Ray區塊鏈與不動產
165 Posts
Ray區塊鏈與不動產
幣圈擼羊毛空投|手把手教學|推特：@RayBlock_house、Youtuber,嗶哩嗶哩：Ray區塊鏈與不動產
Open Trade
U Holder
Frequent Trader
5.1 Years
9 Following
176 Followers
325 Liked
Posts
Portfolio
All
Quotes
TSMC’s $TSM monthly revenue hits a new high!
TSMC’s full-year revenue last year was 3.8 trillion yuan
In the first half of this year alone, revenue has already reached 2.4 trillion yuan
An increase of 35.6% compared with the same period last year.
June revenue was 442.68 billion yuan, up 67.9% year over year, setting a new all-time monthly high again.
TSMC, save us all~
$TSMon
TSMC’s full-year revenue last year was 3.8 trillion yuan
In the first half of this year alone, revenue has already reached 2.4 trillion yuan
An increase of 35.6% compared with the same period last year.
June revenue was 442.68 billion yuan, up 67.9% year over year, setting a new all-time monthly high again.
TSMC, save us all~
$TSMon
White House stock god Trump calls out trades again!
He said: “Go buy a Dell computer, because even his son is nagging for a Dell computer.””
$DELL Dell donates 6.25 billion to the “Trump account”
Trump: “Early investment is a way to create long-term wealth. I believe the stock market will soar to new heights.”
So, this is basically a donation-and-calling-the-trades situation.
By comparison, Micron’s donation of 250 million is a bit small
He said: “Go buy a Dell computer, because even his son is nagging for a Dell computer.””
$DELL Dell donates 6.25 billion to the “Trump account”
Trump: “Early investment is a way to create long-term wealth. I believe the stock market will soar to new heights.”
So, this is basically a donation-and-calling-the-trades situation.
By comparison, Micron’s donation of 250 million is a bit small
SemiAnalysis Report: $NVDA Kyber NVL144 Rack Postponed Over 12 Months to 2028
“Due to the high difficulty of manufacturing the PCB mid-plate, it has been postponed by more than 12 months, and is expected to be released in 2028.”
At the same time, the NVL72x2 back-to-back rack architecture was canceled, as the design format and maintenance burden were opposed by cloud service providers, leading to the termination.
The original 4-chip version of Rubin Ultra has been canceled; only the 2-compute-chip version remains, with actual performance reduced to half of the original plan. Nvidia will increase sales of the Oberon Rubin and Rubin Ultra racks to make up for the shortfall.
However! On June 9, Semi issued an organizational report stating that NVIDIA CPO (co-packaged optics) and large-scale mass production of 800VDC have been delayed to 2028 or later, causing the fiber-optic communications stocks to drop significantly.
About three weeks after the report was released, in late June, SemiAnalysis and Tema ETFs announced a partnership to launch a photonics and optics ETF (LAZR) and other semiconductor ETFs, focusing on the AI infrastructure supply chain.
“Due to the high difficulty of manufacturing the PCB mid-plate, it has been postponed by more than 12 months, and is expected to be released in 2028.”
At the same time, the NVL72x2 back-to-back rack architecture was canceled, as the design format and maintenance burden were opposed by cloud service providers, leading to the termination.
The original 4-chip version of Rubin Ultra has been canceled; only the 2-compute-chip version remains, with actual performance reduced to half of the original plan. Nvidia will increase sales of the Oberon Rubin and Rubin Ultra racks to make up for the shortfall.
However! On June 9, Semi issued an organizational report stating that NVIDIA CPO (co-packaged optics) and large-scale mass production of 800VDC have been delayed to 2028 or later, causing the fiber-optic communications stocks to drop significantly.
About three weeks after the report was released, in late June, SemiAnalysis and Tema ETFs announced a partnership to launch a photonics and optics ETF (LAZR) and other semiconductor ETFs, focusing on the AI infrastructure supply chain.
Meta rents out more than computing power! But is there really excess computing power right now?
$META is reportedly planning cloud business, renting out remaining AI computing resources to compete with cloud service providers such as Amazon AWS, Microsoft Azure, and Google Cloud.
The market reaction to this news is that capital expenditures will be reduced, so upstream AI hardware manufacturers have fallen.
But I think the market reaction is overblown. It’s still too early to interpret this news as “a coming surplus of computing capacity.”
Based on Meta’s current public situation, it seems closer to “continuing to obtain more computing power.”
1. Computing power is scarce (In March, Google limited Gemini’s computing capacity, delaying some projects)
2. It’s so tight that even DDR4 has to be reclaimed for reuse, while DDR5 is mixed in (06-29 ISCA 2026)
3. Meta continues renting and using computing power.
In March, it signed a five-year AI infrastructure agreement with Nebius, worth up to 27 billion USD.
4. In April, it expanded its partnership with CoreWeave, signing a 21 billion USD AI cloud computing agreement.
6. In June, it reached an agreement with Crusoe, renting approximately 1.6 GW of compute capacity.
My guess is that Meta may have seen xAI “monetizing idle servers for rent,” effectively acting as a landlord to collect rent and accelerate cash recovery.
Because Meta’s spending in 2026 is 125–145 billion USD, this kind of approach can ease market concerns about Meta making heavy AI investments and continuously burning cash, while also creating a new revenue stream. After all, using Meta AI to improve ads only increases revenue as much as that.
Most importantly, is Meta’s own AI still so weak that it can only sell computing power?
Similar approaches exist in Elon Musk’s SpaceX renting out idle data centers:
In May, it rented to Anthropic for 1.25 billion USD per month.
In June, it rented to Google for 0.92 billion USD per month.
First, act as the landlord to rent out compute and pocket 2.17 billion per month.
The market didn’t really react to SpaceX renting excess computing power to Anthropic or Google, but when Meta rents out computing power, it seems like “the sky is falling.”?
$META is reportedly planning cloud business, renting out remaining AI computing resources to compete with cloud service providers such as Amazon AWS, Microsoft Azure, and Google Cloud.
The market reaction to this news is that capital expenditures will be reduced, so upstream AI hardware manufacturers have fallen.
But I think the market reaction is overblown. It’s still too early to interpret this news as “a coming surplus of computing capacity.”
Based on Meta’s current public situation, it seems closer to “continuing to obtain more computing power.”
1. Computing power is scarce (In March, Google limited Gemini’s computing capacity, delaying some projects)
2. It’s so tight that even DDR4 has to be reclaimed for reuse, while DDR5 is mixed in (06-29 ISCA 2026)
3. Meta continues renting and using computing power.
In March, it signed a five-year AI infrastructure agreement with Nebius, worth up to 27 billion USD.
4. In April, it expanded its partnership with CoreWeave, signing a 21 billion USD AI cloud computing agreement.
6. In June, it reached an agreement with Crusoe, renting approximately 1.6 GW of compute capacity.
My guess is that Meta may have seen xAI “monetizing idle servers for rent,” effectively acting as a landlord to collect rent and accelerate cash recovery.
Because Meta’s spending in 2026 is 125–145 billion USD, this kind of approach can ease market concerns about Meta making heavy AI investments and continuously burning cash, while also creating a new revenue stream. After all, using Meta AI to improve ads only increases revenue as much as that.
Most importantly, is Meta’s own AI still so weak that it can only sell computing power?
Similar approaches exist in Elon Musk’s SpaceX renting out idle data centers:
In May, it rented to Anthropic for 1.25 billion USD per month.
In June, it rented to Google for 0.92 billion USD per month.
First, act as the landlord to rent out compute and pocket 2.17 billion per month.
The market didn’t really react to SpaceX renting excess computing power to Anthropic or Google, but when Meta rents out computing power, it seems like “the sky is falling.”?
Corning Glass in U.S. stocks $GLW has broken to a new high again
High-speed transmission in AI servers uses optical fiber, and the first company that comes to mind is Corning. It has been making glass for over a hundred years; the glass for Edison light bulbs and the first communication optical fiber must reflect its leading advantages.
Before hearing the introduction, I thought Corning still sold tableware. Later I learned that the tableware brand had long been separated from the current Corning. Today, Corning mainly focuses on touch display glass, specialty materials, and optical fiber communications.
In short, it’s all about:
making glass, making glass, and still making glass.
Purely based on my past impression, plus the feeling that the company’s benefits and atmosphere seemed pretty good, I bought it like that.
High-speed transmission in AI servers uses optical fiber, and the first company that comes to mind is Corning. It has been making glass for over a hundred years; the glass for Edison light bulbs and the first communication optical fiber must reflect its leading advantages.
Before hearing the introduction, I thought Corning still sold tableware. Later I learned that the tableware brand had long been separated from the current Corning. Today, Corning mainly focuses on touch display glass, specialty materials, and optical fiber communications.
In short, it’s all about:
making glass, making glass, and still making glass.
Purely based on my past impression, plus the feeling that the company’s benefits and atmosphere seemed pretty good, I bought it like that.
GLWUS-10.97%
Micron $MU 2026Q3 Financial Report: Revenue, profits, gross margin, and free cash flow have all hit historical highs.
Revenue was $41.46 billion, up 74% quarter-over-quarter and 346% year-over-year.
“Gross margin: 84.9%”
Customers are going crazy—don’t worry if the price is too high; they’re only afraid they won’t get any supply!
The biggest reason for this growth is “a substantial price increase,” while production increases very little. The fabs are still being acquired, setting up production lines, waiting for equipment—everything takes time.
DRAM shipment volume increases only by a single-digit percentage, but the average selling price rises by more than 60%.
NAND shipment volume increases only by a single-digit percentage, but the average selling price rises by more than 80%.
Will next quarter be even stronger? Forward committed orders?
Micron’s outlook for FY2026 Q4:
▪️ Revenue of about $50 billion ▪️ Gross margin of about 86% ▪️ EPS about $31
That means: after Micron’s revenue tops $41.46 billion in a single quarter, next quarter could still grow by around 20%.
Micron has already signed 16 multi-year strategic customer agreements covering DRAM and NAND supply for the coming years.
The lowest disclosed revenue commitment so far is about $100 billion.
In the future, nearly half of Micron’s revenue will come from these long-term agreements. Some contracts also use a “pay-as-a-buyer” model—customers must fulfill minimum purchase commitments even if they don’t use up the originally reserved capacity.
Why does it seem like the AI era always has a shortage of storage?
Because after AI compute power increases, it’s not only that the number of GPUs grows— the memory capacity needed alongside each GPU is rising rapidly as well.
More GPUs → higher HBM demand → DRAM production capacity gets crowded out → general memory prices rise → more data volume → increased demand for NAND and SSDs.
--------------
The most obvious feeling is that there are more and more AI videos on YT—such as China’s AI short dramas. Same script and story: at the beginning, the audio volume is up and down, the characters’ voices and appearances are inconsistent—you can’t really watch it. But now the gap between them and real-life filming is getting smaller and smaller. Lower cost, faster speed.
And also!
TSMC’s near-monopoly advantage in advanced process technology and advanced packaging yields a gross margin of about 66%; but Micron’s gross margin this quarter is as high as 84.9%.
Hope that “this time will definitely be different”—that memory and storage can break away from cyclical industries. But I’m a bit worried that this approach may not be sustainable.
Revenue was $41.46 billion, up 74% quarter-over-quarter and 346% year-over-year.
“Gross margin: 84.9%”
Customers are going crazy—don’t worry if the price is too high; they’re only afraid they won’t get any supply!
The biggest reason for this growth is “a substantial price increase,” while production increases very little. The fabs are still being acquired, setting up production lines, waiting for equipment—everything takes time.
DRAM shipment volume increases only by a single-digit percentage, but the average selling price rises by more than 60%.
NAND shipment volume increases only by a single-digit percentage, but the average selling price rises by more than 80%.
Will next quarter be even stronger? Forward committed orders?
Micron’s outlook for FY2026 Q4:
▪️ Revenue of about $50 billion ▪️ Gross margin of about 86% ▪️ EPS about $31
That means: after Micron’s revenue tops $41.46 billion in a single quarter, next quarter could still grow by around 20%.
Micron has already signed 16 multi-year strategic customer agreements covering DRAM and NAND supply for the coming years.
The lowest disclosed revenue commitment so far is about $100 billion.
In the future, nearly half of Micron’s revenue will come from these long-term agreements. Some contracts also use a “pay-as-a-buyer” model—customers must fulfill minimum purchase commitments even if they don’t use up the originally reserved capacity.
Why does it seem like the AI era always has a shortage of storage?
Because after AI compute power increases, it’s not only that the number of GPUs grows— the memory capacity needed alongside each GPU is rising rapidly as well.
More GPUs → higher HBM demand → DRAM production capacity gets crowded out → general memory prices rise → more data volume → increased demand for NAND and SSDs.
--------------
The most obvious feeling is that there are more and more AI videos on YT—such as China’s AI short dramas. Same script and story: at the beginning, the audio volume is up and down, the characters’ voices and appearances are inconsistent—you can’t really watch it. But now the gap between them and real-life filming is getting smaller and smaller. Lower cost, faster speed.
And also!
TSMC’s near-monopoly advantage in advanced process technology and advanced packaging yields a gross margin of about 66%; but Micron’s gross margin this quarter is as high as 84.9%.
Hope that “this time will definitely be different”—that memory and storage can break away from cyclical industries. But I’m a bit worried that this approach may not be sustainable.
The US is skeptical that the most advanced EUV lithography machine, $ASML , might have slipped into China.
However, ASML denies it. Given the machine's massive size and its automated remote monitoring system, it's nearly impossible for clients to move or operate the equipment without ASML's tech crew on site.
If I remember correctly, the machine locks up automatically if it experiences any unusual vibrations, right?
This whole situation feels like fake news. Maybe Trump is trying to play the same old game, like he did with Intel, to get the US government a stake in ASML.
However, ASML denies it. Given the machine's massive size and its automated remote monitoring system, it's nearly impossible for clients to move or operate the equipment without ASML's tech crew on site.
If I remember correctly, the machine locks up automatically if it experiences any unusual vibrations, right?
This whole situation feels like fake news. Maybe Trump is trying to play the same old game, like he did with Intel, to get the US government a stake in ASML.
Verified
SpaceX's Biggest IPO Ever【xStocks Out of Stock】
Status of Handling and Compensation Across Major Exchanges
binance - 100% refund, plus 1 M SPCXB worth about 35u.
⚠️ Luckily, the hedge position for $SPCX wasn't too heavy, and I managed to dodge a bullet in the end. Naturally, I thought I'd definitely get my coins, but in reality, until the coins are in hand, there's still a lot of uncertainty. A low-probability event like today can still happen!
Status of Handling and Compensation Across Major Exchanges
binance - 100% refund, plus 1 M SPCXB worth about 35u.
⚠️ Luckily, the hedge position for $SPCX wasn't too heavy, and I managed to dodge a bullet in the end. Naturally, I thought I'd definitely get my coins, but in reality, until the coins are in hand, there's still a lot of uncertainty. A low-probability event like today can still happen!
Tonight #SpaceX IPO at $135 per share
$SPCX valuation at $1.76 trillion
Initial circulation at 4.25% of shares
High market cap, low circulation, retail share at 20%
Isn't this the kind of setup we see with familiar VC coins?
$SPCX valuation at $1.76 trillion
Initial circulation at 4.25% of shares
High market cap, low circulation, retail share at 20%
Isn't this the kind of setup we see with familiar VC coins?
Ah, this...
Verified
Binance officially connects to the US stock market!\n"If the stock market can really make money, opening an account is going to be tough" has never been more real.\n\nIt's hard to believe that a crypto exchange can now directly buy US stocks and ETFs, something we couldn't even imagine before! Sure, the fees are a bit high? But at least we can buy now.\n\nOn May 22, the Chinese SEC cracked down on Tiger Brokers, Futu, and Changqiao, labeling them as illegal cross-border operations. They're not just facing confiscation of all illegal gains, but also severe penalties, with a 2-year transition period (during which you can only sell, not buy or transfer funds).\n\nFor Chinese users, opening a Binance account and depositing is way easier than with traditional brokers, right?\n\nWorried about CRS reporting issues? Soon there will be Binance's stock tokens, bstocks, allowing us to bypass that through blockchain.\n\nChange your app language setting to "Traditional Chinese," and the trading interface will directly show US stocks!
Verified
Making bank in the US stock market is tough!
But now Binance has got your back, officially integrating US stocks.
Set your app language to 'Traditional Chinese' to see US stocks on the trading interface.
1. Trading pairs: 7000+ US stocks & ETFs
2. Fees: Up to $350, a $0.35 platform fee; over $350, it's 0.1%
3. Minimum amount: Starting at $5
4. Supported currencies: USDT, USD1, U, BNB.
USDC trades at 1:1 with USD.
5. Trading hours: 5 * 24 trading
Coming soon, bStocks will be on-chain! (Binance's version of Ondo)
But now Binance has got your back, officially integrating US stocks.
Set your app language to 'Traditional Chinese' to see US stocks on the trading interface.
1. Trading pairs: 7000+ US stocks & ETFs
2. Fees: Up to $350, a $0.35 platform fee; over $350, it's 0.1%
3. Minimum amount: Starting at $5
4. Supported currencies: USDT, USD1, U, BNB.
USDC trades at 1:1 with USD.
5. Trading hours: 5 * 24 trading
Coming soon, bStocks will be on-chain! (Binance's version of Ondo)
Huang Renxun $NVDA GTC AI Supply Chain Partners:
Wang's Pork Zongzi
Flower Girl Restaurant
Rich King Pork Knuckles Premium Restaurant
Brick Oven Nostalgic Restaurant
The brick oven food is really delicious
Wang's Pork Zongzi
Flower Girl Restaurant
Rich King Pork Knuckles Premium Restaurant
Brick Oven Nostalgic Restaurant
The brick oven food is really delicious
The myth of the microstrategy's "never sell Bitcoin" approach comes to an end?
The narrative of the Bitcoin perpetual motion machine for the past $MSTR has been simple:
just buy, don’t sell, issue bonds, issue shares, keep buying BTC, wait for Bitcoin to pump, then push the stock price up, creating a cycle.
Saylor mentioned on the conference call: "We won't just sit there saying 'we will never sell Bitcoin' anymore. We want to be net accumulators of Bitcoin, not only increasing our total Bitcoin holdings but more importantly, enhancing each Bitcoin per share, because we believe that’s the most valuable thing for MSTR shareholders in the long term."
He also added: "If selling Bitcoin can buy dollars, reduce debt, or provide gains per Bitcoin share, we will consider doing that in the future."
This is the first official shift in Strategy since Saylor established the 'never sell' approach, moving towards actively managing the balance sheet, with the core goal shifting from pure hoarding to "maximizing the value of each Bitcoin per share," and even suggesting that they might sell some Bitcoin to pay dividends.
Crazy market inoculation to make the market 'immune'
Currently, Strategy has enough cash to cover dividends for over 18 months, and now indicates the potential to sell Bitcoin in the future, which somewhat reduces concerns about a Ponzi bubble. This represents a shift into a selling phase, and whether they will buy back later is another story.
The narrative of the Bitcoin perpetual motion machine for the past $MSTR has been simple:
just buy, don’t sell, issue bonds, issue shares, keep buying BTC, wait for Bitcoin to pump, then push the stock price up, creating a cycle.
Saylor mentioned on the conference call: "We won't just sit there saying 'we will never sell Bitcoin' anymore. We want to be net accumulators of Bitcoin, not only increasing our total Bitcoin holdings but more importantly, enhancing each Bitcoin per share, because we believe that’s the most valuable thing for MSTR shareholders in the long term."
He also added: "If selling Bitcoin can buy dollars, reduce debt, or provide gains per Bitcoin share, we will consider doing that in the future."
This is the first official shift in Strategy since Saylor established the 'never sell' approach, moving towards actively managing the balance sheet, with the core goal shifting from pure hoarding to "maximizing the value of each Bitcoin per share," and even suggesting that they might sell some Bitcoin to pay dividends.
Crazy market inoculation to make the market 'immune'
Currently, Strategy has enough cash to cover dividends for over 18 months, and now indicates the potential to sell Bitcoin in the future, which somewhat reduces concerns about a Ponzi bubble. This represents a shift into a selling phase, and whether they will buy back later is another story.
The UAE (United Arab Emirates) is exiting OPEC on May 1st, gearing up to significantly boost its independent oil output.
Currently, the UAE has the Abu Dhabi oil pipeline, transporting crude from Habshan to Fujairah Port, around 1.5 to 1.8 million barrels a day, bypassing the Strait of Hormuz for exports.
In the short term, oil prices are influenced by war and transit risks.
In the long run, the UAE's production increase will weaken OPEC and drive prices down!
Currently, the UAE has the Abu Dhabi oil pipeline, transporting crude from Habshan to Fujairah Port, around 1.5 to 1.8 million barrels a day, bypassing the Strait of Hormuz for exports.
In the short term, oil prices are influenced by war and transit risks.
In the long run, the UAE's production increase will weaken OPEC and drive prices down!
Rollback? Fork? $LTC
Litecoin just revealed a zero-day vulnerability, with hackers leveraging the exploit to launch a DoS attack on the mining pool. They managed to execute 'phantom withdrawals' to a third-party DEX by submitting invalid MWEB transactions from outdated mining pool nodes.
The dev team is now patching the hole and has completed a reorg of 13 blocks through hash power consensus, eliminating invalid transactions from the erroneous fork.
The essence of PoW is achieving consensus through computational power.
Litecoin just revealed a zero-day vulnerability, with hackers leveraging the exploit to launch a DoS attack on the mining pool. They managed to execute 'phantom withdrawals' to a third-party DEX by submitting invalid MWEB transactions from outdated mining pool nodes.
The dev team is now patching the hole and has completed a reorg of 13 blocks through hash power consensus, eliminating invalid transactions from the erroneous fork.
The essence of PoW is achieving consensus through computational power.
Tether complies with regulatory requirements on the TRON chain, freezing $344 million USDT.
Will Sun's $TRX on-chain transaction printing opportunity come to a halt?
With such high transaction fees on the Tron chain, it's tough to guess who will be using it, huh?
Will Sun's $TRX on-chain transaction printing opportunity come to a halt?
With such high transaction fees on the Tron chain, it's tough to guess who will be using it, huh?
Contract funding fee collection for 'Position Value'
Open a position of 10,000 U with $RAVE
The funding fee is charged every hour, at 2% each time
This means paying 200 U each time
In 24 hours, it will collect 48%, which is 4800 U
In one day, it will almost consume the principal
Open a position of 10,000 U with $RAVE
The funding fee is charged every hour, at 2% each time
This means paying 200 U each time
In 24 hours, it will collect 48%, which is 4800 U
In one day, it will almost consume the principal
Log in to explore more content