觅信聚合

我的关注

2026-08-29 06:13

原文：
If you're curious what I'm doing this weekend:

I'm trying to piece together who has Powerchip (PSMC) wafer allocations and also serves DDR2/DDR3 markets.

then mapping those allocations against operating income inflections / MC to see which ones are most compelling.

So far I found ESMT to be the most interesting, since a lot of the others trade around 4 p/e based on most recent Q2's annualized (ESMT was 1.9x based on July)...

But THE MOST IMPORTANT THING IS July's ESMT earnings gives a good hint about the other's Q3 operating income reports.

The reason I think this trade hasn't been very popular so far is:

> Barely anyone knows about it + limited coverage, esp. around allocations
> High risk of normalized prices, but I do think it tightens further throughout 2027 from my research.
> Hard to model effects of PSMC 45% wafer hikes + inventory (if its sustainable) (I did the math on ESMT, but haven't done so about the rest), which others might be more uncertain about.
> If July was one-off or not (if August earnings delivers $100m+ net income, I think it's going to cook super hard).

But I think markets would do all the price discovery before the earnings so that's where I'm just taking a guess here.

So I'm just going down line by line to double check and I did find a few I missed, so just going through their filings.

I do think this theme can cook given the extreme DDR2 price hikes recently + Winbond withdrawing, so I'm doing heavy research into it.

If you're curious how my current AI portfolio is roughly structured, kinda looks like this rn (not a recommendation):

- 50% photonics/networking
- 25% memory
- 10% robotics + upstream
- bunch of misc (energy, MLCCs, software, non-ai, random exposure like GCS, QD, substrate, etc)

on ~1.3x share only margin, probably only 2.3% of my portfolio is options.

The individual names I'm in are high-beta

2026-08-29 06:13

enough so any additional leverage would be pretty bad.

翻译：
如果你想知道我这个周末在做什么：

我在试着梳理到底是谁拿到了 Powerchip（PSMC）的晶圆配额，同时还覆盖 DDR2/DDR3 市场。

然后把这些配额和营业利润拐点 / 市值对比起来，看看哪些最有吸引力。

到目前为止，我觉得 ESMT 最有意思，因为其他很多标的按最近一季 Q2 年化算，估值大概在 4 倍左右（ESMT 按 7 月数据算是 1.9 倍）……

但最重要的是，ESMT 7 月的业绩能很好地暗示其他公司 Q3 的营业利润表现。

我觉得这个交易到现在还没那么火，原因是：

> 几乎没人知道这件事 + 覆盖很有限，尤其是配额这块
> 归一化价格下行风险很高，但按我的研究看，我确实认为到 2027 年这点还会继续收紧。
> 很难建模 PSMC 提价 45% + 库存的影响（如果能持续的话）（我已经算过 ESMT 的，但还没算其他几家），这点别人可能更没把握。
> 7 月是不是一次性因素（如果 8 月利润能做到 1 亿美元以上，我觉得会非常猛）。

不过我觉得市场会在财报前就把价格发现做完，所以我这里只是在猜。

所以我现在就是逐行往下查，重新核对一遍，确实也找到几家之前漏掉的，于是就继续看它们的公告。

我确实觉得这个主题有机会大爆，最近 DDR2 价格暴涨，再加上 Winbond 撤出，所以我在重度研究这件事。

如果你想看我现在的 AI 组合大概是怎么配置的，差不多是这样（不构成建议）：

- 50% 光子/网络
- 25% 存储
- 10% 机器人 + 上游
- 一堆杂项（能源、MLCC、软件、非 AI、以及像 GCS、QD、基板之类的随机敞口）

在大约 1.3 倍股票多头保证金下，我的组合里大概只有 2.3% 是期权。

我持有的这些个股 beta 都很高，再加任何杠杆都不太好。

2026-08-28 10:24

原文：
只是因为这张图浏览量太高了…

白毛牛来 https://t.co/WiHhXVqtK5

翻译：
白毛牛来 https://t.co/WiHhXVqtK5

2026-08-28 10:06

原文：
白毛牛来? https://t.co/xk1CM3TVCL

翻译：
白毛牛来了？ https://t.co/xk1CM3TVCL

2026-08-28 09:15

原文：
Are Frontier lab relative valuations are too absurd nowadays?

If Anthropic IPOs at $2T… 10% of that can buy up almost every major established consumer brand:

From:

Taco Bell/Pizza Hut/KFC, GAP, American eagle, Levi’s, Victoria’s Secret, Cheesecake Factory, Krispy Kreme…

Calvin Klein, Kohls, AMC, Kura Sushi, Dennys, Nike, Build a Bear, Jack in the Box, IHOP, Papa John’s…

Shake Shack, Apple Bees, Dave & Busters, Under Armour, Petco, Canada Goose.

All together. Using 10% of Anthropic.

And they’d still have $51.6B leftover.

翻译：
现在 Frontier Lab 的相对估值是不是太离谱了？

如果 Anthropic IPO 估值 2 万亿美元……其中 10% 就能买下几乎所有主要的成熟消费品牌：

全部加起来，只用了 Anthropic 估值的 10%，

他们仍然会剩下 $51.6B。

2026-08-28 08:21

原文：
所以… 回到存储这边我们现在有：

$SKHY CEO 表示: “我们预计[存储]短缺将持续到 2030 年底”

$SNDK 表示: “我们看到直到 2030 年，NAND 都存在结构性的巨大需求”非 GAAP 毛利率将维持在约 80%.

$NVDA 表示供应承诺从 1,190 亿美元增加到 2,790 亿美元，主要由存储采购推动…与此同时, 存储定价处于“极端”水平而且还在继续上涨.

Winbond 正在与多家客户就延伸至 2029-2030 年的配额展开讨论.

我之前就说过, 存储需求看起来是结构性的, 而现在从传统 DRAM 到 HBM，需求可见度已经延伸到了未来数年..

所以我确实认为, 整个板块的远期 p/e 倍数都有进一步上升的机会.

翻译：
所以… 回到存储这边我们现在有：

$SKHY CEO 表示：“我们预计[存储]短缺将持续到 2030 年底”

$SNDK 表示：“我们看到直到 2030 年，NAND 都存在结构性的巨大需求”，非 GAAP 毛利率将维持在约 80%。

$NVDA 表示供应承诺从 1,190 亿美元增加到 2,790 亿美元，主要由存储采购推动…与此同时，存储定价处于“极端”水平而且还在继续上涨。

Winbond 正在与多家客户就延伸至 2029‑2030 年的配额展开讨论。

我之前就说过，存储需求看起来是结构性的，而现在从传统 DRAM 到 HBM，需求可见度已经延伸到了未来数年…

所以我确实认为，整个板块的远期市盈率倍数都有进一步上升的机会。

2026-08-28 07:37

原文：
It’s times like these where it would be nice to run a $5B hedge fund…

So you could just buy enormous stakes in companies when the market sells off on stupid commentary.

翻译：
正是这种时候，拥有一家 $5B 对冲基金会很不错…

于是，当市场因愚蠢的言论而抛售时，你只需买入公司巨额股份。

2026-08-28 06:54

原文：
Here's the angle I'm looking at $SIVE at:

Companies/CSPs have been going out of their way way to lock up any qualified CW laser capacity for optical transceivers.

So now, Sivers is coming into the industry with:

- Large capacity from Win Semi
- "tremendous capacity that is available now" from a new foundry (likely qualifications since 2024) since it's engaged "for a while"

And what excited me from this earnings was the "6 pluggable/module" engagements (which is an enormous amount).

The nuance I'm reading is:

- $COHR said it sees no near term ability to sell InP lasers externally because internal transceiver demand consumes all available
- $AAOI said the same about consuming laser capacity internally (wasn't a major merchant supplier before though).
- $LITE has been bottlenecked and been buying lasers off the open market

Your previous merchant players rerouted laser capacity internally.

So a lot of the bigger names (eg. Eoptolink/Innolight as just a random example) are probably looking to source more lasers.

And that kinda matches the quote "capable of very rapid qualification and ramp" (which would not match Series B startup)...

Lot of people are asking why aren't there LTAs to 2030 then?

-> You can't just randomly escape the qualification process that established players have already completed.

Why aren't the customers disclosed?

And as seen with the $MRVL + $POET engagement, you can't just disclose the vendors you're working with.

But the "$1.2B opportunity pipeline" almost doubled relative to the jump of $JBL + $GFS. So it's signals that the new pluggable engagements might be pretty substantial relative to Jabil.

So if $SIVE comes along with enormous amounts of CW DFB laser capacity during a supply shortage...

The industry conditions have changed in a major way that increases conversion rates of engagements.

And with the sheer size

2026-08-28 06:54

from all your ~est. customers jabil, globalfoundries, poet, aeva, lightium, ayar, (maybe lightmatter, celestial, lightelligence), 6 other pluggable players, and others.

I think Sivers is going to cook after connecting the dots.

翻译：
我现在关注 $SIVE 的这个角度：

各家公司/CSP 一直在想方设法锁定任何合格的 CW 激光器产能，用于光收发器。

所以现在，Sivers 带着这些进入了行业：

- 来自 Win Semi 的大产能
- 来自一家新代工厂的“当前可用的巨大产能”（大概率是自 2024 年起就已进入资质认证），因为双方已经“合作了一段时间”

而让我在这次财报里最兴奋的是“6 个可插拔/模块”合作机会（这数量非常惊人）。

我读到的细微信号是：

- $COHR 说，短期内没有能力对外销售 InP 激光器，因为内部收发器需求已经吃掉了所有可用产能
- $AAOI 也说了同样的话，内部在消耗激光器产能（不过它之前并不是主要的商用供应商）
- $LITE 一直被产能卡住，不得不去公开市场买激光器

你们这些以前的商用玩家，已经把激光器产能转去内部用了。

所以很多大厂（比如随便举个例子：Eoptolink / Innolight）大概率都在想办法采购更多激光器。

这也和那句“具备非常快速的认证和放量能力”对得上（这可不像一家 B 轮初创公司会有的表述）……

很多人在问，为什么合同不到 2030 年的长期协议（LTA）还没有？

-> 因为你不可能随便绕过那些成熟玩家已经完成的认证流程。

为什么不披露客户？

而且从 $MRVL + $POET 的合作也能看出来，你不能直接披露你正在合作的供应商。

但“12 亿美元机会管线”相比 $JBL + $GFS 的增长，几乎翻倍了。所以这说明新的可插拔合作机会，规模相对 Jabil 可能相当可观。

所以如果 $SIVE 在供应短缺期间，带着海量的 CW DFB 激光器产能进入市场……

行业环境已经发生了重大变化，这会提高合作转化率。

再加上你那一大堆大致估算的客户：jabil、globalfoundries、poet、aeva、lightium、ayar、（也许还有 lightmatter、celestial、lightelligence）、另外 6 家可插拔玩家，以及其他客户。

我觉得 Sivers 把这些线索串起来之后，会非常能打。

2026-08-27 21:25

原文：
Nah, $SIVE is the most unusual $1B photonics company I've ever seen.

Just for AI DCs alone:

- Reference laser for $GFS SCALE (CPO/NPO/pluggable)
- 7 pluggable engagements disclosed, with $JBL as primary ramp (with H1 2027 timeline)
- ELS path with O-Net
- ELS path with $POET
- Lasers to Ayar for CPO I/O (now in $NVDA nvlink ecosystem)
- Lasers for $AEVA (which now has NPO agreements with optical engine providers for hyperscalers)
- Celestial/Lightmatter/Lightelligence (likely customers in 2023/2024)
- InP lasers on TFLN with Lightium

Then it has 2 substantial allocations for CW DFB laser capacity during an industry shortage.

As well as one of the few CPO-grade laser suppliers out in the industry.

Let me know what other player around this range has so many ongoing qualifications, a large TAM, and in photonics.

$SIVE is the #1 for me in terms of MC relative to qualifications paths across the industry.

翻译：
不，$SIVE 是我见过的最不寻常的 10亿美元光子公司。

仅在 AI 数据中心方面：

- 为 $GFS SCALE 提供参考激光（CPO/NPO/可插拔）
- 已披露 7 项可插拔合作，$JBL 为主要推进（H1 2027 时间表）
- 与 O-Net 的 ELS 路径
- 与 $POET 的 ELS 路径
- 向 Ayar 提供用于 CPO I/O 的激光（现已进入 $NVDA nvlink 生态）
- 为 $AEVA 提供激光（其已与光学引擎供应商签订 NPO 协议，服务超大规模用户）
- Celestial / Lightmatter / Lightelligence（预计在 2023/2024 年成为客户）
- 与 Lightium 合作的 TFLN 上的 InP 激光

此外，它在行业短缺期间拥有 2 项大规模的 CW DFB 激光产能分配。

也是业界为数不多的 CPO 级激光供应商之一。

告诉我还有哪个同档次的公司拥有如此多的在研资格、庞大的 TAM，并且专注光子领域。

$SIVE 在我看来是行业内相对于资格路径的 MC 排名第一的公司。

2026-08-27 19:16

原文：
$SIVE ER transcript just dropped, TLDR:

- 6 NEW pluggable players working with Sivers now.

Probably the biggest news of the entire ER next to new foundry allocations.

"3 are in alpha sample evaluation stage" and another 3 in technical engagement / supply assessment stage."

And now it makes more sense why their $1.2B revenue pipeline ballooned (since this was a bigger leap than Jabil).

- Initial production orders with $JBL expected in the first half of 2027, with the production ramp planned second half of 2027

LFG, Jabil is a massive hyperscaler supplier and finally got clear timelines on revenue from volume ramp.

- "new foundry partner who has brought on tremendous capacity that is available now"

This is what I wanted to hear. Apparently they've been in the works for this for awhile since it's "available now", and THIS IS VERY MATERIAL.

CW lasers are in a massive shortage and $SIVE brought on new supply outside of Win Semi (also de-risks).

As you hear with other qualified CW players (anything they make gets sold), so as Sivers partners finishes their qualifications, I'm expecting the same.

- "long-term capacity model where one-third of manufacturing capacity will be internal, while two-thirds will come from our foundry partners"

Lukewarm on this, it makes sense they need to be vertically integrated like $AAOI / $LITE eventually but implies more capex (better after NASDAQ listing).

As you see with ESMT + DDR2 bottlenecks, the operating income they get from just securing wafers during shortages and doing fabless models is incredible.

And it would make more sense to fund this with cashflow down the road.

- NASDAQ listing ongoing. CFO gave a very lawyer like answer, but on track as usual.

- "We do not see production capacity as a bottleneck at this point in time"

This is very meaningful since with their new foundry partners, implies $SIVE is

2026-08-27 19:16

coming online with a very material CW laser supply to a bottlenecked market.

TLDR:
- NEW substantial FOUNDRY ALLOCATION! (very, very material during CW laser shortage)
- 6 new pluggable players outside of $JBL
- Clear revenue ramp timelines from Jabil

Heavy focus on pluggables -> NPO -> CPO seems like the progression. Only lukewarm piece was building up internal capacity but it's long term positive.

Anyway, very happy after reading the transcript aside from potential capex prioritizing laser capacity (which is fine too during an industry shortage) over IP acquisition.

The 6 new pluggable players + substantial wafer allocation "available now" is a pretty insane disclosure.

翻译：
$SIVE ER 记录刚刚发布，简要概述：

- 现在有 6 家新的可插拔合作伙伴与 Sivers 合作。

这可能是整个 ER 中仅次于新代工分配的最大新闻。

“其中 3 家处于 alpha 样本评估阶段”，另外 3 家处于技术对接/供应评估阶段。

这也解释了为何他们的 12 亿美元收入管道会膨胀（这比 Jabil 的增长更大）。

- 预计在 2027 年上半年收到来自 $JBL 的首批生产订单，生产提升计划在 2027 年下半年进行。

LFG，Jabil 是大型超大规模供应商，终于获得了量产提升的明确收入时间表。

- “新的代工合作伙伴已经带来了巨大的产能，现已可用”

这正是我想听到的。显然他们已经筹备了一段时间，因为现在已经可用，这非常重要。

连续波激光器正处于严重短缺，$SIVE 引入了除 Win Semi 之外的新供应（也降低了风险）。

正如其他合格的连续波激光器厂商所表现的（他们生产的任何产品都会被抢购），随着 Sivers 合作伙伴完成资格认证，我预计情况会相同。

- “长期产能模型为：三分之一的制造产能内部化，另外三分之二来自我们的代工合作伙伴”

对此持温和态度，这说明他们最终需要像 $AAOI / $LITE 那样实现垂直整合，但也意味着更多资本支出（在纳斯达克上市后会更好）。

正如在 ESMT + DDR2 瓶颈中看到的，仅在短缺期间确保晶圆并采用无晶圆厂模式就能获得惊人的营业收入。

这更适合以后用现金流来融资。

- 纳斯达克上市正在进行中。财务总监给出了非常律师式的回答，但进展如常。

- “我们目前不将生产产能视为瓶颈”

这意义重大，因为有了新的代工合作伙伴，意味着 $SIVE 将以非常重要的连续波激光器供应进入一个供不应求的市场。

简要概述：
- 新的重大代工分配！（在连续波激光器短缺期间极其重要）
- 6 家新的可插拔合作伙伴，除 $JBL 外
- 来自 Jabil 的明确收入提升时间表

重点放在可插拔 → NPO → CPO 的进程上。唯一稍显温和的是内部产能的建设，但这是长期的正面因素。

总的来说，阅读完记录后非常满意，除了可能将资本支出优先于激光产能（在行业短缺期间这也无妨）而非 IP 收购之外。

6 家新的可插拔合作伙伴加上“现已可用”的大量晶圆分配，实在是惊人的披露。

2026-08-27 16:36

原文：
$SIVE earnings just got released:

$1.2 billion revenue opportunity pipeline growth up 268% from December (incredible growth).

Last ER: $799M (this was a surprise, likely linked to Jabil)

July 2026: ~$1.2B (hinting more either more customer engagements than what's announced or larger volume ramps within programs).

"Sivers continued to reallocate resources from NRE projects toward product ramp preparation.

Sivers expects the impact of this transition to become visible in Q4 2026 and accelerate through 2027 as multiple programs progress toward volume production."

Makes sense on timelines.

Current financials reflect a qualification-stage player before volume ramp:

- $5.64M in revenue (reducing NRE activity, shifting to volume ramps soon)
-$3.72M in adjusted EBITDA (what to look at given one-off charges, eg. social security accounting expense)

Key thing to look at is earnings call in 30 minutes, very excited so far.

翻译：
$SIVE 财报刚刚发布：

$12亿美元收入机会管线自12月起增长268%（惊人增长）。

上次财报：$7.99亿美元（这是个意外，可能与捷普有关）

2026年7月：约$12亿美元（暗示客户合作比已公布的更多，或项目内的产量提升更大）。

“Sivers 继续将资源从非经常性工程项目重新分配到产品产能准备。

Sivers 预计此转变的影响将在2026年第四季度显现，并在2027年加速，因为多个项目向量产推进。”

时间线上合理。

当前财务显示为量产前的资格阶段公司：

- $5.64M 收入（正在减少非经常性工程活动，近期将转向量产）
- $3.72M 调整后 EBITDA（考虑一次性费用，如社会保障会计费用）

关键是30分钟后的盈利电话会议，目前非常期待。

2026-08-27 07:12

原文：
我感觉我现在大部分粉丝都是讲中文的…

翻译：
我感觉我现在大部分粉丝都是讲中文的……

2026-08-27 05:47

原文：
I've always been bullish on memory with $MU / Samsung / solana:SKHYhSjuRWHgikq8eRKbtBbpABgJSkd7ytQV14i9EQ3 for the 2026-2027 period.

Same opinion, even after $NVDA call.

And I've been fervently defending how high end memory demand is structural + OP for operating income (especially during Iran tensions around LNG/Helium).

Nvidia earnings just reaffirms what we know about extreme demand since commitments went from $119B -> $279B largely driven by memory procurement.

Nvidia's CFO also said: "We are experiencing extreme pricing conditions in memory."

As for what I've done, H1 2026 I was extremely overweight in memory:

With $MU, $SNDK, Phison, $SIMO, Nanya, Macronix, Winbond, and $EWY / SK Hynix (HBM/DRAM + NAND + legacy DRAM/NAND + controllers + NOR Flash).

I trimmed down those positions aside from Samsung/SK Hynix longs, since I do believe many have been rerated (eg. Micron $300 -> $1000+ already).

I think the largest price discovery period has played out, but just a waiting game for the operating income to catch up (esp u samsung)

And I used that period to go overweight on photonics.

But I do believe we're seeing a relatively newer cascade down into the "legacy legacy" memory like DDR2/DDR3 with the "legacy" players like Winbond leaving some of those segments.

Where the price hikes finally hit the 40-60% Q/Q mark, which reminds me of the extreme $SNDK days, across DDR2/DDR3.

Which is why I started up positions in ESMT (1.9x P/E from July annualized) and Etron.

Maybe we'll see a price discovery moment further down the legacy memory stack (could be wrong), but that's the area I've focused on recently.

翻译：
我一直看多存储，尤其是 $MU / Samsung / solana:SKHYhSjuRWHgikq8eRKbtBbpABgJSkd7ytQV14i9EQ3 这类标的，认为 2026-2027 年会是机会窗口。

即便在 $NVDA 财报后，我的看法也没变。

我一直在强调，高端存储需求是结构性的，而且会明显拉动运营利润，尤其是在伊朗紧张局势影响 LNG/氦气供应的背景下。

Nvidia 的财报再次印证了我们对极端需求的判断，因为其承诺额从 $119B 增至

2026-08-27 05:47

$279B，很大程度上是由存储采购推动的。

Nvidia 的 CFO 还说过：“我们正经历存储领域极端的定价环境。”

至于我自己做了什么，2026 年上半年我对存储是极度超配：

持有 $MU、$SNDK、Phison、$SIMO、Nanya、Macronix、Winbond，以及 $EWY / SK Hynix（HBM/DRAM + NAND + 传统 DRAM/NAND + 控制器 + NOR Flash）。

除了 Samsung/SK Hynix 的多头仓位外，我把这些仓位都减了，因为我确实认为很多标的已经完成重估了（比如 Micron 已经从 $300 到 $1000+）。

我认为最大的价格发现阶段已经过去了，接下来就是等运营利润跟上来（尤其是 Samsung）。

而我也利用那段时间加仓了光子学。

不过我确实认为，我们正在看到一轮相对更新的下沉，进入 DDR2/DDR3 这类“更老一代的老旧存储”，而像 Winbond 这样的“传统”玩家正在退出其中一些细分市场。

当涨价最终达到环比 40-60% 时，这让我想起了 $SNDK 当年的极端行情，尤其是在 DDR2/DDR3 上。

这也是我开始建仓 ESMT（按 7 月年化计算，P/E 约 1.9x）和 Etron 的原因。

也许我们会在更下层的老旧存储链条里看到一次价格发现时刻（也可能我错了），但这就是我最近重点关注的方向。

2026-08-27 01:42

原文：
I wanted exposure to DDR2/DDR3 bottleneck ongoing and found ESMT (3006).

A $2.5B MC fabless company with PSMC wafer allocation (esp. focused on DDR2).

Last month's net income was: $109.5M (July month), which annualized is $1.31B net profit (1.9x runrate P/E).

If we look at net profit throughout the months tracking legacy DRAM price hikes:

Jan 2026: ~$16M (est)
Feb 2026: ~$17M
Mar 2026: ~$31M (est)
April 2026: ~$58M
May 2026: ~$60M (est)
June 2026: ~$67M (est)
Jul 2026: ~$110M
(Estimated months are inferred from reported quarterly totals. July is company reported)

This earnings progression reminds me of $SNDK style price hikes + earnings inflection (esp. the report released this month).

And I do expect DDR2/DDR3 capacity to remain constrained throughout 2027 (with legacy players like Winbond withdrawing from certain lines too like DDR2).

It's not quite all inventory liquidation like the other peers. Sure lower cost inventory helped, but the primary driver seems to be the widening spread between wafer costs/supply and legacy dram ASPs?

Their balance sheet is extremely solid as well:

- Cash on hand: ~$395.9M net cash
- $249.6M inventory, $146.3M receivables

And the net income progression... is before further legacy DRAM hikes expected in Q3.

*disclosure own positions, NFA

It's always a bit daunting being early without much commentary around the idea. Wondering if anyone can stress test this thesis?

Since a company that grew monthly net income from $16M -> $100m this year alone (July Annualized would be 1.9x P/E).

And benefits from another projected wave of DDR2/DDR3 hikes this quarter... maybe like 35-40% for DDR2 lines per Trendforce, and DDR3 likely continuing to rise.

Do markets just not know about this bottleneck/company or am I missing something from my research?

翻译：
我想要捕捉 DDR2/DDR3 持续的瓶颈，于是发现了 ESMT（

2026-08-26 21:18

原文：
$NVDA: "We expect to grow revenue by approximately 70% in fiscal 2028. This is a supply constrained outlook"

That statement alone is probably more important than the entire earnings beat just now?

Insane upward revision beat from +43.9% expected to 70%.

"With cloud industry backlog now greater than 2 trillion, CapEx by the top five hyperscalers is expected to reach nearly $800 Billion in 2026 and 1.3 Trillion in 2027"

$1.3 Trillion in 2027. (MS was around $1.2T June, so this is upward projection revisions)

Yeah the high-beta parts of the supply chain growth is going to like this.

翻译：
$NVDA：“我们预计2028财年收入将增长约70%。这是一种供应受限的展望”

仅这句话本身可能比刚才整个盈利超预期更重要？

从预期的+43.9%疯狂上调至70%。

“随着云行业积压订单已超过2万亿美元，前五大超大规模云服务商的资本支出预计在2026年接近8000亿美元，2027年将达1.3万亿美元”

2027年为1.3万亿美元。（微软6月约为1.2万亿美元，这是一项上调的预测）

是的，供应链中高贝塔部分的增长将会是这样的。

2026-08-26 21:15

原文：
$NVDA: "We expect to grow revenue by approximately 70% in fiscal 2028. This is a supply constrained outlook"

wtf? Isn't that ~$110-130B of above prior consensus.

That statement alone is probably more important than the entire earnings beat just now?

"With cloud industry backlog now greater than 2 trillion, CapEx by the top five hyperscalers is expected to reach nearly $800 Billion in 2026 and 1.3 Trillion in 2027"

$1.3 Trillion in 2027. (MS was around $1.2T June, so this is upward projection revisions)

Yeah the high-beta parts of the supply chain growth is going to like this.

翻译：
$NVDA：“我们预计在2028财年收入将增长约70%。这是一种供应受限的前景。”

这什么情况？不是比之前共识高出约110-1300亿美元吗。

仅这句话可能比刚才整个盈利超预期更重要？

“随着云行业积压订单已超过2万亿美元，前五大超大规模云服务商的资本支出预计在2026年接近8000亿美元，2027年达到1.3万亿美元。”

2027年1.3万亿美元。（微软6月约为1.2万亿美元，所以这是上调的预测）

是的，供应链中高贝塔部分的增长会是这样。

2026-08-26 20:59

原文：
$NVDA earnings TLDR:

Revenue: $96.22B vs. ~$92.17B (beat)
Adj. EPS: $2.22 vs. $2.09-$2.10 (beat)
DC: $89.0B vs. ~$86.3B (beat)
Adj. gross margin: 75% vs. 75%

Hyperscaler Revenue: $48.71B from $43.05B last quarter (custom ASIC growth hasn't really prevented this revenue from accelerating)

For guidance:
Revenue: $108B vs. ~$104.2B
Adj. gross margin: 74% vs. 75% (kinda the only soft spot)

Revenue ramp has been genuinely absurd over last 4 quarters.
-> $68.1B
-> $81.6B
-> $96.2B (we are here)
-> next quarter guidance: $108B, despite assuming zero China DC compute revenue.

Fun thing to note is Nvidia says its commitments jumped from $119B last quarter to $279B, primarily related to procurement of memory (for next few years).. So memory goes brrr.

TLDR: AI keeps on going brrr. Nvidia is clearly leading the charge and no obvious signs of demand slowing.

Given Nvidia is already a $5T+ company, I think most of the alpha comes from how Nvidia's architecture/capacity decisions impact elsewhere in the supply chain (eg. CPO, memory, 800V) nowadays.

Rather than simply finding mispricing in Nvidia itself.

Regardless, all the fun stuff happens in the earnings call in a few min.

翻译：
$NVDA 财报要点：

收入：962.2亿美元 vs. 约921.7亿美元（超预期）
调整后每股收益：2.22美元 vs. 2.09‑2.10美元（超预期）
数据中心收入：890亿美元 vs. 约863亿美元（超预期）
调整后毛利率：75% vs. 75%

超大规模云服务商收入：487.1亿美元，上一季度为430.5亿美元（自研 ASIC 增长并未阻止收入加速）

指引：
收入：1080亿美元 vs. 约1042亿美元
调整后毛利率：74% vs. 75%（唯一的软点）

过去四个季度收入增长异常惊人。
-> 681亿美元
-> 816亿美元
-> 962亿美元（当前）
-> 下季度指引：1080亿美元，且假设中国数据中心计算收入为零。

有趣的是，英伟达称其承诺金额从上季度的1190亿美元跃升至2790亿美元，主要用于未来几年内的内存采购……所以内存在“嗡嗡”增长。

简要概述：人工智能持续“嗡嗡”增长。英伟达显然领头，暂无需求放缓迹象。

鉴于英伟达已是市值超过5万亿美元的公司，我认为大部分超额收益来自英伟达的架构/产能决策对供应链其他环节（如 CPO、内存、800V）的影响，而不是单纯在英伟达本身寻找错价。

无论如何，所有有趣的内容都将在几分钟后的财报电话会议中揭晓。

🔄

Loading...