neway，up，up！

[沐沐精品日记 @mumu339960]

群P精神小妹淫乱现场，新疆妹子的颜值简直吊炸天

[Sawyer Merritt @SawyerMerritt]

Here’s my first full Cybercab ride. This vehicle is a game changer.

[Venu @Venu_7_]

The biggest stock market winners tend to share the same 7 characteristics before their biggest moves:

C - Current quarterly earnings up 25%+
A - Annual earnings growing 25%+
N - New product, new service, new highs
S - Supply & demand: tight float, big volume
L - Leader in a leading group
I - Institutions accumulating
M - Market in a confirmed uptrend

That's it. That's CANSLIM.

Most people obsess over the first 6 letters and ignore M.

But 3 out of 4 stocks follow the market. Get M wrong and the rest don't matter.

[李岳 @liyue_ai]

卧槽，现在开海外银行卡都这么简单了吗？

不到10分钟就开卡成功，而且还支持+86手机号，免开卡费。支持各种海外App订阅，ChatGPT、Claude Code、Grok、X Premium。

所有海外AI订阅一张卡全搞定，还能扫微信付款码把海外的钱在国内花掉。

还没有申请到海外银行卡的兄弟可以关注一下，这张神卡就是Starryblu。

为什么推这张卡：

1、背后是熊猫速汇母公司 WoTransfer，持新加坡 MAS 大型支付机构牌照，红杉资本投资，用户资金托管在 OCBC 银行，不是什么野鸡卡；
2、开卡只需要+86手机号，中国护照和身份证，0开卡费；
3、过审即得虚拟 Mastercard，可另申请实体卡寄到国内/海外；
4、支持美元、港币、欧元、英镑、新币、日元、离岸人民币等 10 种货币；
5、App 内可扫微信商户码/展示付款码，海外账户余额可直接在国内消费，开卡操作也非常简单，我自己走了一遍确实10分钟不到，前提是一定要填邀请码，可以加速审核。

开卡步骤：

1、应用商店搜 Starryblu（地区选中国），或用这个网页链接注册：https://t.co/0Hx8IGi1IF
2、邀请码填 MDDK027 ，能拿60新币消费券（不填只有10新币消费券），而且不填审核会很慢，填了邀请码会加速审核，我也是用了别人的邀请码，几分钟搞定。
3、证件选中国护照，地址证明上传身份证正反面即可；
4、过审后先申请虚拟 Mastercard；
5、入金优先选港卡，汇丰银行，ZA Bank等，秒到免手续费。

⚠️ 重点提醒：
一定要填邀请码！！！
一定要填邀请码！！！
一定要填邀请码！！！

自己裸申的兄弟排队有排到半年的，填完注册后还在排队审核的把你的 Starryblu ID 私信给我，我帮你问，加快审核。

兄弟们，有护照和身份证就能开，这条路不知道什么时候就又堵上了。

[Fang @FLMdongtianfudi]

冲冲冲！兄弟们！这波信息差一定要抓住！

最近才发现 Starryblu 这个全球支付 App，居然把不少海外支付场景一次性整合了。ChatGPT、Claude、Cursor 这类海外订阅能用，多币种也能直接管理，美元、欧元、英镑、日元、港币这些随时换。

更离谱的是，海外账户的钱回到日常消费场景也能直接用，支持微信/支付宝等支付方式。

而且现在通过我的链接注册，还有 60 新币消费券，消费直接抵扣。之前甚至看到有人在闲鱼倒腾白名单，这种信息差懂的都懂。

这种羊毛能薅就先薅，冲冲冲！
我的注册链接：https://t.co/87vRqS9MIW
邀请码：MDDK029

#Starryblu #GlobalPayment #MultiCurrencyAccount #PayWithStarryblu

[娜美知识库 @fhwofjow51260]

A 股玩家可以收藏了，这个开源项目有点猛。

一个本地面板直接搞定：全 A 选股 + 实时盯盘 + 策略回测 + AI 复盘。

内置 18 种策略，MA、MACD、RSI、KDJ 一键扫全市场；回测连 T+1、手续费、滑点、止损都算进去，盘后还能自动扒出连板梯队、概念轮动、炸板翘板。

更爽的是，日 K 不填 Key 也能用，AI 还能接 DeepSeek、通义、Ollama。

🔗：https://t.co/a48giCESEI

[neway，up，up！ @e8829163]

A股牛市宣言海报设计 https://t.co/76EEo1oK9Q

[迷口 @MixMico3]

生命滋养.

[Venu @Venu_7_]

One of the biggest mistakes new investors and traders make:

Buying a stock that gaps down after earnings.

It looks like a discount. It's usually a trap. To avoid falling into that trap, follow one simple rule:

The 3-Day Rule.

When a stock gaps down on earnings, wait at least 3 trading days before even thinking about buying.

Why?

Institutions holding millions of shares can't exit in one day. The gap down isn't the end of their selling - it's the beginning.

- Day 1: Bounce = dip buyers + short covering, not real demand.
- Days 2-3: You'll see whether the selling pressure actually dries up.
- After Day 3: The stock will tell you if it's stabilizing or still bleeding.

You're trading with evidence, not hope.

And remember:
Our focus should always be on earnings gap ups, not gap downs.

A gap up on heavy volume usually signals institutional buying like $PLTR - and their buying often takes days or even weeks to complete. That's the fuel behind many of the market's biggest winners.

Gap down = big money leaving.
Gap up = big money arriving.

Own what they're accumulating. Don't try to catch what they're distributing.

[Venu @Venu_7_]

The Rule of 40 - The Quality Test for Software Businesses

One formula: Revenue Growth % + FCF Margin %

Above 40% = high-quality software business.

Why it works: It rewards companies that can grow profitably, not just grow at any cost. The formula doesn't care how you get there - only that the combination of growth and profitability is strong.

Simple example:
• Revenue growth: 28%
• FCF margin: 15%
• Rule of 40 = 28 + 15 = 43% -> pass

Now look at $PLTR's Q2:
• Revenue growth: +93% YoY
• FCF margin: 62%
• Rule of 40 = 155%

That's not passing the test. That's lapping it - nearly 4x the bar.

Most software companies fight to stay above 40. The rare ones compound growth AND cash generation at the same time. That's what the formula is built to find.

[美研芒格君 @Kay2289123]

07/10 Meta出租算力，我已经陪大家解读了市场几条严重的逻辑错误，相信大家也赚到了。

先别着急骂meta，背后还有一条非共识的机会，搞懂它赚多赚少只是时间问题—— 算力供不应求，大家在加价抢，价高者得

根据我们的消息，SpaceX 出售给 Claude 的算力租赁价格远远高于市场平均价，这也是为什么 Meta坐不住了，每个月都有人来问小扎，能不能把算力租给他们

买越多的 GPU 赚的钱就越多，这是实话，但是还没有形成市场的共识，目前市场交易的逻辑还是：买越多的 GPU，现金流越差，越应该被惩罚。那些因为买了太多 GPU 被市场惩罚的公司，例如 $ORCL $AMZN $MSFT 等一众云服务厂，以及tpu的 $GOOG

那问题来了，这类公司什么时候才能洗掉冤屈，不再被惩罚呢？我认为还在等几个机会：
1. 第一个机会：GPT 5.6 以及后续更多的模型出来。

这个面一旦铺开，不管 token 的账单能不能打平，只要能向市场证明 Oracle 的长期订单不会被割掉，是非常稳定的。

2. Meta 出租算力实锤：不论是托管（用自己的算力托管先进模型），还是直接出租算力，市场都要见到：真的是买的越多，GPU 赚的钱越多

能让我打破这个逻辑链唯一的原因，就是消费端拒绝购买更多的算力 Token 来放到企业应用，在硅谷来看，这个市场非常火爆，超乎了大部分人的想象

最后分享给大家，我在硅谷看到的 SaaS 的现状。
The information说的没错。大部分的 SaaS 软件都会被强大的模型直接给吞掉了。

一切生产力最后还是会落在算力、网络和存储，把握机会！

[Sun Liao @sunxliao]

Why is $VPG my favorite robotics ticker?

Simple... EVERY humanoid robot needs precision force sensing. It needs to know how hard it's gripping a cup... how much weight it's carrying... how much torque its joints are producing... how its balance is shifting when it walks.

That is the fundamental layer that makes a humanoid actually functional in the REAL world.

Vishay Precision $VPG makes exactly that. Strain gauges. Load cells. Precision resistors. Force sensors. Motion measurement systems. The kind of components that get engineered into every humanoid platform being built right now.

Q1 2026 was the inflection quarter. Revenue hit $84.4M, up 17.6% year over year and above the $77M consensus. Bookings surged to $102.1M, the third highest quarterly level in company history. Book-to-bill came in at 1.21 with the Sensors segment hitting 1.36. Q2 guidance is $85M to $90M, well above the $79M Street expectation.

They're now actively engaged with FOUR different humanoid developers in various stages of design and qualification. Each of those relationships represents multi-year revenue potential once those platforms scale into commercial production.

The math on scaling matters. If humanoid production goes from thousands of units to millions of units over the next 5 years, and each humanoid uses dozens of precision sensors, the addressable market for a qualified supplier like $VPG expands by orders of magnitude.

Disclosure... we have $VPG positions.

We will buy/hold/sell/trade, as we see fit.

Not financial advice; do your own research.

Very few pure play humanoid supplier names exist in the public market right now...

$VPG is one of the cleanest IMO.

And BTW, gap filled. Bounce time?

[Serenity @aleabitoreddit]

Very interesting statement today: $MU CEO predicts a multi-decade memory demand cycle driven by humanoid robots.

"Humanoid robots, he says, will require roughly ten times more memory than today’s Level 2+ autonomous vehicles."

"And that demand wave is set to begin before the decade is out."

Something as well as was "Over time, we expect the value of on-device AI combined with pent-up unit replacement demand to drive memory demand growth"

Which is also another trend (Apple Intelligence is currently dog, but I'm sure we'll see innovations with localized/edge AI).

Feels like all the industry leaders from $TSM Chairman, $TSLA Elon Musk, to $MU CEO see humanoids as the next major trend so physical AI is probably next.

I wonder if the world is going to have enough memory. Or if we'll see enough breakthroughs to shrink memory usage.

[川沐｜Trumoo🐮 @xiaomustock]

下一代AI芯片用玻璃基板代替pcb，康宁是专业的，
下一代AI芯片pcb要换玻璃基板，康宁是专业的，
下一代电互连要换光互连，康宁是专业的，
康宁 $GLW 在玻璃这一块有无与伦比的话语权。

[川沐｜Trumoo🐮 @xiaomustock]

至于日月光半导体 $ASX
可以当成1000亿市值的小号台积电 $TSM ，
想象力无比夸张，也进入了供需紧张的提价周期。
现在手里就3个股票 #海力士 $GLW $ASX
别的没了，海力士正股占1/2，glw和asx各占1/4

[Paradis @ParadisLabs]

Humanoids will need a lot more memory:

Which will drive the next leg for memory suppliers like $MU / $SNDK / SK hynix.

From Micron's CEO earnings transcript:

"Humanoid robots carry 10 times the amount of memory as an average L2+ vehicle."

"We expect a sustained, substantial, multi-decade memory demand cycle to begin in the latter part of this decade."

Things can change very quickly on a couple fronts though:

1. Will Micron etc. have enough capacity to meet demand from the humanoid ramp up?
2. Will there be some kind of innovation in humanoids that drives less overall memory usage?

Right now, Micron are expanding capacity to supply volumes to their 16 long-term agreements. Which represent ~20% of DRAM volumes & ~33% of NAND volumes.

So there'll potentially be some overlap between those LTAs and the upcoming humanoid ramp.

Regardless, it all paints a picture that memory demand will remain high even if hyperscalers like $GOOGL / $AMZN etc slow down capex one day.

[Shay Boloor @StockSavvyShay]

$MU generated ~$365M in profit per day last quarter 🤯

[Michael | Hypermarkets @itsmichaelluu]

In 1 year, $AMKR spiked from $18 to $90 so 500% but it can easily hit $300-$500.

Here's 8 reasons why it hasn't ran like $AVGO $AMD $NVDA $MRVL $GOOG $AMZN:

1. Advanced packaging sits directly on the critical path of AI infrastructure growth and is becoming more important as Moore's Law slows.

2. Management expects revenue to grow from $6.7B in 2025 to approximately $11B+ by 2030.

3. EPS is projected to increase from $1.50 to roughly $5.00+, implying more than 3x earnings growth.

4. AI servers, GPUs, ASICs, HBM memory, and chiplets all require increasingly complex packaging solutions.

5. $AMKR has 18 active 2.5D engagements and multiple HDFO programs already entering production.

6. The Arizona facility creates a strategic U.S. packaging monopoly-like position among scaled OSAT providers.

7. Deep relationships with $NVDA, $AAPL, $AMD, $AVGO, $QCOM, $TSM and other semiconductor leaders reduce customer risk.

Higher-value advanced packaging should drive gross margins from 14% toward 22%+ by 2030.

♻️ RESHARE this post and write 1 comment, I'll DM you 2 more plays that hasn't exploded under $10 still.