中国生产的芯片大多数都是mature node的产品。用tsmc的marketing话术来说的话就是10nm以下的芯片中国都有能力大量生产。不过一些特别的芯片如inari主要客户的产品，中国可没有这方面的技术。这是我们可以留意的。
顺带一提，因为中美贸易战的关系，原本跟中国下单的公司都转移来马了。这个是我们的moat，没有地渊政治问题的moat。这也是为什么我会说可以留意大马的封装测试了。

Yup, 你说的没错。
我比较喜欢把公司的发展跟技术摆在第一位。这样我晚上睡觉也不用太担心。

我比较喜欢用低价买入有潜质的公司，而且我买的timing超烂的。所以只要价格合适，公司未来也不错的情况下，我都会入手。

但是失去的是机会，因为投进去的钱会卡死在里面。

Earning Hightlight:

Mass Capacity demand improved sequentially, led
by stronger nearline Cloud more than offsetting
seasonally lower VIA demand

Nearline revenue increased sequentially for the
second consecutive quarter reflecting higher sales
into both US and China cloud customers and
stabilized enterprise demand

VIA market inline with expectations; anticipating VIA
to reflect more typical seasonal patterns through
CY24 with the March quarter representing the low
point

Over the next several years, the volume of AI-generated content is expected to increase and also shift toward more imagery and videos, which can be up to 1,000 times larger than text. These trends bode well for HDD demand over the long term, as HDDs remain the most cost-effective means to house and subsequently use mass capacity data.

for HDD, we see it as all the data that's going to feed that process is going to sit on HDDs. And obviously, once those models get trained, they're going to turn out data that's 85% plus of that is going to be stored on HDD.

DUFU 周年股东大会

✔️ 管理层表示今年的盈利保持在低二位数增长率

✔️工厂的utilisation rate 在2023年平均在60%，而现今工厂的utilisation rate 是70%

✔️管理层表示他们预计接下来的几年内，HDD的盈利增长为10%；NON HDD的盈利增长是20%

✔️在总盈利中，HDD还是占据了高达70%营业额，而其余的30%是来自于其他业务。

✔️Bukit Minyak 的新工厂已经在去年投入运营，专注于Sheet Metal Fabrication 的业务。管理层表示他们预计今年的业务会增长在20%，而接下来五年会维持在10-20%的增长率。

✔️总结来说，管理层会继续开拓业务市场以增加市场份额。与其同时也会考察新的领域，寻求发展机会。

WY, Stockaholics
Growth • Through • Sharing

我从fb偷过来的，credit to Stockaholics。

@Adama, you mean Sam Engineering is it haha.
Very good company, but the current price is not something that I am comfortable with.
Just a heads up, I think most likely they will need to do some corporate activities like issuing right issues or even private placement in the future to support their aggressive growth plan.

@HK Chong
Yes, definitely.
The management purposely release the QR after the AGM session. They also don't wish the price to plummet due to the poor result xD
Investing is all about the improving earnings in the future, not the previous nor current one.
Since the management already mentioned that they will grow at least 10% in HDD and 20% in their metal stamping division, I expect the upcoming QRs will achieve YoY and QoQ growth.

Haha good stuff never cheap. Dont u think they deserve some premium since they did not suffer a huge revenue/profit drop like many other bursa semicon related company. Althought their hdd/sdd equipment segment suffer too last year but their front end segment still continue growing fast.

Their price is already premium in my opinion. It's a great company managed by great managers. No doubt they will continue to grow in this year. But the price isn't that attractive to me at this moment. Just my two cents, cheers.

Yes 你说的对。
现在large capacity 的 hdd最多就只有10片disks。
我想要分享一下我的见解，如果有问题麻烦纠正。我先从downsides说起，
1。HAMR里面用的是titanium而不是smr里面用的aluminium，理应该来说可以fetch higher price。不过之前hdd的价格plunge了很久了，所以不能expect spacer的价格能现在立刻就起价。
2。在未来hdd的shipment会越来越少，而且主要顾客都是CSP，而且ship出去的product都几乎会是nearline hdd。所以spacer的需求不能expect会一直成长下去。
3。现阶段csp的主要capex都比较注重于AI server，意思就是囤gpu，而hdd是属于conventional data center里面的其中一个component，所以我也不expect hdd的demand在这个阶段因为“我对你 ‘AI AI AI' 不完”而有significant的提升。

当然downsides还有很多，不过我看到的upsides是
1。CSP在囤完了gpu，搞出了超强的大模型llm后，就会开始思考如何把大模型的产出store起来。逻辑上，这种类型的储存都是存在hdd里面。毕竟省钱是large capacity hdd最主要的卖点。
2。hamr是一个很重要的技术，如果一个plate做不到5tb甚至更高的话，在不久的将来cost per bit就会真的比ssd还要贵了。hamr里面需要很耐用的材料，而titanium spacer可以fetch更高的价钱。今年seagate才刚开始ship hamr hdd给customers，他们预计会ship 1个million个disks。
3。dufu本身也有做metal stamping这个业务，agm也说了每年要成长20％。个人觉得这个是令人鼓舞的，做不做得到就要继续观察了。。。