Buy Them Coffee

[Buy Them Coffee @BuyThemCoffee]

$MU 空頭一直在壓，每個週五都會下跌，每週五苦苦堅持都把周K線畫得很難看。

但存儲價格和公司利潤一直往上爬，這些空倉總有一天需要面對現實。

不要再低估存儲龍頭未來兩年的現金流
如果把一半自由現金流去回購股份
美光將會是$3000以上的股票

[Buy Them Coffee @BuyThemCoffee]

@carl88187210579 你试想想一个外国人用支付宝打算在贩卖机前打开雪柜买一枝冰水，扫完码之后说你没有芝麻信用，然后柜门不让你打开

[Ian Andrews @IanAndrewsDC]

Probably worth buying the dip in chip stocks today. Panic over China DUV machines seems overblown:

[Buy Them Coffee @BuyThemCoffee]

很多股都有向下突破17/7之低位
$MU 已拒絕再下跌
7月17日的盤中低位 $804 明顯出現超跌
現時在830-870之間找Higher Low支撐

[Buy Them Coffee @BuyThemCoffee]

笑話一則： 美光 $MU
在6月24日公佈其每股季度盈利為USD 25.11
其後股價於6月25日USD1,255見頂 ！

至7月17日, 股價下跌36%+ 至USD 800
用了15個交易日跌了約USD450,
相等於跌了4.5年 (18個季度)的盈利。

[Buy Them Coffee @BuyThemCoffee]

Many stocks are trading below the 17July low, but Someone is buying $MU.
It is now at 877 which is still 8-9% above the intra-low(804) since 17July.

[Buy Them Coffee @BuyThemCoffee]

Micron $MU announced its quarterly EPS of USD 25.11 on June 24. It reached ATH at USD 1,255.
On 17July, it fell to USD 800. It took 15 trading days to drop about USD 450(-36%), equivalent to 4.5 years (18 quarters) of earnings.

[Buy Them Coffee @BuyThemCoffee]

@Alpha_Cat 可惜現在872了

[Buy Them Coffee @BuyThemCoffee]

@LuBtc888 拿了外国护照 改了另一个名 ，CRS根本连不上税务居民编号，只是新闻发出来给宠百姓看，感觉舒服一点。实际上你信托找海外居民代持，代持文件在律师手，银行又不知道是代持，那怎会连把资料连到你国内的信号。富人还是有富人的方法。你少担心。

[Buy Them Coffee @BuyThemCoffee]

@Stanleysobest 這種黑工 現在香港默許了嗎？不怕坐獄？

[Buy Them Coffee @BuyThemCoffee]

@alexdata99 那你擴產完還是跟不上需求的增加 HBM不是週期商品 是基建持續 物理AI推動下 需求只會更多

[Shay Boloor @StockSavvyShay]

Bank of America added $MU to its list of "best investment ideas" saying low cost Chinese AI models are expanding adoption and HBM demand.

Kimi K3 still requires ~1.4 TB of HBM per instance reinforcing Micron’s long term outlook.

[Athu @athuinvests]

UBS projects Micron $MU to generate over $400B in cumulative free cash flow between 2027 and 2029. 🤯

$1,625 price target (UBS analyst Timothy Arcuri)

It could use that cash to repurchase more than 40% of its shares.

[心灵书舫 | SoulSloop @SoulSloop99]

当你越来越优秀时，你就开始理解，其实每个人没有好坏没有对错，它只是处在不同的频率，做了不同的选择。
有个好的心态，走的路会更广阔。

世界上最珍贵的三个词：认真、坚持、感恩。认真的人改变了自己，坚持的人改变了命运，感恩的人留住了幸福。
如果生活有奖励，请你先努力，好的运气都深藏在自己的努力里，任何道理不如自己摔一跤。

我们还年轻，长长的人生可以受一点风浪，在最黑暗的那段人生是我自己把自己拉出深渊，没有那个人，我就做那个人。

最近比较触动我的一段话，深深鼓励了我，希望也可以鼓励到你们：我知道你最近很累，是那种看不见的，身体上和精神上的疲惫感，但请你一定要坚持下去，就算天不遂也 好，技不如人也罢，千万别让烦躁和焦虑毁了你本该有的热情和定力！

别贪心，我们不可能什么都有，也别灰心，我们不可能什么都没有。

[TheBigBerbowski @TheBigBerbowski]

These companies are down 40%.

What changed?

Nothing.

We got various confirmations that supply is short in this decade and will only slightly improve in the next one. SK Hynix will double capacity but it will not be enough.

@michaeljburry thinks this is the peak of the cycle. High margin, low forward p/e. If I'm not mistaken he shorted $NVDA and is short $ORCL, but wouldn't it make sense to short HBM/NAND producers if it's indeed a peak of the cycle?

In my book, nothing changed when it comes to $SKHY $MU $SNDK $KXIAY $SSNLF.
Does it mean they can't go lower? Ofc not.

Will it go higher? If the business shows strength over and over again then nothing will be able to keep the price down for long.

Long $SKHY $MU $SNDK $KXIAY

[Shay Boloor @StockSavvyShay]

Morgan Stanley projects $MU gross margins could reach 90% in 2027 supported by 16 supply agreements with floor pricing and HBM sold out through 2027.

That contracted structure combined with a mix shift toward higher-margin HBM and data-center memory should structurally lift company-wide profitability while advanced packaging remains a huge chokepoint.

Also important to call out that HBM4 for $NVDA Vera Rubin platform began shipping in March 2026 and is ramping at roughly twice the pace of HBM3E 12-high scaling Micron’s highest-margin product into its strongest demand platform.

[Athu @athuinvests]

Even if the street prices $MU based on historical PE ratios, $3000 seems inevitable by 2028. 🤯

I did the math based on EPS projections. It is absolutely easy and crazy at the same time!

Average PE ratio of $MU for the last 10 years has been ~20. We'll consider that as the base case.

2027: EPS (est.): $150 → Price: $3000 ($150 x 20 PE)
2028: EPS (est.): $180 → Price: $3600 ($180 x 20 PE)

Beyond 2029, EPS could go past $200. That opens the door to $4000+, especially if margins hold. Even in a slower scenario, it is hard to see the company reverting to old valuations.

Catalysts:
> Innovation driving demand across applications and tech developments.
> Capex ramp through 2030 with more SCAs signed.
> Custom memory solutions sustaining demand.
> Supply imbalance continuing or stabilizing at tighter levels.

Downside risks:
> Hyperscaler capex coming in lower than expected.
> Oversupply in memory products.
> Competitor capacity pressuring ASPs.

Also underrated: fabs are genuinely hard to build fast. New capacity takes years to break ground, ramp, and staff with scarce skilled labor. People overestimate how quickly supply comes online, which is why this cycle stays tighter for longer than expected.

On a personal note, memory has become as crucial as GPUs/CPUs for AI. Especially HBM, where AI essentially thinks. That is Micron's flagship product right now.

The lifespan of AI accelerators has compressed from 5-7 years to roughly 18-36 months. Since HBM is permanently integrated on the package, you replace it too. Heat from bigger models and workloads accelerates burnout, forcing ongoing refreshes.

Even if supply eases, manufacturers will keep producing HBM, DDR, LPDDR, NAND, etc. This creates a continuous revenue base with upside from new tech and custom solutions.

HBM and HBF stay tied directly to the GPU on the same package, so we stay bullish there (and on $SNDK/Kioxia even if DDR eases). LPDDR is now foundational system memory for Nvidia AI servers. New products and HBM should drive the majority of sustained upside through the AI buildout.

Unlimited memory demand from AI agents, millions of daily chats, and RAG keeps NAND strong. Software optimization might trim some NAND needs but will not hit HBM the same way. Bullish on memory controllers too. $SIMO is my pick there as the brains behind NAND.

Other HBM applications in robotics and AVs have barely started scaling. Micron's CEO highlighted this in the last earnings call. Plenty of unknown upside left.

Ultimately, if supply meets demand, I see two paths for ASPs:

1. Sustainable, stable demand with continued (slower but steadier) increases from tech developments, custom solutions, new products, and shorter compute lifecycles.

2. Prices lower but not to historical lows. More units produced balance revenue, with EPS growth from hotter new products.

Worse cases feel like a far-off 2030+ story.

[Buy Them Coffee @BuyThemCoffee]

@StockSavvyShay Is it really MS projection?

[Ehrmantraut Capital @EhrmantrautCap_]

When you see those estimates on $MU from Morgan Stanley, I don't see how anyone would panic sell.

I mean look at these estimates:
Q1 2027 operating income: $54.1 billion
Q2 2027 operating income: $60.2 billion
Q3 2027 operating income: $67.9 billion

Current market cap: $959 billion. They will earn a significant portion of their market cap in the coming few years, let alone the terminal value of their cumulative earnings in the period after.

AI demand is not going anywhere, and what happens to memory demand once physical AI hits momentum? When basic LLMs already require astronomic amounts of memory, just imagine how much memory robotics will require.

$MU $SNDK $SKHY $DRAM $EWY

[Buy Them Coffee @BuyThemCoffee]

@_FORAB 现在又说存储不缺了 盈利见顶了

[Buy Them Coffee @BuyThemCoffee]

@btcbears 现在又说存储不缺了 盈利见顶了

[Buy Them Coffee @BuyThemCoffee]

笑話一則： 美光 $MU
在6月24日公佈其每股季度盈利為USD 25.11
其後股價於6月25日USD1,255見頂 ！

至7月17日, 股價下跌36%+ 至USD 800
用了15個交易日跌了約USD450,
相等於跌了4.5年 (18個季度)的盈利。