· 方法论

市场在瓶颈间轮动,基本面未变仅估值波动,建议理性看待。

涉及标的:

相关概念: InP 磷化铟衬底Pluggable 可插拔光模块光子学 Photonics

中文翻译

是的,我依然看好存储芯片,如 $MU / 三星。正如我之前所说,市场倾向于在瓶颈之间轮动。本周看来,光子学领域从 $AXTI 到 $LITE 再次成为焦点。关键在于……主要变化的是股价,随后辅以一些叙事和零星更新。对于光子学:> 在7月的下跌中,我们已经知道 $COHR / $LITE 的激光器已完全售罄,未来两年无货。> 我们从上季度财报电话会中已知 $AAOI 的需求失衡。7月崩盘期间,除了清算后的挂牌价外,基本面没有任何恶化。然而,当 $AAOI 跌至 $75 或 $AXTI 跌至 $35 时,许多人称其为“骗局”……但当它们涨回 $140 或 $80 时,人们又看多了,尽管光模块/磷化铟(InP)衬底瓶颈未变,甚至可能更糟(例如美国拟禁止中国新光模块,需求预测上调)。对于存储:我目睹大量散户投降,但这些人一个月前在 $MU 签署16份SCA并给出卓越指引时曾极度看多。或曾庆祝三星拥有全球最高营业利润。这里有一些更新,例如 Rubin Ultra 的存储优化(Nvidia每代都在追求),价格不再极端高于预期。但在当前价格下,营业利润相对于市值简直荒谬,尤其是存储成为结构性需求。明年的需求失衡应该更严重。人们倾向于在板块下跌时投降并跟随叙事(例如伊朗战争期间的氦气/LNG),即使瓶颈或基本面未变(例如 $SPCX Elon 财报电话会重申存储紧张)。我不能告诉别人该做什么:但 $140 的 $AAOI 和 $75 的 $AAOI 是同一家公司。市值 $1.5T 的三星和 $980B 的三星是同一家公司。只是估值和叙事(通常是噪音)在变,市场在板块间轮动。

英文原文

Yes, I'm still bullish on memory like $MU / Samsung. As I said earlier, markets tend to rotate from bottleneck to bottleneck. This week it looks $AXTI to $LITE in the photonics sector is the focus again. The thing is... the primary thing that changed are the stock prices, followed by some narratives + updates sprinkled in here and there. For photonics: > We already knew $COHR / $LITE lasers were completely sold out for the next 2 years during July's drop. > We knew about demand imbalance from $AAOI from last quarter's earnings calls. Nothing deteriorated fundamentally during July's crash, other than listed price after liquidations. Yet tons of people called $AAOI a "scam" when it dropped to $75, or $AXTI a "scam" on its drop to $35... But are bullish again at $140 or $80, when the transciver/InP substrate bottleneck hasn't changed at all, but maybe even got worse... (eg. draft for US ban on new china optical transceivers, scale up demand projections) For Memory: I'm witnessing a lot of retail capitulation, but the same people I'm seeing were mega bullish after $MU signed 16 SCAs and gave exceptional projections a month ago. Or were celebrating Samsung having the highest operating profit in the world. There's updates here and there eg. Rubin Ultra with memory optimizations (which Nvidia strives for every generation), with prices no longer being hiked way above expectations to the extreme. But the operating income relative to MC is just absurd around current prices, especially memory becomes structural. And the demand imbalance should be even worse next year. People tend to capitulate and follow narratives when a sector drops (eg. Helium/LNG back in Iran war), even if the bottleneck or fundamental situation hasn't really changed much (eg. $SPCX Elon earnings call reiterating memory tightness). I can't tell others what to to do: But $AAOI at $140 and $AAOI at $75 are the same company. Samsung at a $1.5T MC and Samsung at a $980B MC are the same company. Just valuations and narratives (often noise) change, and markets rotate from sector to sector.

在 X 上查看原推 ↗