觅信聚合

我的关注

2026-09-15 20:46

的收购，并尝试映射其供应商（这些不是当前供应商，仅是约 2022 年他们联盟研发时的历史供应商）：

- Almae Technologies
- Adtran
- CEA-Leti
- Bright Photonics
- TEEM Photonics
- ficonTEC
- Mellanox（已被 $NVDA 收购）
- Fraunhofer IZM
- AUTH
- Tektronix

大多为私营公司，ficonTEC / Mellanox 被众多企业使用。对投资私营公司的朋友来说，这些名称可能值得关注。

但从 Credo 现有关系来看，$TSEM 用于 PIC、$TSM 用于 DSP、$FN 作为 OSAT（根据高管评论）似乎难以回避（虽未确认）。

值得注意的是，他们的新 1.6T 收发产品使用了 2 个外部连续波激光器，这表明 Credo 正在从商用 CW 激光器供应商采购。

例如规格明确写明（外部连续波激光源）。

这让我更想了解 $CRDO 正在向哪些供应商采购，因为 $SIVE 最近提到 6 项可插拔合作。

（$CRDO / $COHR 是美国的潜在变量，中国的可插拔玩家可能最多，因为他们正在设立中国办公室并从 CIOE 2026 渠道检查开始与其合作开发）。

我想知道他们是否已经锁定产能，去年早些时候的 BOA 大会上，他们提到为旧产品提前锁定激光产能。

但目前还没有关于这条 1.6T 产线的消息，后续活动中需要关注他们的配额分配。几天前的 Goldman Sachs 大会也未提及。

无论如何，这对商用 CW 激光器供应商来说是个积极信号（虽然 1.6T 降低了单个模块所需的激光数量，但整体产量提升，带来更大的总体需求）。

而且连续波激光器一直供不应求，之前的商用供应商如 $COHR 已转向内部使用并转向公开市场买家，这限制了供应商名单。

至于 Credo 本身，它拥有自研的 DSP、硅光 PIC 以及集成 IP，提升了利润率。这对公司非常有利（但我未持续跟进，暂无方向性观点）。

最终，我仍然看好

2026-09-15 20:46

原文：
$CRDO announced their new 1.6T optical transceiver today. Just some thoughts / Dustphotonics stuff:

Obviously Credo and high-beta co seems to be getting nuked from rate hikes tomorrow + macro fears.

Combined with very high increase in short interest across a lot of these names from August into September. So very muted market reaction in general.

But I just got curious and did digging into their DustPhotonics acquisition and see if i could map their suppliers (these are not current suppliers, just historical back in ~2022 from their Consortium R&D efforts):

- Almae Technologies
- Adtran
-CEA-Leti
-Bright Photonics
-TEEM Photonics
-ficonTEC
-Mellanox (acquired by $NVDA)
-Fraunhofer IZM
- AUTH
- Tektronix

mostly private, then ficonTEC/Mellanox are used by a ton of companies. But for people who invest in private companies, a lot of these names might be interesting to loo.

But it does seems like $TSEM for PIC/ $TSM for DSPs / $FN as OSAT (from exec comments) is kinda hard to see not being used (not confirmed though) just from Credo's existing relationships.

Thing that stood out to me is that their new 1.6T transceiver product uses 2 external CW lasers so this confirms Credo is buying off merchant CW laser suppliers.

eg. specs explicitly state (External continuous-wave laser sources).

Obviously this makes me more interested who $CRDO is buying off of since $SIVE mentioned 6 pluggable engagements recently.

($CRDO / $COHR are US wildcards, China pluggables players were most were likely since they're setting up an China office and starting development with them from CIOE 2026 channel checks).

I was curious if they secured capacity already, their BOA conference earlier this year, they mentioned securing laser capacity ahead of time for their older products.

But nothing yet on this 1.6T line, so will need to hear in future events about their

2026-09-15 20:46

allocations to map more links. Goldman Sachs conference few days ago didn't mention anything either.

But regardless, this is positive for merchant CW laser suppliers in a nuanced way (their 1.6T reduces laser content needed, but they produce more, so creates more overall demand).

And given CW lasers have been very bottlenecked + former merchant suppliers like $COHR were redirecting lasers internally now and turned to a buyer off open markets. So that limits supplier lists.

As for Credo itself, it has its own DSP, silicon photonics PIC, and integration IP, which drives higher margins. And this is very materially positive for the company (but haven't been keeping up to date so no directional opinion here).

But in the end the reason I still like laser companies as it becomes increasingly important + valuable from an IP/acqusition standpoint.

And that laser light sources are being vertically integrated via M&A.

eg. $AVGO / $LITE / $COHR / $AAOI (vertically integrated lasers into their own optical programs)

Innolight/Eoptolink/Credo/Cambridge/Jabil/etc. Buy CW lasers externally for pluggables, and these types would be the type to pull off a M&A.

TLDR:
- supplier research was kinda a nothingburger, but just posting it anyway since I got curious
- $CRDO architecture uses few lasers/module which looks bearish, but nuance is that it increases overall CW laser demand regardless
- Credo is buying external CW lasers for their 1.6T offering, so that's bullish for the remaining small pool of CW merchant laser suppliers (since markets have another buyer)

If you're wondering about markets:
- Very macro focused right now, don't have a clue about market reaction to FED decision tomorrow, we'll see.

翻译：
$CRDO 今日宣布其新款 1.6T 光收发模块。以下是一些想法 / Dustphotonics 相关内容：

显然 Credo 和高贝塔公司明天将因加息和宏观担忧被打压。

再加上从八月到九月这些公司的空头持仓大幅上升。因此整体市场反应相当平淡。

但我出于好奇，调查了他们对 DustPhotonics

2026-09-15 12:00

原文：
$CCXI releases their Digit 5 next-generation humanoid.

Thoughts? https://t.co/r0Gb05cnd5

翻译：
$CCXI 发布了他们的 Digit 5 下一代类人机器人。

有什么想法？ https://t.co/r0Gb05cnd5

2026-09-15 08:59

原文：
哇,中国的 AI 视频生成能力可能已经是世界第一了吧？字节跳动的 Seedance 也太强了, 强到让我开始担心…

日本的动画师会不会很快就被 AI 取代？

但想象一下你可以随时生成新的剧情,让故事一直延续下去, 想看多久就看多久

这还挺酷的

翻译：
哇，中国的 AI 视频生成能力可能已经是世界第一了吧？字节跳动的 Seedance 也太强了，强到让我开始担心…

日本的动画师会不会很快就被 AI 取代？

但想象一下你可以随时生成新的剧情，让故事一直延续下去，想看多久就看多久

这还挺酷的

2026-09-15 01:54

原文：
Some color on current memory environments for the $MU / $SKHY bros:

"There is zero risk of memory prices dropping" per Team Group GM Gerry Chen.

- Memory module makers have received notices warning of drastic cuts to their 2027 supply allocations. Some could receive 50% or less of their 2026 allocation.

- Upstream manufacturers have reportedly sold out nearly all of their 2027–2028 capacity.

- DDR5 supply, in particular, is projected to come in at less than 80% of previous expectations.

- Chen expects prices to continue rising 20–30% quarter-over-quarter into 2027 (source Digitimes)

- Chen's commentary on $MU is that even before their new MegaFab capacity comes online Q4 2027 at the earliest, orders for 2028 through 2029 are already fully booked.

So that there's virtually no relief in sight to current shortages.

Sharply lower allocations, fully booked capacity into 2029, and future price hikes doesn't quite sound like the AI buildout is slowing down.

翻译：
关于 $MU / $SKHY 兄弟们当前的内存环境，来点信息：

“内存价格绝不会下跌”，Team Group GM Gerry Chen 如是说。

- 内存模组厂商已收到通知，警告其 2027 年供应配额将被大幅削减。有的甚至只能拿到 2026 年配额的 50% 或更少。

- 上游制造商据称已几乎售罄 2027–2028 年的产能。

- DDR5 供应预计将低于此前预期的 80%。

- 陈预计价格将在 2027 年每季度环比上涨 20–30%（来源 Digitimes）。

- 陈对 $MU 的评论是，即使他们的新 MegaFab 产能最早要到 2027 年第四季度才上线，2028 至 2029 年的订单已经全部订满。

因此目前的短缺几乎没有缓解的迹象。

配额大幅削减、产能到 2029 年已订满、未来还会涨价，这听起来 AI 需求并没有放缓。

2026-09-15 00:42

原文：
I've seen comments from Jim Cramer today supporting hedge funds that short Leopold's "nonsense" $SNDK / $DRAM / SK Hynix type calls.

As well Cramer calling $CRWD a "Must Buy".

I'm not sure if Jim Cramer sees a $241B company labelled "Cybersecurity" and thinks of it as safe... while ignoring all their fundamentals.

But with Crowdstrike, you have a ~145x FY2028 fwd P/E company, growing revenue 26% Y/Y. That Cramer thinks is a "must buy".

And then you turn your eyes to the AI capex beneficiary theme that Leopold is long on.

With names like $SKHY / Samsung in $DRAM... Both FY2027 ~2.8-3x fwd P/E off analyst projections.

-> Samsung LTAs to 2031 (70% capacity) for durability... -> Minimum price floor + no price ceiling type contracts from SK Hynix.
-> Memory shortage much tigher for the next year.
-> DRAM/NAND % of hyperscaler capex goes higher (est. 68% per Trendforce)
-> 2027 hyperscaler capex projections all hiked.

Then with $SNDK, 50% FCF margins, 80% gross margin projections into 2030.

Only Situational Awareness position I'm in disagreement with is $CRWV due to debt interest eating FCF, but I can make a similar bullish argument with $BE, $INTC, $AMD, and others.

Not quite sure about short-term, but over the medium run:

It's hard to not see Leopold's positions pulling an Inverse Cramer due to overwhelming fundamentals.

翻译：
我今天看到 Jim Cramer 的评论，支持做空 Leopold 所谓的“胡说” $SNDK / $DRAM / SK Hynix 类股票的对冲基金。
Cramer 还把 $CRWD 称为“必买”。

我不确定 Jim Cramer 是否把一家市值 2410 亿美元、标榜为“网络安全”的公司视为安全……却忽视了它的基本面。

但 Crowdstrike 的市盈率约为 FY2028 前瞻 145 倍，收入同比增长 26%。Cramer 却认为它是“必买”。

再看 Leopold 持

2026-09-14 11:20

原文：
$IREN CEO on Anthropic call for frontier slowdown:

“Even if models never improved from here…

just rolling out what they can already do would take more compute than the world can build for years”

I agree, and current narrative is reminiscent of the $NVDA DeepSeek selloff or the $GOOGL TurboQuant scare.

Guess what happened?

These scares once or twice a year all turn out to be immaterial to the overall AI buildout.

Especially when we have contracted demand all the way into the 2030’s now from advanced packaging to memory procurement.

翻译：
$IREN CEO 对 Anthropic 呼吁前沿放缓的回应：

“即使模型从此不再提升…
仅仅部署它们已经能做的事，也需要的算力超过全球多年能建造的算力”

我同意，当前的叙事让人想起 $NVDA DeepSeek 抛售或 $GOOGL TurboQuant 恐慌。

猜猜会怎样？

这些恐慌一年一两次，最终对整体 AI 建设影响微乎其微。

尤其是我们已经签订了从先进封装到存储采购一直延伸到 2030 年的需求合同。

2026-09-14 08:17

原文：
NOR flash “Severe Supply Shortage” with high density continued increases of 90-110% for H2.

I wonder who sells both high density NOR and SLC NAND? *cough cough, $ESMT*

For H1 hikes:
NOR: +100–120%
SLC NAND: +130–150%

For H2 hikes:
High-density NOR: 90-110% in H2
SLC NAND: ~70–75%.

Obviously there’s others from Winbond / Macronix / Gigadevice, but my idea and position with ESMT is that it’s the largest beneficiary of the shortage relative to size. (they sell NOR: >256Mb)

The moderate bearish part is:

“For mid- and low-capacity products of 128Mb and below, additional capacity from Chinese suppliers is gradually coming online.

Price increases for consumer-grade products are therefore expected to moderate”

However, this update is overwhelmingly bullish for earnings of legacy memory players.

Let the ASP hike games continue.

翻译：
NOR flash “严重供应短缺”，高密度在 H2 仍将上涨 90-110%。

我想知道谁同时销售高密度 NOR 和 SLC NAND？*咳咳，$ESMT*

H1 涨幅：
NOR：+100–120%
SLC NAND：+130–150%

H2 涨幅：
高密度 NOR：90-110%（H2）
SLC NAND：约 70–75%。

显然还有 Winbond / Macronix / Gigadevice 等供应商，但我对 ESMT 的看法和仓位是，它相对于规模是此次短缺的最大受益者。（他们销售的 NOR >256Mb）

适度的看空因素是：

“对于 128Mb 及以下的中低容量产品，来自中国供应商的额外产能正逐步投产。

因此消费级产品的价格上涨预计会趋缓”

然而，此次更新对传统存储厂商的业绩极为利好。

让 ASP 涨价游戏继续。

2026-09-14 06:57

原文：
Man Ray: are you long Samsung?

Patrick: Yup.

Man Ray: Samsung has memory LTAs extending to 2031?

Patrick: Yup.

Man Ray: And it’s trading at ~2.8× 2027E P/E?

Patrick: Yup.

Man Ray: So you’re buying Samsung?

Patrick: No, I’m selling on a 4% drop because X said memory will crash from AI slowdown

翻译：
Man Ray: 你持有三星多头吗？
Patrick: 是的。
Man Ray: 三星的存储长期协议延伸到2031年吗？
Patrick: 是的。
Man Ray: 而且它的交易估值约为2027年预期市盈率的2.8倍？
Patrick: 是的。
Man Ray: 那么你在买入三星吗？
Patrick: 不，我在跌幅4%时卖出，因为X说内存会因AI放缓而崩盘。

2026-09-13 22:27

原文：
Be Anthropic:

Call for: “limiting the ingredients that go into frontier models, such as training compute”

> secure $517 Billion in compute deals over last 11 months

Proclaim: “We have tried to prioritize caution over speed and prudence over profit.”

> actively speedrun NASDAQ IPO in October while others like OpenAI delayed theirs to 2027

Suggest: “employee-like access to a team of embedded third-party evaluators” for regulatory oversight of frontier labs

> after becoming one of the two leaders…

翻译：
成为 Anthropic：

呼吁：“限制进入前沿模型的成分，例如训练算力”

> 在过去 11 个月中确保 $5170 亿美元的算力交易

宣称：“我们尝试将谨慎置于速度之上，将审慎置于利润之上。”

> 在十月积极加速完成 NASDAQ IPO，而其他如 OpenAI 则将其推迟至 2027 年

建议：“为前沿实验室的监管监督提供类似员工的嵌入式第三方评估团队访问”

> 在成为两大领袖之一之后…

2026-09-13 20:12

原文：
I assign essentially 0% probability that Anthropic or OpenAI will slow down frontier development in any meaningful way.

And that majority of the "regulation" they try making is regulatory capture to slow down competition.

1. Anthropic/OpenAI will not willingly hand over their lead to Gemini, xAI, Meta, or another frontier lab.

-> Elon will want a slowdown to catch up with OpenAI.
-> OpenAI will want a regulatory framework to slow down xAI + competition.

Every lab has an incentive to support rules that constrain competitors more than they constrain itself

2. The US is not going to meaningfully slow frontier AI development while China continues advancing. (as seen with Trump's post today about rejecting AI slowdown).

It would be probably the most flagrant national security risk if China leapfrogged the US in capability.

Especially when both sides have called frontier AI capabilities "cyber nuclear weapons"

3. China will not slow down frontier AI development. That is a given.

It will be theater if any of these parties agrees so publicly, because behind the scenes:

-> They'll likely keep on progressing, whether it's OpenAI, the US Government, or China.

If they wanted to slow down, they could have done so themselves at any point in time:

Yet everyone is still fighting over the scarce resources to run AI as seen with memory procurement from Samsung going all the way to 2031.

Or the $NVDA / $AVGO supply constrained 70-100%+ Y/Y growth protections.

So I suggest people unfollow the clowns on X that predict AI stocks from $NVDA to $TSM to $MU will crash for engagement.

All this does is cause retail to panic short term and hand over the keys to the kingdom over to others.

翻译：
我几乎把Anthropic或OpenAI会以任何有意义的方式放慢前沿研发的概率定为0%。

而且他们所谓的“大多数监管”都是监管俘获，用来压制竞争。

1. Anthropic/OpenAI不会主动把领先优势让给Gemini、xAI、Meta或其他前沿实验室。

-> Elon

2026-09-13 13:24

原文：
Tech people in real life: I have shares in “ $TSLA, $GOOGL, $SPCX. “

People on my X feed/comments:

- Leveraged Micron, Sk Hynix calls all the way from $1M to $28M down to $4M

- calls on short dated tanker freight futures because $AXTI 3000% 1Y returns wasn’t enough.

- Long $PLTR short $LULU because inverse Burry indicators

- Long Swedish high beta CW laser suppliers for CPO because it reminds them of $NVDA bottlenecking EML

- “AI is going to end human civilization” so this is a signal for longer AI capex duration + trade wars so long $EQR cause tungsten.

Gap between finance X and normal finance/tech people is also insane.

I can’t help but feel proud that I’m partly to blame.

翻译：
现实中的技术人士：我持有 $TSLA、$GOOGL、$SPCX 的股票。

我在 X 上的动态/评论里的人：

- 从 $1M 到 $28M 再到 $4M 的杠杆 Micron、SK Hynix 做空

- 做空短期油轮运费期货，因为 $AXTI 1 年 3000% 回报还不够

- 做多 $PLTR 做空 $LULU，因为逆向 Burry 指标

- 做多瑞典高贝塔 CW 激光供应商用于 CPO，因为它让人想起 $NVDA 对 EML 的瓶颈

- “AI 将终结人类文明”，于是这被视为 AI 资本支出周期更长 + 贸易战的信号，因而做多 $EQR 因钨

金融 X 与普通金融/技术人士之间的差距也太大了。

我不禁感到自豪，因为我也有点儿责任。

2026-09-13 08:12

原文：
It’s the weekend but surprised I missed this from $ORCL earnings.

Implications for $NBIS to $IREN are pretty positive since:

Oracle said all GPU capacity for renewal were resold +20% premium above previous contracts.

Majority of the equipment was 4Y+ and older…

So another L for the Burry’s rapid GPU depreciation thesis.

And W for $NVDA and the Neocloud party.

翻译：
周末了，但我惊讶自己错过了 $ORCL 财报中的这点。

对 $NBIS 到 $IREN 的影响相当正面，因为：

Oracle 表示，所有待续约的 GPU 容量都以比之前合同高出 20% 的溢价转售。

大部分设备已使用 4 年以上…

这又为 Burry 的快速 GPU 折旧论点再添一个 L。

而对 $NVDA 和 Neocloud 派对来说是 W。

2026-09-12 08:56

原文：
I think there’s a gap to fill…

Since I believe retail deserves the same research and early thematic ideas that institutions access first.

I feel like everyone understandably gravitates institutional as they grow.

But helping retail has always been my priority. And that hasn’t changed even as I grew to 1M followers.

I’m excited to see what we can cook up together soon chat.

翻译：
我觉得还有一个空白需要填补……

因为我相信散户应当获得与机构第一手获取的相同研究和早期主题想法。

我觉得随着成长，大家自然会倾向于机构。

但帮助散户一直是我的首要任务。即使我的粉丝增长到 1M，这一点也未改变。

我很期待看到我们很快一起聊天并共同打造的成果。

2026-09-11 17:57

原文：
What's the next $BWET if you guys have ideas?

Need more weird ideas like short dated tanker freight future ETFs...

I didn't know this existed until last week. https://t.co/kguL6svZNj

翻译：
如果你们有想法，下一只 $BWET 会是什么？

需要更多奇怪的想法，比如短期油轮运费期货ETF……

我直到上周才知道这东西的存在。 https://t.co/kguL6svZNj

2026-09-11 17:15

原文：
Apparently my "bottleneck thesis" has made it into a quant fund research pipeline?

Hope the institutions enjoy reading my takes on DDR2, CW lasers, and InP substrates...

Didn't expect to end up in the same paragraph as Trump on a Friday. https://t.co/j5leamPwJV

翻译：
显然我的“瓶颈论”已经进入量化基金的研究流程？

希望机构们喜欢阅读我对 DDR2、CW 激光器和 InP 基板的见解……

没想到会在周五的同一段落里和特朗普并列。 https://t.co/j5leamPwJV

2026-09-11 15:16

原文：
Looks like Situational Awareness is back:

Sources say Leopold bought options in $AMD, $BE, $CRWV, and $DRAM.

And was active in other stocks such as $SKHY, and $SNDK.

Little surprising since option IV across the board is quite high. https://t.co/P2Pg5szbvO

翻译：
看起来情境意识又回来了：

消息称 Leopold 在 $AMD、$BE、$CRWV 和 $DRAM 上买入了期权。

并且在其他股票如 $SKHY 和 $SNDK 上也很活跃。

有点令人惊讶，因为整体期权隐含波动率相当高。 https://t.co/P2Pg5szbvO

2026-09-11 12:12

原文：
X should go out and ban every network ring that pays others to quote/comment on their posts.

If you want to see how it’s done:

> companies pay influencers to comment/quote their posts (now sometimes with labels)

> they spawn new accounts like ___ “finance” or ___ “”with a new affiliated account

> when each company gets big enough they just interact with one another eg. quote posts to boost engagement.

By the end each affiliate account ends up with tens/hundreds of thousands of followers.

And they use it to boost their other account visibility.

My worry is that X is just going to end up dominated by engagement networks marketing company products.

Instead of having independent authentic voices.

翻译：
X 应该出台禁令，封禁所有付费让他人引用/评论其帖子的大型网络圈子。

如果想看看是怎么运作的：

> 公司付费给网红，让他们评论/引用公司的帖子（现在有时会加标签）

> 他们会创建新账号，例如 ___ “finance” 或 ___ “”，并关联一个新账号

> 当每家公司规模足够大时，它们就会相互互动，例如引用帖子以提升互动率。

最终每个关联账号会拥有数万甚至数十万粉丝。

他们利用这些粉丝提升其他账号的曝光度。

我担心 X 最终会被营销公司产品的互动网络所主导。

而不是拥有独立真实的声音。

2026-09-11 10:33

原文：
Haha I feel bad for the guy at the $SIVE booth.

Genuinely thanks everyone for helping crowd source channel checks, really appreciate it!

Think the most interesting update overall from Innolight channel checks was explicit confirmation that lasers across 70mW-200mW power range have seen ASP going up.

Sivers also disclosed they raised prices.

That should support gross margin expansion for laser suppliers in future quarters as higher pricing flows through.

翻译：
哈哈，我为 $SIVE 展位的那位感到难过。

真心感谢大家帮助众包渠道检查，真的很感激！

我认为 Innolight 渠道检查中最有趣的更新是明确确认 70mW-200mW 功率范围的激光器的平均售价在上升。

Sivers 也透露他们提价了。

这应当在未来几个季度支持激光供应商的毛利率扩张，因为更高的价格会传导下去。

Loading...