方法论
研究框架、选股思路、思维方式 · 共 651 条 · 滚到底自动加载更多
-
说如果单位量超出预期，估值会更高。
@Liudas 我是在展示基础模型：如果它们的单位出货量打到上限（比如 400 万台而不是 350 万台），估值就会到 20。 而我确实预期下一轮架构“超级周期”会把很多预测都甩开，就像现在 $LITE 及其供应链正在发生的那样。
英文原文
@Liudas Was showing baseline modeling, if they hit upper bounds of unit projections (eg. ~4m units instead of 3.5), it hits 20. And I do expect the next architectural "supercycle" to blow away a lot of projections like what's happening with $LITE and their supply chain rn.
-
说自己在原文 thesis 里尽量简化给散户。
@Yolo365247isme @DynamicMoats 我会在原始 thesis 帖里尽量把事情讲得简单一点，这样散户投资者就不会看到太多这类细节。
英文原文
@Yolo365247isme @DynamicMoats I try to make things simpler for retail investors in the original thesis post so I don't include a lot of stuff like this.
-
说法案不会被通过得太理想，散户得看清现实。
大概不会，JP Morgan、$SOFI 和大银行游说太强了。 我只是把这项法案到底会做什么，讲给散户听，免得他们吃亏。 里面确实有一些好内容，比如用美国国债做 1:1 储备之类的（看看 Tether）。 但随后又塞进了银行想要的收益条款……然后政客就对散户继续疯狂洗脑，试图让他们相信这对行业有利。 可实际上，一旦这法案通过，你大概率再也拿不到储蓄的真实利率，而且很多流动性也会被抽干。
英文原文
Probably not, can't beat JP Morgan, $SOFI, and big bank lobbying. I'm just giving the unfortunate truth to retail investors about what the bill actually does. There's some good parts about 1:1 reserve with US treasuries, and such (looking at you Tether). But then they throw in clauses with yields that banks want... Then politicians just straight gaslight retail, trying to convince them it's positive for the industry. But in reality, once this is passed, you'll likely never get true interest rates on your savings and a lot of liquidity will be wiped out.
-
直接把 Clarity Act 说成银行游说法案。
把话说清楚：Clarity Act 是一部银行游说法案。 JPM 和其他机构游说双方，想控制数字资产 / $CRCL 稳定币。 $COIN 大概率会把整个行业卖掉，因为他们拿到了有条件的银行牌照批准。 但实际上，这等于禁止收益竞争： -> 银行可以继续给 .3% 利率的活期账户。 而不是把 4.3% 的国债收益发给用户，然后自己拿走那 4% 的差价。 -> 他们掌握了入金 / 出金通道和稳定币的卡点。 -> 禁止任何非银行机构提供类似服务，比如通过持有稳定币给出 4.3% 收益的新创公司。 他们的理由？ 1. “安全”。这些机构本身就是部分准备金模式，真遇上银行挤兑，和 1:1 抵押代币比起来更扛不住。 2. “那你就去当银行吧”。但他们却在背后偷偷游说，阻止任何新的竞争性公司真正拿到银行牌照。 如果你想看这个政府底下的事情到底是怎么运作的： 就看看现在发生了什么。 任何真正帮到散户的东西，最后大概率都会被禁。
英文原文
Let's get this straight: The Clarity Act is a bank lobbyist bill. Where JPM and others lobbied both sides to get control over digital assets / $CRCL stablecoins. $COIN is probably going to sell out the industry because they got their conditional banking charter approved. But effectively, this bans competition in yields: -> So banks can continue offering .3% interest checking accs. Instead of handing out 4.3% treasury yields and keep the 4% difference. -> Hands them chokepoints over on/off-ramps and stablecoins. -> Bans any non-banks from offerings, like startups giving out 4.3% yields by holding stablecoins. Their claim? 1. “Safety”. These are the same institutions that operate on fractional reserve models and would GG on a bank run vs. 1-1 collateralized tokens. 2. “Just become a bank”. While they secretly lobby behind the scenes to prevent any new competitive firm from becoming a bank. If you want to see how it's things are doing under this administration: Just look at how things are going. Any genuinely helpful to retail are just going to get banned.
-
认为市场比大家想的更有韧性。
市场比人们想象得更有韧性。 当然，局势升级是令人担忧的，但现在做空的人太多了（这反而会火上浇油），而且指数年内还是下跌的。 OpenAI 已经筹了 1220 亿美元以上的 capex 支出，所以我还是会保持多头（最近主要的传染恐慌，其实和他们无法履行合同义务有关）。
英文原文
Markets are more resilient than people think. Obviously escalation is concerning, but there's too many people short right now (which adds fuel to the fire) and indexes are still down YTD. OpenAI raised $122B+ for capex spend, so I'd remain long (main contagion fears recently were related to them not being able to fulfill contract obligations)
-
说要解除封锁，就得封锁封锁。
@zephyr_z9 要解除封锁，你就必须先封锁封锁本身。
英文原文
@zephyr_z9 To unblock the blockade, you must blockade the blockade.
-
说未来会有很多更便宜甚至开源的会计/创业工具。
@b00gersburner @pocimont 我也在用 QuickBooks，不过未来会出现很多创业公司竞争者（而且更便宜，甚至免费 / 开源），大家可以用 Claude vibe coding 自己搭出来。 所以这可能会来自多个方向的颠覆。
英文原文
@b00gersburner @pocimont I use quickbooks as well, but there's going to be a lot of upcoming startup competitors (and cheaper or even free/open source) that people can spin up vibe coding with Claude. So it's possibly disruption from multiple fronts.
-
认为 Intuit 会受 Claude 轻度冲击，而部分软件会受益于 AI 降本。
这个问题比较微妙。我确实认为像 $INTU 这样的公司会因为 Claude 以及其他工具帮你做税务和报税而受到一点冲击，而且这并不是一个很好的逆向抄底机会。 但反过来，像 $RDDT / $SNAP 这种软件名，可能会因为 AI 带来的费用下降而变得更赚钱（前提是 Evan 的老婆别再像 piggy back 一样一直用 Snapchat）。 颠覆是真实存在的，但也会有少数赢家。
英文原文
It's nuanced, I do think companies like $INTU are going to get slightly disrupted by Claude and others doing taxes + filing for you. And it's not a great contrarian buy-the-dip opportunity. On the flip side certain software names like $RDDT / $SNAP are likely going to end up more profitable with AI due to opex cuts (only if Evan's wife stopped using Snapchat like a piggy back). Disruption is real, but there's a few winners.
-
提到美军封锁霍尔木兹海峡，并想到 VIRT。
美国海军要封锁任何试图进出霍尔木兹海峡的船只。 我不太确定这里的计划是什么？ 不管怎样，我完全忘了去买像 $VIRT 这种做市商的看涨期权。 因为在这种环境下，他们大概率会一边欢呼一边从更高波动率中获利……
英文原文
The US navy to blockade any ships trying to enter or leave the Strait of Hormuz. Not quite sure what the game plan is here? Regardless, I completely forgot to take call options on market makers like $VIRT. Since they are probably cheering and profiting off of increased volatility in these environments…
-
说机构在高位买入说明 thesis 本身是成立的。
如果 thesis 本身不扎实，Point72 和其他机构就不会在历史高位买入。 每个人都想找下一个 10 倍股，而当我终于用 $AXTI 或 $IQE 这样的名字把它指出来时，很遗憾，总有人会拿市值或粉丝数来试图削弱原始 thesis。
英文原文
Point 72 and other institutions would not be buying at ATHs if the thesis were not fundamentally sound. Everyone wants to find the next 10x, and when I finally point it out with names like $AXTI or $IQE, it's unfortunate so many people try and undermine the original thesis by citing MC/follower count.
-
说新闻多半是表面动作，真正的信号要看别处。
你可能到处都能看到这种新闻： “美国和伊朗在巴基斯坦历史性和平会谈后未能达成协议” 公开谈判破裂本来就是预期内的，这样双方都能在表面上显得强硬。 任何真正的协议大概率都在背后渠道里完成（比如制裁放松）。 要跟踪这件事，像航班跟踪器（例如 Oman、Qatar）、Ghost Fleet 执法，或者保险承保人： 大概才是更好的信号，而不是主流媒体头条。 周日还是别过度反应，享受一下比较好。
英文原文
You might be seeing news everywhere: "US and Iran fail to reach agreement after historic peace talks in Pakistan" Deal falling through publicly is expected, so each side looks tough. Any actual deals are probably done through backchannels (eg. sanctions relief). To track this: stuff like flight trackers (eg. Oaman, Qatar), Ghost Fleet enforcement, or insurance underwriters: Are probably your best signals, rather than MSM headlines. Probably better to just enjoy your Sunday rather than overreact to headlines.
-
强调融资结构差异。
@nickthehodler 融资结构才是关键。 $NBIS 以更大的市值，稀释 20 亿美元去买 $NVDA 相关投资，或者用比当前行权价高 20%+ 的可转债。 这和 $IREN 把自己一半市值印出来，再卖到公开市场给散户，是结构上完全不同的。
英文原文
@nickthehodler Financing structure is what matters. $NBIS diluting $2B for an $NVDA investment off a much larger MC. Or convertible notes 20%+ above current strikes. Is structurally different than $IREN printing half their MC and selling that into the open market on retail shareholders.
-
说 6B ATM 的运作逻辑已经证明了一切。
@nickthehodler 这背后就是一个 60 亿美元 ATM 被打印出来并卖到公开市场的运作逻辑。 $NBIS 61%+ 的表现和 $IREN 年内 -8% 的差距已经证明了一切。
英文原文
@nickthehodler It's backed by the logic of how a $6,000,000,000 ATM that gets printed and sold into the open market works. The performance disparity between 61%+ with $NBIS and -8% YTD with $IREN is proof of it.
-
说没有融资就不可能变现 4GW+ 容量。
对，我真的不明白，他们在没有融资的情况下，根本不可能把 4GW+ 产能变现。 而他们却选择了 ATM，而不是像 $NBIS 那样用可转债 + $NVDA 投资。 大多数人都不理解稀释的结构机制，所以才会盲目地买进一个 60 亿美元、往公开市场狂印钱的机器。 如果他们当初听我的，大家的组合本来能从 $NBIS 涨 60%+，而不是被 $IREN 拖到 -8%+。
英文原文
Yeah, I genuinely don't understand, there's 0 chance they can monetize 4GW+ capacity without financing. And they chose ATMs over convertible notes + $NVDA investment like $NBIS. Majority of people don't understand structural mechanics of dilution, which is why they keep blindly buying into a $6B money printer sold into the open market. People's portfolios would have been up 60%+ from $NBIS instead of down -8%+ from $IREN if they listened.
-
建议先做税务收割，再长期持有真正能升值的资产。
@NieMehrKrieg 有时候，最好的做法就是先做税务收割，然后长期持有一个真的可能升值的资产。 现在的持有者，不过就是 $IREN 60 亿美元 ATM 的退出流动性而已。那个社区还会尽力鼓励别人跟着一起抱着不放。
英文原文
@NieMehrKrieg Sometimes it's best to just tax harvest and long an asset that might actually appreciate in value. Current holders are just exit liquidity for $IREN $6,000,000,000 ATM. And the community tries their best to encourage others to baghold with them.
-
说直接供货 hyperscaler 的公司通常不会低于 10 亿市值。
@fabiorbb 你自己去判断吧。我们就说，直接供货 hyperscaler 的公司，通常不会在前沿供应链里还低于 10 亿美元市值。
英文原文
@fabiorbb Up to you to figure out. Let’s just say companies that supply directly to hyperscalers usually don’t trade below $1B MC in frontier supply chains.
-
说信息发现会吸引注意力，不是发帖的人。
我有这么强吗？ 我觉得大概只是信息综合与发现而已。要是我发一篇关于一家纸巾公司 8 倍市盈率的帖子，没人会在意。 但当 $SIVE 是 $JBL 和 $MRVL 的光源，市值只有大约 4 亿美元…… 而且只交易像 $LWLG 这种更投机名字估值的四分之一…… 而其他像 $MTSI 或 $LITE 之类做 CW DFB 的公司市值都在几十亿美元时…… 最后吸引大家注意力的，是信息本身，而不是发信息的人。
英文原文
am i that powerful? pretty sure it's just information synthesis + discovery. If I posted about a napkin company trading at 8 p/e, nobody would care. When $SIVE is the light source for $JBL and $MRVL at ~400m... and trades at like 1/4th the valuation of something more speculative like $LWLG. while all the other CW DFB companies like $MTSI or $LITE are all in the tens of billions. Information ends up drawing people's attention, not the person.
-
批评只看技术图形的分析，强调基本面、催化和现金流建模。
那些收 500 美元一个月、只会画线的人，大概是我最烦的存在了……尤其是当他们还会在我的 thesis 帖子下面评论，比如 $AEHR 财报那种。 股价是由基本面和催化推动的。技术分析只是自我实现的占星术；如果 $AMZN 有一笔巨大的采购订单，没人会在乎技术面说的“超买”，真正重要的是把这会带来多少自由现金流，相对于市值要怎么建模。 $IREN 也是一样，居然还有人拿技术图去画它，而市场里明明有 60 亿美元的稀释，图表根本没把这部分算进去。
英文原文
Swiggly line drawers that charge $500 a month are possibly the greatest annoyance to me... Especially when they comment underneath my thesis posts like $AEHR earnings. Stocks move based on fundamentals and catalysts. TA is astrology self-fulfilling prophecies but if $AMZN has a massive purchase order, nobody cares about "overbought" based on TA it's about modeling in how much FCF that brings relative to MC. Same with anyone who tries drawing a TA for $IREN when there's literally $6,000,000,000 in dilution that a chart doesn't account for.
-
解释自己不设付费墙，因为信息发现和整合应该是正和的。
很高兴听到这个！我不设付费墙，因为我认为股票和信息的整合 / 发现是正和的。 我希望能打破那种只让机构受益的传统模式： $5000 的分析师报告 -> 机构通过暗池 / 冰山单慢慢吸筹 -> 散户 500% 之后才知道。
英文原文
Glad to hear it! I don’t add paywalls since I believe stocks and information synthesis/discovery is positive sum. Hoping to disrupt that traditional model benefiting institutions only from: $5000 analyst reports -> institutions accumulate through dark pools/iceberg -> retail finds out 500% later.
-
列举一组光子学相关标的，并说明自己更偏好集中持有一只 Win Semi。
先列几个： 2455 TWSE：Visual Photonics 268A TSEJ：Rigaku 3017 TWSE：Asial Vital 3363 TEPX：Foci 3450 TWSE：Elite Advanced Laser 5289 TPEX：Innodisk 6451 TWSE：Shunsin 6777 TSEJ：Santek 我通常会更专注于像 Win Semi（3105 TPEX）这样我最有把握的一只，做更简单、更集中的配置。
英文原文
Just a few: 2455 TWSE: Visual photonics 268A TSEJ: Rigaku 3017 TWSE: Asial Vital 3363 TEPX: Foci 3450 TWSE: Elite Advanced Laser 5289 TPEX: Innodisk 6451 TWSE: Shunsin 6777 TSEJ: Santek I usually try and focus on one like Win Semi (3105 TPEX) that I have the most concentration in and do more simpler concentrated on those.
-
说明自己主要持有股票，且会把核心思路公开，但不会把全部内容定价太高。
@qdgk123 1. 不，我只做股票。除非是像 $MSFT 这种低隐含波动率标的。 2. 主要是我对很多标的的思考过程。我把价格设到了最低的 1 美元，但核心想法其实还是都会发在主时间线上。
英文原文
@qdgk123 1. No, shares only. Unless it's low IV like $MSFT. 2. Just my thought process behind a lot of names. I set it to the lowest amount at $1. But post all my core ideas to main timeline anyway.
-
认为把钱投向供应链中的瓶颈环节，是市场还没完全意识到的机会。
@CoukosStephen 这本质上就是投资供应链，并去寻找市场可能还没注意到的瓶颈环节。
英文原文
@CoukosStephen It’s just investing in supply chains and finding chokepoints in them that markets might not realize yet
-
认为 AEHR 的快速上涨源于对资格认证周期和量产爬坡的误读。
哇，$AEHR 一周内就几乎翻了三位数回报。 这简直是在“用 Serenity 101 逃离永久下层阶级”这门课里高速刷进度。 先不说这个，理解资格认证周期对于看股票或看财报真的很有帮助。 很多散户在财报后直接做空，只看当前收入，而没有看到量产爬坡 + 所有超大规模客户都在测试订单这件事。
英文原文
Woah, $AEHR almost triple digit return in 1 week. This is speed running “Escape the Permanent Underclass with Serenity 101”. That aside, it helps to understand qualification cycles when you look at stocks or ER. A lot of retail investors went short after ER by looking at current revenue rather than volume ramp + all ongoing hyperscalers testing out orders.
-
解释自己的 thesis 灵感和研究不会付费墙。
@hekoya1 主要就是我在主时间线上写 thesis 帖子时的思路过程。 或者是我对一些最后并没有建仓的股票做的随机研究。我不会把自己任何核心想法加付费墙。
英文原文
@hekoya1 Just thought processes behind how I make my thesis posts on the main timeline. Or random research on stocks that I don’t end up taking positions on. I don’t paywall any core ideas I have.
-
感谢 9200 付费订阅，并表示大家喜欢自己的研究思路。
哇，已经有 9200 位付费订阅者了…… 我真的非常感激，谢谢大家！ 很高兴很多人觉得我从 $JBL 到 $ALRIB 这些随机标的上的思路对他们有帮助，只要 1 美元就能看到。
英文原文
Woah 9200 paid subscribers… I’m super humbled, thank you everyone! Im glad a lot of people found my thought process on random names from $JBL to $ALRIB helpful for $1! https://t.co/LXGVgFPDJt
-
看好 AEHR，并指出很多散户误解资格认证周期和量产爬坡。
是的，我后来把仓位更集中在 $AEHR 上了，因为它拿到了像 $LITE、$AVGO 或 $COHR 这样的新硅光一线客户。 很多散户误解了资格认证周期 -> 量产爬坡，所以才去做空。 看到散户试图做空 $AAOI 和 $AEHR，最后却在超级周期里遭遇无限亏损，真的很夸张。
英文原文
yeah i decided have a higher concentration on $AEHR after the new silicon photonics t1 customer like $LITE, $AVGO, or $COHR. a ton of of retail investors misunderstood qualification cycles -> volume ramp so they went short. its crazy to see retail try and short $AAOI and $AEHR just to face infinite losses in a supercycle.
-
认为 AEHR 上涨是市场误读资格认证周期和只看当前收入所致。
$AEHR 这表现也太猛了。 今天又涨了 20%。 对所有空头来说，这就是你误解半导体名字里的资格认证周期会发生什么。 然后还只盯着当前收入数字去看。
英文原文
Not too shabby with $AEHR. Up +20% today. For all the bears, this is what happens when you misunderstand qualification cycles in semi names. Then try and look at current revenue numbers only. https://t.co/9hU4UwNyRE
-
认为很多有粉丝的人却不懂光子学和收发器需求，只会看历史营收和画线。
有很多粉丝的人居然会如此自信地看错光子学和光收发器需求，真的很离谱。 他们只会看过去的营收，再在图上乱画几条线……然后一周之内就把自己的粉丝带着一起亏掉 53%+，在 $AAOI 上无限止损。 这只说明很多人根本不知道自己在说什么。
英文原文
It's crazy that people with a lot of followers are so confidently wrong about photonics and optical transceiver demand. They just go off former revenue than draw squiggly lines on a chart...And then proceed to lose all their followers 53%+ in a week with infinite losses on $AAOI. Just goes to show so many people don't know what they're talking about.
-
总结最近的 thesis，并补充自己长期看好的相关标的。
我最近分享的 5 篇 thesis 帖子： 1. $HPS.A（17.7 亿美元）- 变压器 / 开关设备的 DC 瓶颈 2. $ARM（1520 亿美元）- AI CPU 放量 3. Win Semi（57 亿美元）- 面向 CW 激光及从 SpaceX 到人形机器人等供应链的晶圆代工 4. $SIVE（2.95 亿美元）- 2026 年下半年及 2027 年的 CW 激光放量 5. $TSEM（220 亿美元）- 光子学晶圆代工 除此之外，我还看多并正面提到过的名字，比如 $MRVL、$AAOI、$RDDT、$NBIS、$RPI、$AEHR、$LITE、$COHR、SK Hynix、$LASR、$SOI、$IQE 等，也都可能是不错的补充。
英文原文
Most recent 5 thesis posts I've shared: 1. $HPS.A ($1.77B) - Transformer/Switchgear DC bottleneck 2. $ARM ($152B)- AI CPU ramp 3. Win Semi ($5.7B) - Foundry for CW lasers and other supply chains from SpaceX to humanoids 4. $SIVE ($295M) - CW Laser ramp for H2 2026 and 2027. 5. $TSEM ($22B) - photonics foundry Apart from those, names I've positively mentioned like $MRVL, $AAOI, $RDDT, $NBIS, $RPI, $AEHR, $LITE, $COHR, SK Hynix, $LASR, $SOI, $IQE, and others might be decent additions.
-
强调自己买入时会公开 thesis 和入场点。
@JungleStocks 我是 X 上少数几个会在第一次买入股票时就公开 thesis 和入场点的人之一。
英文原文
@JungleStocks I’m one of the only people on X that posts a thesis and entry point when I buy a stock for the first time.