Xi?

I got this thesis completely right.

But really, Japan should start to cut off upstream machine supply like Ulvac for humanoids. And red phosphorus like Rasa/NCI to China, if they're not shipping.

And US should pressure EU to take back their their photonic monopolies like ficonTEC that Chinese companies like Robotechnik managed to buy.

[Not Match of the Day @NOT_MOTD]

I was not expecting that.

[Not Jerome Powell @alifarhat79]

Financial advisors putting your money in index funds