美股投资young

美股投资young

美股科技投资者 ai狂热者 分享每日快讯. all in ai

303Following
796followers

Feed

美股投资young
美股投资young
I just got a share of nmd😅 Are you kidding me $ANT
美股投资young
美股投资young
$NVDA has fallen for five consecutive days. It opened at 159.19 yesterday and closed at 150.96, down 2.89%. From 166.19 on 11/12 to now, it has dropped 9% over five days. It couldn't hold after the surge, similar to how $SPY rose above 733 intraday yesterday but fell back by close. However, $QQQ still closed higher, up 0.39%. $AMD rose 5.31%, $AVGO up 3.65%, and $SOXX also gained slightly by 0.28%. The leading weighted stocks are falling, and the index is being supported by a few others. Before this morning's market open, semiconductors heated up again: $SOXX up 3.6% $AMD up 3.6% $AVGO up 3.2% $NVDA up 2.2% $QQQ up 2.3% $SPY up 1.3% The semiconductor gains are more than double those of $SPY. $QQQ pre-market is at 674.5, still below its 200-day moving average of 681.8. $SOXX is even further below its own 200-day moving average of 160.65. Whether $SPY can hold 733 until the close today depends largely on semiconductors.
美股投资young
美股投资young
The Dow Jones fell 561 points yesterday, not because of tech stocks, but because of healthcare insurance. CMS proposed on Monday to expand healthcare insurance audits. Currently, these are just proposed rules, with public comments open until January 19 next year, and no implementation yet. But the numbers CMS provided are alarming: audit results will be pushed back, starting from payments in 2018, and the insurance industry may have to return about $9 billion annually. Insurance stocks crashed in one day: $UNH fell 10.21% $HUM fell 12.98% $ELV fell 7.63% $CVS fell 5.20% $UNH has a weight of about 4.6% in the Dow, and by itself contributed roughly 40% of the Dow's decline. $AMGN, $JNJ, and $MRK also fell. $DIA closed at 501.73, breaking below its 200-day moving average. On the same day, the Nasdaq rose 0.42%. This morning, $UNH is basically flat in pre-market, $XLV is still slightly down, and the gap has not been filled at all. The rules are not yet in effect, but the money has already fled. If insurance stocks don't stabilize, it will be very difficult for the Dow to return above the 200-day moving average.
美股投资young
美股投资young
$TJX rose 8% yesterday, and retail stocks followed suit with gains. Same-store sales increased by 4%, better than the company's own plan. Revenue was $16 billion, up 6%. Earnings per share were $1.39, and excluding the 5 cents contribution from tariff refunds, it was $1.34. The full-year same-store sales guidance was also raised to 4%. $TGT fell 2% pre-market but closed up 4.01%. $LOW dropped over 4% pre-market but closed up 1.05%. $XRT rose 2.28%, $XLY rose 1.79%. Yesterday, retail was one of the few sectors holding up in the market. However, $RSP still fell slightly by 0.15%, and consumer stocks supported some breadth but couldn't lift the entire index. $XLP also declined; the gains were in discretionary consumer stocks, not the entire consumer sector. Today it's $WMT's earnings report, with almost no movement pre-market. Whether consumer stocks can help $SPY reclaim 733 depends heavily on $WMT alone, which carries more weight than those few companies combined yesterday.
美股投资young
美股投资young
$ANT was listed on Nasdaq today. Last night it was priced at $83, higher than the previous range of $70 to $78, raising about $81.3 billion, with a fully diluted valuation of approximately $1.5 trillion. This price is not cheap. The company’s annualized revenue is about $81 billion, corresponding to a price-to-sales ratio of over 18. And this $81 billion is annualized based on usage from the most recent month, not the full-year revenue already realized. Actual revenue for the first three quarters was $33.2 billion. The moat for large models is very shallow. Today you lead, tomorrow others cut prices, and the day after open-source models catch up. OpenAI and $GOOGL are competing for customers on the same track. The Hong Kong stock market has already gone through this for us: Zhipu ($2513.HK) was listed at HKD 116.2 in January, surged to 2980 in June, and closed at 359.8 today, down 88% from the peak. MiniMax ($0100.HK) was listed at HKD 165 in January, reached 1330 in March, and closed at 114.8 today, already below the issue price. Look at $SPCX again. Listed on June 12, issue price $135, rose 19% on the first day, and surged to 225.64 a few days later. Then it declined steadily, hitting a low of 104.83 in August, falling below the issue price. Now at 143.50, more than 30% below the peak. And the market was much better back then. On June 12, VIX was 17.68, $SPY was 8% above the 200-day moving average, and the 10-year US Treasury yield was 4.49%. Now VIX is 24.56, $SPY is near the 200-day line, the 10-year yield is 5.07%, and a rate hike in December is almost certain. $SPCX surged and then fell back in a better environment. It’s normal for $ANT to have a short-term emotional surge, and the likely outcome is the same: a sharp fall from the peak to a relatively reasonable price. It’s much more cost-effective to short after the surge than to chase it now.
美股投资young
美股投资young
$ANT is said to be the most hardline AI company in the US attitude towards China, not a label given by outsiders, but one it wrote out itself line by line. In September 2025, it changed its terms of service: Any company directly or indirectly controlled more than 50% by Chinese enterprises, regardless of which country it is located in, is strictly prohibited from using Claude. Mainland China and Hong Kong were never on its service area list, and still are not. Founder Dario Amodei publicly supports tightening chip exports to China, drawing a line between "democratic countries and authoritarian countries," saying the most critical technologies must stay on the US side. In February this year, it publicly named DeepSeek, Moonshot, MiniMax, saying these companies used about 24,000 violating accounts, chatted with Claude over 16 million times, and used it to distill their own models. The September report escalated: saying activities related to $BABA were the largest scale it had seen, with over 151 million conversations from May to July. These are all its unilateral accusations, not court rulings. Why does it dare to claim it is "the safest"? It relies on a framework it built itself: using a set of written principles to constrain the model's responses; the stronger the model, the higher the safety level it must pass before going online; plus its status as a public benefit corporation and a trust that can elect the majority of the board. These are its own positioning, with no third party certifying it as safer than others. So why do Chinese users still use it after accounts are banned? Because programming is really useful. Claude Code has a great reputation in the developer community; its ability to write code and run intelligent agent tasks is its strongest asset. It even said that over 90% of the company's internal code in September was written by Claude. Many people go out of their way to use it; when one batch is banned, another batch comes. Is it afraid of losing this part of the business? Not really. China was never its official market, so there is no revenue from this on the books. US customers account for about 70% of revenue, the remaining 30% from other countries. It can't be said to rely on the government either. Government contracts account for about 1% of revenue. In 2025, it received up to $200 million contracts from the US Department of Defense, but this year, because it refused to use the model for fully autonomous weapons and large-scale domestic surveillance, it was listed as a supply chain risk and even sued over it. Banning Chinese users is mainly to prevent distillation and protect the model, not to spite any market.
美股投资young
美股投资young
The Federal Reserve's October meeting minutes have been released: almost all members believe that further rate hikes are "very likely appropriate." The meeting from October 27 to 28 kept the rate steady at 3.75% to 4.00%, with a vote of 10 to 2; Cook and Harker wanted to raise by 25 basis points. The minutes also mention one or two members saying that if inflation data is unfavorable, the next hike could exceed 25 basis points. Several others felt that holding steady in October was already a very close decision. On the same day, Waller spoke more directly on CNBC: a 50 basis point hike in December is very likely, and he also mentioned the possibility of needing to shake the market. This statement is not in the minutes; it is his personal judgment. But since he is a board member, his words carry significant weight. The market reaction was somewhat counterintuitive. The 10-year US Treasury yield did not rise but fell, closing at 5.066%, down about 4 basis points; the 30-year closed at 5.412%, down 5 basis points. Because 25 basis points were already priced in. FedWatch this morning gave about a 90% probability of a 25 basis point hike in December, and about 10% for a 50 basis point hike. When the minutes first came out, the probability for a 50 basis point hike was only about 7%. The remaining 10% is the part not yet priced in. The meeting will be held from December 8 to 9, with a batch of data in between; the first is today's initial jobless claims. The minutes cover the October meeting; for December, new data will be considered.
美股投资young
美股投资young
Pre-market futures are all rising, but the gains are very uneven. Nasdaq futures +2.3% S&P futures +1.3% Russell futures +0.8% The Nasdaq's gain is about three times that of the Russell. The ETFs are the same: $QQQ +2.3% $SPY +1.3%, back above 733 $IWM +0.7% $RSP +0.6% $DIA +0.7%, just touched the 200-day moving average Yesterday's close was already split: Nasdaq up, S&P and Dow down, with the S&P down for five consecutive days. $IWM is still below the 200-day moving average. The pre-market structure today is still the same, tech is pulling up, small caps and equal weight just moved along. VIX is around 22 this morning, lower than yesterday's close of 24.56, but the market hasn't closed yet. The rebound driven only by the Nasdaq has already been tested once yesterday: it rose above 733 intraday but fell back by the close. Whether small caps and equal weight follow this time will be clear after the market opens.
美股投资young
美股投资young
On 11/18 in the US stock market, the Nasdaq closed higher while the Dow dropped over 500 points. The Nasdaq rose 0.42%, $QQQ up 0.39%. The S&P fell 0.23%, marking its fifth consecutive day of decline. The Dow dropped 1.11%, about 561 points; $DIA closed at 501.73, breaking below its 200-day moving average. On the same day, the three indices moved in three different directions. Equal-weight and small caps also did not follow tech. $RSP fell 0.15%. $IWM rose 0.40%, but still remained below its 200-day moving average. Breaking down by sector makes it clearer: $XRT rose 2.28%, $XLY up 1.79%, retail and discretionary consumption were the strongest. $XLV fell 2.20%, $XLE dropped 3.27%, $XLF down 1.14%, and $XLP also fell 0.31%. $GLD dropped 3.20% in one day. The main drag on the Dow was healthcare; insurance stocks fell double digits in one day, not tech. The 10-year US Treasury yield dropped 4 basis points, closing at 5.066%. The rate decline did not lift financials and energy. Money is shifting between sectors, not flowing in together, nor running out together. VIX closed at 24.56, with pre-market quotes returning to around 22 this morning. Pre-market, $QQQ rose 2.1%, $DIA up 0.7%, just touching the 200-day moving average; $IWM and $RSP only up 0.6%, $XLV still falling. Yesterday’s close was split, and pre-market today remains split.
美股投资young
美股投资young
$ANT is Anthropic, the company behind Claude. Founded in 2021, the founders are siblings Dario Amodei and Daniela Amodei, who came from OpenAI. There are three main products: the Claude conversational app, the programming tool Claude Code, and workflow solutions for enterprises. How does it make money? The main revenue comes from usage-based fees: enterprises and developers pay based on how much they call the model via API. In the first three quarters, this accounted for about $26.3 billion, or 79%. The remaining 21% comes from subscriptions—Pro, Max, and enterprise packages—about $6.9 billion. About 44% of revenue is sold through cloud platforms of $AMZN, $GOOGL, and $MSFT. The biggest money burner is Claude Code and intelligent agent tasks, which involve massive calls, driving up customer spending. Growth is staggering. Revenue in the first three quarters was $33.24 billion, compared to only $2.55 billion in the same period last year, a roughly 12-fold increase. The company says annualized revenue as of the end of October is about $81 billion, but this is an annualized figure based on recent usage. Profitability needs to be viewed separately. Just counting the compute used to serve customers, the gross margin is about 85%. But total compute and infrastructure expenses are $18.8 billion, including training and free usage. Operating loss is $1.26 billion. The net loss on the financial report is $104 billion, which looks scary, but $94.3 billion of that is fair value changes on convertible bonds, an accounting figure, not actual cash spent. Operating cash flow in the first three quarters was positive $6.57 billion. Cash on hand is $95.6 billion, with an additional $36.3 billion in short-term investments. The real burden is the already signed compute bills. Various compute and infrastructure commitments total about $895 billion, with about $69.8 billion due in the next 12 months. Customers are very concentrated. The top five customers account for about a quarter of revenue. There are over 1,900 customers spending more than $1 million annually. Employees number about 5,300. $AMZN and $GOOGL are both shareholders and its cloud and compute providers. This IPO was accompanied by a $26 billion private placement, with participation from $GOOGL, $NVDA, $AMD, $MSFT, $AVGO, $MU, and $TSM. Control is held by the founders. Founders hold about 61% of voting rights through special shares, and a long-term interest trust can elect the majority of the board. Ordinary investors get Class A shares with little say. Many say this is the "world's second largest IPO," but by fundraising amount, it is actually the largest. Last night’s pricing raised $81.3 billion, more than $SPCX’s $75 billion in June, and more than three times Saudi Aramco’s $25.6 billion in 2019. Only when including $SPCX’s later exercise of the overallotment, totaling about $85.7 billion, does it rank second.