萌小主
38 Posts
萌小主
Square Verified我是TokenMore🍣一个不擅长分析的BTC分析师
Occasional Trader
8.8 Years
67 Following
95 Followers
202 Liked
Posts
Trading-side data continues to grow, with the entire network DEX growing by 34%
Hyperliquid earns 5M in a day—definitely a printing machine
The lending side is also heating up, with average growth of 20%
On-chain leveraged debt is also expanding.
Personally, I think the data is preparing for even more aggressive growth
#Hyperliquid bitcoin:native https://t.co/QPiqL38p9n
Hyperliquid earns 5M in a day—definitely a printing machine
The lending side is also heating up, with average growth of 20%
On-chain leveraged debt is also expanding.
Personally, I think the data is preparing for even more aggressive growth
#Hyperliquid bitcoin:native https://t.co/QPiqL38p9n
Hong Kong Exchanges and Clearings is raking it in
Made 10.5 billion in the first half; 2Q profit in a single quarter was 5.38 billion
HSI stock trading volume surged to HK$289.5 billion.
IPO financing in Hong Kong in the first half raised HK$212.4 billion, ranking second globally.
There are 345 companies lining up to get listed
Haitong and CICC are calling the shots!
Mainland insurance funds can buy Hong Kong ETFs via Stock Connect
Plus, U.S.-listed Chinese concept stocks have returned to list
PE is only 26x! What do you think?
#港交所 #香港IPO全球第二 https://t.co/FeDwvjVhDw
Made 10.5 billion in the first half; 2Q profit in a single quarter was 5.38 billion
HSI stock trading volume surged to HK$289.5 billion.
IPO financing in Hong Kong in the first half raised HK$212.4 billion, ranking second globally.
There are 345 companies lining up to get listed
Haitong and CICC are calling the shots!
Mainland insurance funds can buy Hong Kong ETFs via Stock Connect
Plus, U.S.-listed Chinese concept stocks have returned to list
PE is only 26x! What do you think?
#港交所 #香港IPO全球第二 https://t.co/FeDwvjVhDw
Ethereum turns bullish brothers (massive pump)
BlackRock has been buying for 5 straight weeks
Leveraged futures have shrunk in volume; the drop is down 70%
All of the web liquidations blew up
ethereum:native https://t.co/lqURgpdELs
BlackRock has been buying for 5 straight weeks
Leveraged futures have shrunk in volume; the drop is down 70%
All of the web liquidations blew up
ethereum:native https://t.co/lqURgpdELs
One general’s success leads to ten thousand skeletons
This kind of battle is really exhilarating
#牛来 https://t.co/UKpai7xqI4
This kind of battle is really exhilarating
#牛来 https://t.co/UKpai7xqI4
Nvidia’s 13F filing reveals 8 holdings
Total value: $63.44 billion, with the largest being Intel at about $30 billion
The second-largest is SpaceX at about $21 billion.
Nvidia’s holdings are almost entirely focused on chips, compute power, communications, and cloud infrastructure.
One of the companies is Generate Biomedicines, which uses compute power to design new drugs; in the future, drug R&D will also become a compute business.
Generate handles protein drugs, while Nvidia sells compute power and tools.
Biopharma companies become GPU’s biggest customers.
#NVDA #SpaceX
Total value: $63.44 billion, with the largest being Intel at about $30 billion
The second-largest is SpaceX at about $21 billion.
Nvidia’s holdings are almost entirely focused on chips, compute power, communications, and cloud infrastructure.
One of the companies is Generate Biomedicines, which uses compute power to design new drugs; in the future, drug R&D will also become a compute business.
Generate handles protein drugs, while Nvidia sells compute power and tools.
Biopharma companies become GPU’s biggest customers.
#NVDA #SpaceX
August 14 earnings report: the state-owned “National Team” clears out Moutai! Institutions move in and take over turnover.
Central Huijin: fully exited the position. Previously held 10.4 million shares
CSIC Securities (China Securities Finance Corporation): fully exited the position. Previously held 4.04 million shares
HuaXia SSE 50 ETF: fully exited the position. Previously held 4.57 million shares
Huatai-Pinebridge CSI 300 ETF: fully exited the position. Previously held 5.04 million shares
Northbound funds reduced holdings by 8.55% to 53.71 million shares
China Merchants CSI Liquor ETF reduced holdings by 10.25% to 4.56 million shares
China Life: newly bought 5.58 million shares
Huatai Asset Management FOF: newly bought 4.03 million shares
Dacheng Fund FOF: newly bought 3.72 million shares
Guizhou Financial Holdings: newly bought 3.49 million shares
Conclusion: Moutai is currently taking back pricing power and sales rights from distributors. In the past, the terminal price of Feitian Moutai was 3,000 yuan, with a factory-gate price of 1,169 yuan. The entire price spread of 1,800 yuan in between was captured by distributors and scalpers.
Now Moutai connects directly with consumers through iMoutai. In the first half of the year, it collected 40.2 billion yuan directly into its own pocket.
Central Huijin: fully exited the position. Previously held 10.4 million shares
CSIC Securities (China Securities Finance Corporation): fully exited the position. Previously held 4.04 million shares
HuaXia SSE 50 ETF: fully exited the position. Previously held 4.57 million shares
Huatai-Pinebridge CSI 300 ETF: fully exited the position. Previously held 5.04 million shares
Northbound funds reduced holdings by 8.55% to 53.71 million shares
China Merchants CSI Liquor ETF reduced holdings by 10.25% to 4.56 million shares
China Life: newly bought 5.58 million shares
Huatai Asset Management FOF: newly bought 4.03 million shares
Dacheng Fund FOF: newly bought 3.72 million shares
Guizhou Financial Holdings: newly bought 3.49 million shares
Conclusion: Moutai is currently taking back pricing power and sales rights from distributors. In the past, the terminal price of Feitian Moutai was 3,000 yuan, with a factory-gate price of 1,169 yuan. The entire price spread of 1,800 yuan in between was captured by distributors and scalpers.
Now Moutai connects directly with consumers through iMoutai. In the first half of the year, it collected 40.2 billion yuan directly into its own pocket.
Semiconductor Manufacturing International Corporation (SMIC) Valuation: Data from Six Investment Banks, Three Valuation Frameworks, One Conclusion
Investment Banks’ Target Prices: Substantial Disagreement
Goldman Sachs: HK$135
BNP Paribas: HK$94.1
UBS: HK$96.2
Citigroup: HK$90
Morgan Stanley: HK$85
Nomura: HK$75
Three Valuation Approaches Compared
1. P/E Method — the Most Optimistic Play
Goldman Sachs estimates SMIC’s EPS for 2026–2028 rises from US$0.15 to US$0.28, which implies the P/E multiple drops from 58x to 30x.
2. P/B Method — Most Institutions Use This
Current net asset value per share is about HK$19.8, with a P/B of 3.57x. Citigroup uses 4x → HK$90, while UBS uses 3.3x → HK$96.2, because net assets roll upward. The 3.57x range is slightly above the historical midline—neither cheap nor at the top.
3. DCF — Cannot Produce Positive Value
Free cash flow in 2025 is -US$4.9 billion, 2026 is forecast at -US$2.4 billion, 2027 at -US$3.1 billion, 2028 at -US$1.0 billion, and it turns positive only in 2029. Net debt is US$16.9 billion. Building fabs costs US$8 billion per year—any DCF model cannot yield a positive valuation.
New Development: Q2 EBITDA Profit Margin 70%, up 13 percentage points quarter-over-quarter.
Management says depreciation peaks only in 2027; in other words, the company’s profitability improves, and depreciation absorbs part of those gains.
Personal Summary:
SMIC is not a stock whose value can be neatly calculated with a DCF. If you look at P/E, it’s expensive at 58x; if you look at P/B, it looks reasonable (around 3x); and if you use DCF, the valuation comes out to no value—cash flow is negative for five consecutive years.
SMIC cannot be valued with conventional methods. This is a product created in response to supply-chain bottlenecks—something that embodies the country’s lifeline in semiconductors.
#中芯国际 #semiconductors
Investment Banks’ Target Prices: Substantial Disagreement
Goldman Sachs: HK$135
BNP Paribas: HK$94.1
UBS: HK$96.2
Citigroup: HK$90
Morgan Stanley: HK$85
Nomura: HK$75
Three Valuation Approaches Compared
1. P/E Method — the Most Optimistic Play
Goldman Sachs estimates SMIC’s EPS for 2026–2028 rises from US$0.15 to US$0.28, which implies the P/E multiple drops from 58x to 30x.
2. P/B Method — Most Institutions Use This
Current net asset value per share is about HK$19.8, with a P/B of 3.57x. Citigroup uses 4x → HK$90, while UBS uses 3.3x → HK$96.2, because net assets roll upward. The 3.57x range is slightly above the historical midline—neither cheap nor at the top.
3. DCF — Cannot Produce Positive Value
Free cash flow in 2025 is -US$4.9 billion, 2026 is forecast at -US$2.4 billion, 2027 at -US$3.1 billion, 2028 at -US$1.0 billion, and it turns positive only in 2029. Net debt is US$16.9 billion. Building fabs costs US$8 billion per year—any DCF model cannot yield a positive valuation.
New Development: Q2 EBITDA Profit Margin 70%, up 13 percentage points quarter-over-quarter.
Management says depreciation peaks only in 2027; in other words, the company’s profitability improves, and depreciation absorbs part of those gains.
Personal Summary:
SMIC is not a stock whose value can be neatly calculated with a DCF. If you look at P/E, it’s expensive at 58x; if you look at P/B, it looks reasonable (around 3x); and if you use DCF, the valuation comes out to no value—cash flow is negative for five consecutive years.
SMIC cannot be valued with conventional methods. This is a product created in response to supply-chain bottlenecks—something that embodies the country’s lifeline in semiconductors.
#中芯国际 #semiconductors
Tencent’s second-quarter results are out—I read all three big investment bank reports.
No more nonsense, let’s get to the data.
All the money they earned has been poured into AI.
In Q2, Tencent invested RMB 10.5 billion in AI. Analysts predict the full-year capital expenditures will reach RMB 210 billion, nearly RMB 60 billion more than at the beginning of the year. Cash flow: -13.8 billion.
To build AI, Tencent is willing to give up short-term GPU rental revenue and focus first on getting foundational models and AI applications up and running.
Priority order for computing power allocation: model training > AI applications > cloud services.
In other words: even if it makes less money in the short term, they’d rather get AI rolling.
Three data points worth watching:
1: Advertising revenue year over year +22%.
2: Domestic gaming revenue year over year +17%—stable cash flow.
3: In 2026, watch whether AI can really scale up and generate gains.
Final summary:
Goldman Sachs → target price HKD 670.
Nomura → target price HKD 727.
Morgan Stanley → target price HKD 690.
“Hunyuan 4” will be released in the second half of the year. WorkBuddy’s paid-user retention rate, and WeChat Xiaoa’s data—these three metrics should be monitored continuously.
Once the numbers catch up, Tencent with a 13x PE looks cheap.
#腾讯 #Tencent quarterly report
No more nonsense, let’s get to the data.
All the money they earned has been poured into AI.
In Q2, Tencent invested RMB 10.5 billion in AI. Analysts predict the full-year capital expenditures will reach RMB 210 billion, nearly RMB 60 billion more than at the beginning of the year. Cash flow: -13.8 billion.
To build AI, Tencent is willing to give up short-term GPU rental revenue and focus first on getting foundational models and AI applications up and running.
Priority order for computing power allocation: model training > AI applications > cloud services.
In other words: even if it makes less money in the short term, they’d rather get AI rolling.
Three data points worth watching:
1: Advertising revenue year over year +22%.
2: Domestic gaming revenue year over year +17%—stable cash flow.
3: In 2026, watch whether AI can really scale up and generate gains.
Final summary:
Goldman Sachs → target price HKD 670.
Nomura → target price HKD 727.
Morgan Stanley → target price HKD 690.
“Hunyuan 4” will be released in the second half of the year. WorkBuddy’s paid-user retention rate, and WeChat Xiaoa’s data—these three metrics should be monitored continuously.
Once the numbers catch up, Tencent with a 13x PE looks cheap.
#腾讯 #Tencent quarterly report
Partly True
The credit card industry’s average annual interest rate is 21.47%. Trump wants to push it down to 10%, but banks oppose it—this year, the chance of approval is only 10%.
Institutional outlook
Stablecoin annual payment flows: $1.6 trillion in January 2024, rising to $5.66 trillion by 2030.
Why USDC looks set to break through on three fronts:
1. The first to obtain both US and EU dual compliance licenses, with 1:1 reserves covering all assets, and monthly disclosures
2. USDT was blocked due to its place of registration in El Salvador, forcing it to launch a new token to take a different compliance route
3. Meta directly chose USDC for creator payments, and USDC gained another distribution channel
USDC’s market-cap share is about 28%, but its share of trading volume has already exceeded 60%. In full-year 2025, it processed $4.5 trillion in transaction volume,
while in just the first half of 2026 it already processed $5.3 trillion—up 140% compared with 2025.
At an investors’ call, Circle said something quite straightforward: large institutions won’t use Tether; they use USDC—what comes next, you all understand.
base:0x833589fcd6edb6e08f4c7c32d4f71b54bda02913 binancecoin:native
Institutional outlook
Stablecoin annual payment flows: $1.6 trillion in January 2024, rising to $5.66 trillion by 2030.
Why USDC looks set to break through on three fronts:
1. The first to obtain both US and EU dual compliance licenses, with 1:1 reserves covering all assets, and monthly disclosures
2. USDT was blocked due to its place of registration in El Salvador, forcing it to launch a new token to take a different compliance route
3. Meta directly chose USDC for creator payments, and USDC gained another distribution channel
USDC’s market-cap share is about 28%, but its share of trading volume has already exceeded 60%. In full-year 2025, it processed $4.5 trillion in transaction volume,
while in just the first half of 2026 it already processed $5.3 trillion—up 140% compared with 2025.
At an investors’ call, Circle said something quite straightforward: large institutions won’t use Tether; they use USDC—what comes next, you all understand.
base:0x833589fcd6edb6e08f4c7c32d4f71b54bda02913 binancecoin:native
The US spent nearly $600 billion in 2026 on AI
That’s about 2% of GDP, but the net boost to GDP is only 0.1–0.3 percentage points.
An analysis by institutions attributes it to three reasons:
1. Most of it is imported hardware, so it isn’t counted in GDP statistics.
2. Enterprises’ software spending they save is considered intermediate services and isn’t included in GDP.
3. Various crowding-out effects replace other investment—competing for building resources and pushing up interest rates add up to only $50 billion, far less than people imagine.
89% of surveyed companies say their AI costs make up just 1–5% of the IT budget, and one-third comes from new spending.
Deeper reasons:
The GDP accounting rules were created for the industrial era—imported hardware counts as costs, while intermediate services count as consumption. The real productivity gains brought by AI aren’t captured in the same framework.
What happens next?
Most of the money goes to buying GPUs and building data centers—those count as spending in GDP.
Later, when more spending shifts to software and applications, helping businesses save money and make money, GDP will reflect it more accurately.
One-sentence summary: In the first phase they buy hardware (GPUs, PCBs, optical modules). In the second phase they buy ASICs and servers. In the third phase they buy software and applications. In the fourth phase they buy robots and automation.
Right now, the first phase has mostly topped out, and the turning points for the second and third phases are approaching.
#AI binancecoin:native
That’s about 2% of GDP, but the net boost to GDP is only 0.1–0.3 percentage points.
An analysis by institutions attributes it to three reasons:
1. Most of it is imported hardware, so it isn’t counted in GDP statistics.
2. Enterprises’ software spending they save is considered intermediate services and isn’t included in GDP.
3. Various crowding-out effects replace other investment—competing for building resources and pushing up interest rates add up to only $50 billion, far less than people imagine.
89% of surveyed companies say their AI costs make up just 1–5% of the IT budget, and one-third comes from new spending.
Deeper reasons:
The GDP accounting rules were created for the industrial era—imported hardware counts as costs, while intermediate services count as consumption. The real productivity gains brought by AI aren’t captured in the same framework.
What happens next?
Most of the money goes to buying GPUs and building data centers—those count as spending in GDP.
Later, when more spending shifts to software and applications, helping businesses save money and make money, GDP will reflect it more accurately.
One-sentence summary: In the first phase they buy hardware (GPUs, PCBs, optical modules). In the second phase they buy ASICs and servers. In the third phase they buy software and applications. In the fourth phase they buy robots and automation.
Right now, the first phase has mostly topped out, and the turning points for the second and third phases are approaching.
#AI binancecoin:native
Europe — the Hidden Winners
Enough talk, get straight to the point
1. Europe’s EPS grew 14% in the first half, the strongest in three years; full-year growth has been raised to 15%—who says Europe isn’t growing?
2. Europe’s ROE climbed from 9% in 2021 to 15% in 2026 step by step, with banks being the biggest driving force behind the turnaround.
3. Since 2022, European bank stocks have outperformed the U.S. Big Tech’s ‘Seven Giants’ by 2x; this year, the STOXX 600 total return is 14%, beating the S&P 500.
4. Median European stocks are up 13%, basically in line with the index’s gains—not just a handful of big names; it’s more the rally in mid- and small-cap stocks, so don’t obsess over U.S. stocks every day.
5. Even with China’s exports surging, European automakers account for only 1% of market value—there aren’t many industries that are truly taking a serious hit.
One-sentence summary: In recent years, Europe has been quietly making money—most people just still hold onto the impression from ten years ago.
#EuropeOutperforms #HiddenWinner binancecoin:native
Enough talk, get straight to the point
1. Europe’s EPS grew 14% in the first half, the strongest in three years; full-year growth has been raised to 15%—who says Europe isn’t growing?
2. Europe’s ROE climbed from 9% in 2021 to 15% in 2026 step by step, with banks being the biggest driving force behind the turnaround.
3. Since 2022, European bank stocks have outperformed the U.S. Big Tech’s ‘Seven Giants’ by 2x; this year, the STOXX 600 total return is 14%, beating the S&P 500.
4. Median European stocks are up 13%, basically in line with the index’s gains—not just a handful of big names; it’s more the rally in mid- and small-cap stocks, so don’t obsess over U.S. stocks every day.
5. Even with China’s exports surging, European automakers account for only 1% of market value—there aren’t many industries that are truly taking a serious hit.
One-sentence summary: In recent years, Europe has been quietly making money—most people just still hold onto the impression from ten years ago.
#EuropeOutperforms #HiddenWinner binancecoin:native
AI era: Differences in China and the US in Power Supply
In 2025, the total electricity consumption in the United States reached 4.58 trillion kilowatt-hours, a historic high
Industry projections indicate that the power shortfall for North American data centers in 2025–2028 will reach 44 GW.
China’s annual electricity generation averages about 10 trillion kilowatt-hours. China does not face nationwide power shortages, but rather structural, time-based, and regional tightness: there is surplus power in the northwest, while coastal areas are tighter.
China’s industrial electricity price of 0.2–0.3 yuan per kWh in the west is a cost advantage in the global AI compute industry.
For every 1 kWh, when sold through compute services, the accompanying GDP value far exceeds selling electricity directly.
This is using electricity as leverage: the same amount of electricity, if not used for compute, is only an industrial price of 0.3 yuan per kWh.
Put it into GPUs, run to produce tokens, and sell to the world—thereby magnifying the value of electricity by dozens or even hundreds of times.
#AI binancecoin:native #power
In 2025, the total electricity consumption in the United States reached 4.58 trillion kilowatt-hours, a historic high
Industry projections indicate that the power shortfall for North American data centers in 2025–2028 will reach 44 GW.
China’s annual electricity generation averages about 10 trillion kilowatt-hours. China does not face nationwide power shortages, but rather structural, time-based, and regional tightness: there is surplus power in the northwest, while coastal areas are tighter.
China’s industrial electricity price of 0.2–0.3 yuan per kWh in the west is a cost advantage in the global AI compute industry.
For every 1 kWh, when sold through compute services, the accompanying GDP value far exceeds selling electricity directly.
This is using electricity as leverage: the same amount of electricity, if not used for compute, is only an industrial price of 0.3 yuan per kWh.
Put it into GPUs, run to produce tokens, and sell to the world—thereby magnifying the value of electricity by dozens or even hundreds of times.
#AI binancecoin:native #power
In the first 7 months, China’s integrated circuit exports amounted to US$216 billion.
In July alone, exports were US$38.74 billion.
Now let’s look at Cambricon’s interim report.
Revenue was RMB 5.996 billion.
Net profit was RMB 2.311 billion, up 122.6% year-on-year.
Q2 single-quarter net profit was RMB 1.298 billion, 28% higher than in Q1.
Cambricon’s inventory reached RMB 8.248 billion, and prepayments were RMB 2.914 billion.
The export amount doubled, while the number of units remained almost unchanged.
The era of low-price volume sales is gradually coming to an end.
#domesticchips
#Cambricon
In July alone, exports were US$38.74 billion.
Now let’s look at Cambricon’s interim report.
Revenue was RMB 5.996 billion.
Net profit was RMB 2.311 billion, up 122.6% year-on-year.
Q2 single-quarter net profit was RMB 1.298 billion, 28% higher than in Q1.
Cambricon’s inventory reached RMB 8.248 billion, and prepayments were RMB 2.914 billion.
The export amount doubled, while the number of units remained almost unchanged.
The era of low-price volume sales is gradually coming to an end.
#domesticchips
#Cambricon
Just finished the Huatai companion robot report. A robot isn’t only able to do laundry and cook. Its biggest purpose is (to cure loneliness).
The Japanese LOVOT won’t do any household chores.
It will warm up, look at you, come close to you, and hug you. After 3 years, the retention rate is about 90%, and users hold it every day for more than 1 hour.
One unit sells for nearly 30,000, with a monthly fee of a few hundred. Clothes, glasses, and harnesses can also keep being sold using this model—so the whole thing has already proven to work.
China has even more opportunities. Thanks to the supply chain and global manufacturing capacity, a companion robot can be replicated for just a few thousand yuan,
UBTECH: https://t.co/q672Alkdzg
Anker’s Home companion robot for China; https://t.co/O4vdUwX2kR
Guangyingshi network in China; https://t.co/jFNTFxuHAl
Ecovacs in China; https://t.co/O61rr9Q6cx
Casio in Japan; 6952.T
The most pure original brand is Japan’s LOVOT—its parent company, Groove X, hasn’t
been listed yet.
The next round of competition in the robotics space isn’t just about physical AI model evolution—it's about who can buy a robot home and treat it as a family member.
#AI陪伴机器人 #具身智能 #机器人经济 #LOVOT #优必选 ＃卧安机器人 ＃humanoid robot binancecoin:native
The Japanese LOVOT won’t do any household chores.
It will warm up, look at you, come close to you, and hug you. After 3 years, the retention rate is about 90%, and users hold it every day for more than 1 hour.
One unit sells for nearly 30,000, with a monthly fee of a few hundred. Clothes, glasses, and harnesses can also keep being sold using this model—so the whole thing has already proven to work.
China has even more opportunities. Thanks to the supply chain and global manufacturing capacity, a companion robot can be replicated for just a few thousand yuan,
UBTECH: https://t.co/q672Alkdzg
Anker’s Home companion robot for China; https://t.co/O4vdUwX2kR
Guangyingshi network in China; https://t.co/jFNTFxuHAl
Ecovacs in China; https://t.co/O61rr9Q6cx
Casio in Japan; 6952.T
The most pure original brand is Japan’s LOVOT—its parent company, Groove X, hasn’t
been listed yet.
The next round of competition in the robotics space isn’t just about physical AI model evolution—it's about who can buy a robot home and treat it as a family member.
#AI陪伴机器人 #具身智能 #机器人经济 #LOVOT #优必选 ＃卧安机器人 ＃humanoid robot binancecoin:native
FOMO➕DLINK
The strongest competitor to PUMP is here
A weekly income of 3.25 million
Used FOMO for 2 days—didn’t really get used to it at first
Now it feels pretty good
#FOMO solana:pumpCmXqMfrsAkQ5r49WcJnRayYRqmXz6ae8H7H9Dfn $DLINK https://t.co/IPuCz2ftOn
The strongest competitor to PUMP is here
A weekly income of 3.25 million
Used FOMO for 2 days—didn’t really get used to it at first
Now it feels pretty good
#FOMO solana:pumpCmXqMfrsAkQ5r49WcJnRayYRqmXz6ae8H7H9Dfn $DLINK https://t.co/IPuCz2ftOn
As of the latest A-share data (as of August 7), ETF holdings ranking
In the All-A ETF market, the last 1/3/5 days saw net outflows of 23.878 billion, 80.439 billion, and 74.984 billion units respectively; for A500, holdings decreased by 12.674 billion, 43.407 billion, and 43.227 billion units respectively. Broad-based funds continue to be redeemed.
The thematic side is not doing much better: AI saw decreases of 6.021 billion, 15.740 billion, and 12.776 billion units over the last 1/3/5 days.
Robotics fell by 0.617 billion, 1.282 billion, and 0.873 billion units respectively. CPO, optical communications, and semiconductor equipment are all among the leaders in share reductions.
There are few directions with inflows: energy storage increased by 224 million units in a single day, and grid equipment added 250 million units over the past ~5 days. It’s more like localized probing; it’s not time for a broad-based rebound yet.
#ETF #A股
In the All-A ETF market, the last 1/3/5 days saw net outflows of 23.878 billion, 80.439 billion, and 74.984 billion units respectively; for A500, holdings decreased by 12.674 billion, 43.407 billion, and 43.227 billion units respectively. Broad-based funds continue to be redeemed.
The thematic side is not doing much better: AI saw decreases of 6.021 billion, 15.740 billion, and 12.776 billion units over the last 1/3/5 days.
Robotics fell by 0.617 billion, 1.282 billion, and 0.873 billion units respectively. CPO, optical communications, and semiconductor equipment are all among the leaders in share reductions.
There are few directions with inflows: energy storage increased by 224 million units in a single day, and grid equipment added 250 million units over the past ~5 days. It’s more like localized probing; it’s not time for a broad-based rebound yet.
#ETF #A股
Log in to explore more content