Welcome to Taiwan Era.
Hello everyone, I'm Chong Cheng Hua.
We saw the Philadelphia Semiconductor Index undergo a rather sharp correction from late
June to mid-July.
But don't worry, everyone says this isn't an AI bubble, just a cleanup of funds and
deleveraging.
Has the deleveraging ended?
What about the fundamentals?
Have there been any new changes in the fundamentals?
Do AI models from China pose a threat to USI developers?
Is the competition intensifying?
Today, I am very pleased to invite Junong Capital CEO, CEO Shen Jun, to analyze this
for us.
Hello,CEO,BrotherChong,Hello to all the viewers
Right,looking back,from late June to mid July,the SOX index corrected by about 13%,
while some funds flowed into the so-called magnificent seven,
with their average stock price rising by 4.5%.
Tesla and Alphabet released their earnings reports this week,
so let's carefully examine this so-called i-logic.
We are seeing a shift.
In the past,we focused on capital expenditure and scale.
The more you invested,the more the stock rose.
Will we be more stringent in examining these AI earnings now?
在此,AI有所造成的異常在我們的實現
Google的 CEO曾經說
AI曾經說1%的人持有99%的貨幣
這不是我們的貨幣
它決定了一個非常強烈的貨幣
也會引起大增加和貨幣
貨幣貨幣貨幣貨幣貨幣貨幣貨幣
我們看到了如此重新的貨幣
最近的貨幣貨幣貨幣
為什麼之前的強烈賞出了,甚至於是一種股市場,
while the M7 stocks that weren't rising began to catch up?
It's simply because the gap between the winners and losers was too large.
You can see it yourself. By mid-year, as of June 31st, the stocks had surged by nearly 80%.
Among all US stock indices, except for small cap stocks which might have risen by 20%.
Others like the NASDAQ were actually up less than 20%, but the stocks alone rose by 80%.
所以,市場感受到太多重
更多
有個例子,有個例子
有個例子,是有個例子
我們可能會有四年來的屁股
但他們有四年來的屁股
他們說有150,000
你能夠借到1.5萬
是那些稍稅
這情況有所在的
有關影響的影響
從我們的影片
從一位的主意
the key focus is the extreme distortion of resources which will not be resolved in the future
let me share a number with you
we all know that csps are expected to spend 700 billion dollars or 800 billion dollars this year
that's the book
let me mention that this year the semiconductor sector including tsmc
ase and memory combined will earn nearly 700 billion dollars
oh so that's it it's moved from here to there
and this is exactly what causes extreme volatility in stock prices
Right,so coming back to the question,is there an AI bubble?
We just mentioned whether AI would lead to a bubble.
I want to share that in the AI space,some are working hard for their money,like the
humanoid robots sector.
Or even those we consider winners,like OpenAI,they are working hard for their money.
They haven't made a profit yet,they might be losing $10 billion or even $20 billion
every quarter.
They are taking huge risks,and the humanoid robotics companies they pull along are also
taking huge risks,yet their profit margins remain relatively fixed.
嗯
conversely for those selling shovels wow their profits keep piling up
even the director of sk hinnick said excuse me i feel this isn't quite normal we're making too much money could it be
they even said the prices are a bit high and they should probably be a bit healthier
if you read the whole thing he sounds a bit guilty wondering how they could be earning so much
he didn't actually see a downturn coming yet
so from this perspective we really need to be mentally prepared for the investment market knowing that it will never be balanced
When extreme stock price volatility occurs, sometimes people get cautious, causing the
price to rebalance, but in the long run, it will still lead to polarization.
Yes, but when we look back at the memory industry, we see they are continuously expanding capacity.
We know that the demand for memory remains very high, but it will certainly take time
for the supply side to catch up. I just want to remind everyone, because many ghost stories
are surfacing right now, as memory manufacturers everywhere are expanding plants, whether it's
在香港,臺灣,或在中國,包括YMTC和CXMT,
都在增加了
人們可能會想想
如果您撥開工程
決定,它們將在增加了什麼我們的名義
是一種連續的連續
它們現在變成了一種連續
在最後一種連續的連續
至於2030,
there will still be a very significant shortage.
This can help strengthen confidence,
confirming that if the memory sector corrects,
it is perfectly fine in terms of both stock price and industry fundamentals.
Like recently, Sammy, who used to love to talk down the market and so on,
suddenly came out and said he believes memory will remain in shortage all the way through.
His perspective, or for example,
like even when you look at TSMC's earnings call mentioning envious gross margins and so on,
actually tells you one thing,
why some people will tell you decisively that memory demand will be excellent and will continue to be in short supply
you might think if the price gets too high i just won't buy
the conventional investment logic for business has always been if it goes up i won't use it or i'll just stop
so how could demand possibly be limitless
what i want to share with everyone that they are ignoring or why they see what others are ignoring
is that we now see tsmc's capex at 60 billion this year maybe 80 billion next year
or nvidia's gpus and amd's gpus doing great
我們當然忘記 that for every GPU or CPU produced
它 must be paired with a unit
a fixed unit of memory
它是不 sold separately
memory is not sold alone
所以我們的 AI future becomes clear
with so many GPUs coming out
如果我們不存在的話
我們不存在的話
我們不存在的話
所以我們看它從 TSMC's perspective
從 the semiconductor perspective
我們的 demand for memory 4 years from now
should be at least 4 times what it is today
嗯
在最少
Yes
And this premise is assuming my GPU volume hasn't even increased
If I produce 100 units of GPUs this year and 100 units 4 years from now
Although there is migration and they might be more powerful in 4 years
If it's still 100 units, the memory I need to pair with them will still be 4 times as much
Unless I downgrade, for instance, fine, I'm driving a Bentley, but I don't want air conditioning
I don't want ABS brakes, then your performance will definitely suffer
So later on you will keep hearing people talk about downgrading specs
你可能會認為 memory是不好的,
但不,因為有沒有的,
你會認為大家是在中國的增長和產品的增長。
如果看來,產品是在增長的增長並且在2028,
或者我們叫它,四年以後,
加上了它的增長。
嗯嗯。
我們就說,
在這四次,
所以,
在這一年,
四次。
如果我們的資料設計是在四年以後,
it is now, wouldn't TSMC be finished? Wouldn't Nvidia be finished? That's impossible.
We're all optimistic that GPUs will keep trending upward, so investors who are worried about memory,
if you are truly worried that memory will crash, then you should be like Michael Burry and not buy
any semiconductors at all. You can't say you're bullish on TSMC but bearish on memory. Let's get
this logic straight. You can't have it both ways. If it looks like memory is going to collapse in
未來,這些世界都會崩潰。
當然,我們不可能會否有機會,
我們不可能會否會發展開四年
和世界的世界返回到大陸,
讓 AI後來的確定。
我們不可能會否會發生。
但你必須理解釋這個 logic,
如果 AI還會升幅度的數量,
如果數量的數量會升幅,
就會增加了。
這就是這些人所說的,
不只是給你一個路線或是領導的
而我們剛才提到的
《病》
我們還沒有說了
我們還沒有說了
《軟體》
《軟體》
《軟體》
《軟體》
《軟體》
《軟體》
《軟體》
《軟體》
《軟體》
As we've mentioned before, TSMC estimates that future applications for AI will require 8 times the current computing power of these, brains.
Yes, the memory demand is simply unimaginable, it's terrifying.
To be honest, in our reports from visiting Samsung and Hinex, even they can't calculate.
Exactly how much is needed, they don't know how to negotiate with clients.
如果他們太高,他們會負敗的交易,
但在實際上,他們會負敗的貨幫助,
他們會負敗的貨幫助,
他們會負敗的貨幫助。
當然,他們會負敗的貨幫助,
他們會負敗的貨幫助,
但根據上一個討論的評論,
他們會負敗的貨幫助。
另一個重點的評論,
包括了一種同時的評論,
我們看到最近,
人們覺得這些關係的關係
是被高傳的高傳領的高傳領的
and their ratings are increasing so it's a situation where if gpus increase racks increase
and servers increase including from ai from communications passive components and pcbs
all demand is moving upward if we talk about computing power as we just mentioned cpus and
gpus are paired with memory they are a set in the future like the hvdc power supply for servers
it is paired with passive components passive components are used for power so if you believe
如果您相信,您的需要提供力量的力量,
您的支援力量是否有力量的力量,
您的支援力量是否有力量的力量。
如果您相信,您的支援力量是否有力量的力量。
如果您相信,您的支援力量是否有力量的力量。
您相信,您的支援力量是否有力量的力量。
您相信,您的支援力量是否有力量的力量。
所以,您的支援力量是否有力量的力量。
所以,您的支援力量是否有力量的力量。
您的支援力量是否有力量的力量。
您的支援力量是否有力量的力量。
所以,您的支援力量是否有力量的力量。
優優優優優優優優優優優優優優優優優優優優優優優優優優優優優優優優優優優優優優優優優優優優優優優優優優優優優優優優優優優優優優優優優優優優優優優優優優優優優優優優優優優優優優優優優優優優優優優優優優優優優優優優優優優優優優優�
but shortages are normal because after all this massive demand, the supply simply cannot keep up.
Now, as the first wave pulls back and finishes consolidating,
once the actual shortages begin, the stock price will follow the logic of supply shortages.
So, what should its valuation be?
Actually, in due time, the capital market will give it a reasonable valuation.
Okay, so looking at it from an economic perspective or from the fundamentals of the industry,
memory is currently in short supply.
The supply side is severely lacking, and this will likely continue for a long time.
但是,有一個不斷的經濟發生在這個業務中的
美國的美國人說是中國人的美國人
像 Samsung是傳續於美國人的美國人
而是太多貨
的供貨力是完全在美國人的手段
而且是中國人的美國人現在太高
所以政府是購買的貨帶
你認為這個人的這個?
我認為這個人認為他們認為貨帶的貨帶
但在實際上,我們的認為我們的課計
難以為他們的貨帶貨帶的貨帶
With US companies, they can ask for equity, like they did with Intel or others.
But for the US government to take equity in a Korean company, that's a bit difficult.
We haven't seen them take an equity stake in TSMC, right?
However, they could follow the TSMC model, meaning if you don't come and set up a plant here.
If you don't come here to build what they call a US-based memory facility,
because everyone knows there's a shortage,
it's only natural for them to want these companies to expand their capacity here.
honestly if these companies don't face such overwhelming demand pushing them
they really aren't willing to expand because for memory manufacturers they know one thing
i don't need to expand my plants when the time comes and you have a shortage i'll just double
the price and i'll make the same amount of money or even more and i can do that for a long time
of course they want a healthy market so their customers don't go under but for them over
肥料肥料肥料肥料肥料肥料肥料肥料肥料肥料肥料肥料肥料肥料肥料肥料肥料肥料肥料肥料肥料肥料肥料肥料肥料肥料肥料肥料肥料肥料肥料肥料肥料肥料肥料肥料肥料肥料肥料肥料肥料肥料肥料肥料肥料肥料肥料肥料肥料肥料肥料肥料肥料肥料肥料肥料肥料肥料肥料肥料肥料肥料肥料肥料肥料肥料肥料肥料肥料肥料肥料肥料肥料肥料�
KIMI-K3
This was actually one of the factors that led to a market correction last Friday.
People thought, okay, last month the White House decided that Anthropix Claude 3.5 was
too powerful, so it restricted access for non-Americans.
And then, shortly after, China released Quen 2.5, which scored nearly the same as 3.5,
but it's actually open source and launched immediately.
How big of a shock is this to the current USI model landscape, and has market competition
其實, everyone is just looking at the final evaluation results.
It has 7 metrics, and Quen won 6 of them.
And one look tells us that Chinese people are really good at taking tests.
What we are best at is test-taking, whether it's the SATs or our joint entrance exams,
we eventually just master the question bank.
Essentially, they looked at those 7 performance metrics and tuned their model to be perfectly
suited for those, whether for AI agents or image generation.
You can actually see how their models work
Currently, whether it's Anthropic or OpenAI, they are all very powerful
So powerful, in fact, that sometimes they become powerful enough to be stupid
For example, if you use Gemini in a coffee shop
It starts off strong with great proposals and estimates
But after a month, it gets slow because every time it makes a decision
It reconsiders every previous decision
you can imagine how that piles up
by the second month
it starts making errors
like ordering thousands of rubber gloves for a cafe
it starts ordering carrots
when there isn't even a kitchen
so in the end
it just crashes
the problem with AI now
is that while it might be powerful
even single task powerful
i have to tune it to handle these 7 specific metrics
so you have to train it specifically for that
看起來,不像 DeepSeq
DeepSeq is like a needle, a small model that had a huge impact when it came out.
What does that mean?
It means I don't need to be that powerful, and I don't need that much computing power.
But with Quen, I have massive computing power, yet I perform better than you, and the way
I perform better is by tuning it specifically to handle the performance you're looking for.
Coming back to the point, even though it performs well, people don't need to worry
太多了因為它還需要大量的空氣力
然後它將會很快的空氣力
有一個東西人都沒有認為
它的價值
它的價值是一樣的價值
它是0.95%和1.07%
但安thropic是超費的
$3
但看起來
是 OpenAI really doing worse?
它是非常好
的開放
為
因為人們說它是費的
are enterprises willing to pay?
this is what we need to go back and discuss regarding AI
it's not just about performing well
or generating pretty pictures or answers
if that were all
it would just be for us non-paying or low-paying users
而是我們的不費用
原因是因為它做得到的事情
它是做得到的
它是需要的需要
像是需要的
如果用它是用它
或是關於你的研究
你不能夠做得到的
你會看到,甚至是很貴的,而且它的效果不太好,為什麼它是很高的?
你可能說,這些人們不懂得用 Open Source,但我認為不覺得他們的獨自貴。
他們應該能夠實現實現實現實現實現實現實現實現實現實現實現實現實現實現實現實現實現實現實現實現實現實現實現實現實現實現實現實現實現實現實現實現實現實現實現實現實現實現實現實現實現實現實現實現實現實現實現實現實現實現實現實現實現實現實現實現實現實現實現實現實現實現實現實現實現實現實現實現實現實現實現實現實現實現實現實現實現實現實現實現實現實現實現實現實現實現實現實現實現實現實現實現實現實現實現實現實現實現實現實現實現實現實現實現實現實現實現實現實
是在那裡的事,
之後K was released,
它的機器 immediately hit the limit.
它 seems unable to allow new users
去上網上用了新的新版本來使用了新版本來的
因為它的背景已經不足了。
所以,它不是在這裡的底色
它可以用 AI models on lighter hardware.
它們的額外 demand是其實大量的。
所以,是這其實好事?
是這其實好事?
有更多的影響,
是更多的更多的更多
為什麼是更多的更多?
因為我們剛剛說
如果是個體驗的問題
而你的模特兒是有點困難
而你使用它在不足的地方
但實際上的使用者
不會讓我能夠享受到更多的效果
而我不會讓我能夠
稍為的更多的效果
如果我知道這些資料
在正確的地方
要有更多的執行
是當然好
那些人會有盡量的
所以你能看到
這次的壞壞壞在售貨物店裡的壞壞壞壞壞壞壞壞壞壞壞壞壞壞壞壞壞壞壞壞壞壞壞壞壞壞壞壞壞壞壞壞壞壞壞壞壞壞壞壞壞壞壞壞壞壞壞壞壞壞壞壞壞壞壞壞壞壞壞壞壞壞壞壞壞壞壞壞壞壞壞壞壞壞壞壞壞壞壞壞壞壞壞壞壞壞壞壞壞壞壞壞壞壞壞壞壞壞壞壞壞壞�
the real concern might be when we go back to examine OpenAI
it becomes a situation where, sure,
CLOD's benchmark scores might not be significantly higher than GPT-4's
but they are able to make much more money from enterprise clients
Exactly
As we mentioned, OpenAI's costs and efficiency are similar to CLOD's
so CLOD 3 might not hurt Anthropic,
but it could definitely hit OpenAI
有甚至今的說話,如果有一個AI瀏規, OpenAI會是說的一種類似的一種類似的。
所以,我們看來的責任 OpenAI在中國的責任?
如果它不影響 Anthropik,它會直接射擊 OpenAI,
尤其是跟 Oracle和 Cloudflare的背景。
你覺得這個動作?
其實,我們不相信AI會流動,
但最弱的原因是在《AI》和《Oracle》中的主要是《OpenAI》。
你能看到了解的目標
像它的借貨倫、購乎率、購乎率、
或是在購買購買貨。
這些是否有任何的影響。
只要是,《OpenAI》需要找到一種商品的商品作為
所以,看著他們在做的
他們要進行進行的
他們想要做一個話題
他們要做一個話題
他們要做一些其他東西
他們要做到販賣物的物品
我們會把這個做法
很清楚
他們真的不能夠保持
unless美國給他們
他們的公司不可能會有問題
所以我們會看 closely
因為他們的商品
如果這個認識的認識是沒有發展,
在這個中處的認識是沒有發展,
在這個中處的認識是沒有發展,
在這個中處的認識是沒有發展,
在這個中處的認識是沒有發展,
在這個中處的認識是沒有發展,
在這個中處的認識是沒有發展,
在這個中處的認識是沒有發展,
在這個中處的認識是沒有發展,
在這個中處的認識是沒有發展,
在這個中處的認識是沒有發展,
如果他們開始發展,
如果他們發展,
他們發展,他們發展,
他們發展,他們發展,
如果能否購買貨貨貨貨貨貨貨貨貨貨貨貨貨貨貨貨貨貨貨貨貨貨貨貨貨貨貨貨貨貨貨貨貨貨貨貨貨貨貨貨貨貨貨貨貨貨貨貨貨貨貨貨貨貨貨貨貨貨貨貨貨貨貨貨貨貨貨貨貨貨貨貨貨貨貨貨貨貨貨貨貨貨貨貨貨貨貨貨貨貨貨貨貨貨貨貨貨貨貨貨貨貨貨貨貨貨貨貨
在美國的美國,我們覺得我們的貨幣是否會有效果,
我們的其他人,不一定會有效果。
其實,我們覺得我們的貨幣未來的未來,
現在,當然,我們的貨幣未來的貨幣未來,
那是確實。
現在,如果你貨幣,你貨幣。
而且,你更多貨幣,像更多貨幣,
像TSMC,他們不能算貨幣,
Eventually, they will make money.
Yes, it's hard-earned money, and they are truly carrying human hope forward, like OpenAI and Anthropic.
They have to figure out how to make money themselves.
Why do I say the real winners haven't appeared?
Because it's about the application side.
Because once this AI, brain, is smart enough, robots will emerge.
AI pets will emerge, AI autonomous vehicles will emerge.
you might be able to use what others have already spent so much money on
developed and made ready
and i just need to pay a reasonable price
i can integrate it into my system
it's a bit like apple now
apple chose gemini to integrate
if it's smart enough
it might just make their iphone a huge hit
these things could happen on the application side in the future
in the application sector
those who can seize the initiative and hit that sweet spot
will likely be the first wave winners
但便是企業的,在一旦會發生,
在一旦會發生的一流程。
這就是為了我們已經看著當中的AI模式
是完全適合的。
像是在手機上,
它們已經在8-10年來
讓它們發生了一些
手機,
手機,
手機,
最後手機。
現在,
它們都已經在三年。
一切都要做到,每公司都要推出了,就像每公司都要推出了,像每公司都要推出了,像 Inventech的手法上,在那天
BenQ smartphones and the rest all died off
Why?
People won't buy from everyone
It will be the same in the future
There might be a hundred brands of robots
That's when a real bubble forms
You'll need all sorts of parts
And there will be massive shortages
But in the end
There will be a pile of robots that no one buys
That's exactly where the real bubble lies
So, before those applications truly emerge
That's why many say we are in the early stages of AI
Like the top of the third inning
那是他們所說的,OK?
很多 pessimists might think
you're lying to me again, trying to trick me every day, saying it's only the third inning when stock prices have already risen like this.
They think that because they see the massive scale of the applications coming later.
Yes, exactly.
So we see that at this point in time, it is indeed still the top of the third inning.
For those of us selling shovels,there's no need to worry.
Once the leverage is removed and funds are rebalanced,buying the dip is a very good entry point.
Thank you very much to the CEO for joining us for this interview,and thank you to our viewers for watching.
Taiwan's Grand Era
請訂閱 訂閱 轉發 打賞支援明鏡與點點欄目