Oli.
Oli.
🍓Web3投研 🍑人工智能 🚀《干翻狗庄》系列工具作者
5Following
256followers
Feed
Livestream
Feed
Robinhood's trading volume continues to grow, but I want to break it down first to see exactly which types of customers these trades come from.
The disclosed nominal cryptocurrency trading volume already includes both Robinhood App and Bitstamp business. Interpreting the total volume as if U.S. retail investors suddenly love trading crypto more can easily lead to misjudging the customer structure and overestimating how much revenue the growth can bring.
Different businesses may have different fees, customer types, and trading habits. Even if the transaction volume increases by the same amount, the revenue left in the end may not be the same. The total volume is large and looks lively, but it still needs to be compared with revenue and changes in customer funds.
I do have expectations for Robinhood expanding its business scope. With more products, users have the opportunity to keep more financial activities on the same platform, and revenue sources can also be more diversified.
But I also worry that the market estimates every new type of trade with the same profitability. Especially when the market is active, transaction volume can easily grow; when the heat cools down, which customers remain and which funds have not left are closer to the platform's true strength.
In the next financial report, I will look at trading revenue, net customer inflows, and the performance of different products together.
Of course, trading platforms want everyone to trade more, but shareholders cannot be excited just by transaction numbers. After scaling up, the question of how much money each business actually leaves will sooner or later return to the financial statements.
#Robinhood交易业务增长能否延续？
Anthropic's $80.7 billion net loss over six months cannot be directly understood as burning through $80.7 billion in cash during that period.
According to the IPO prospectus reviewed by the media, about $72 billion of that comes from accounting expenses related to valuation changes of convertible bonds and other financing instruments. Labeling all of it as compute power burn would seriously misrepresent the operational situation; but excluding this part does not immediately lead to the conclusion that the company has no funding pressure.
What I care more about this time is what different numbers in the listing materials are actually answering.
Net loss reflects accounting results, cash flow tells you how money moves in and out, and long-term compute commitments involve how much expenditure must be borne in the future. These items need to be viewed together; you cannot pick the scariest or the most favorable number to judge the entire company.
Especially the expenses caused by the rise in valuation of financing instruments may occur simultaneously with the company's valuation growth. It sounds counterintuitive but does not mean the numbers are wrong. What really needs further verification are operating costs, revenue quality, and future payment arrangements.
I am interested in the progress of Claude products, but that does not exempt me from the difficult questions in the prospectus. A good product and a good stock purchase are separated by a long calculation.
What is least lacking now are opinions like "daring to go public despite such losses" or "it's all just paper losses, no need to worry." I prefer to wait for the official documents and verify the accounts that can be checked first.
#AnthropicIPO草案披露巨额算力承诺
There are changes to the stablecoin reward rules this time, but don't directly interpret it as "issuers can finally pay interest to token holders."
The final rules of the GENIUS Act announced by the OCC still retain restrictions on stablecoin issuers paying interest. Regulatory treatment of third-party platform reward arrangements has been adjusted, but this does not mean that any renamed yield scheme is automatically compliant. The rules are planned to take effect on January 18, 2027, and specific products still depend on the source of funds and implementation methods.
What interests me more is how this will change competition among platforms.
If users hold the same stablecoins, but different platforms offer different rewards, redemption conditions, and user experiences, funds may be reallocated accordingly. Issuance scale is important, but who can retain user balances is becoming increasingly worth watching.
From the user's perspective, you can't just compare the largest annualized number on the page. Where the rewards come from, whether the platform can adjust them, and whether there are withdrawal restrictions all affect the actual outcome.
I am willing to see platforms share more profits with users and hope competition will make stablecoins more practical. However, just because yields look close to deposits does not mean the guarantees are the same as deposits.
This news is worth celebrating, but don't be so happy that you skip over the product terms. Clearer rules can reduce speculation; where exactly the money is placed still needs to be examined case by case.
#GENIUS法案细则发布，稳定币奖励限制调整
Oil prices remain high and hard to come down; now even heating oil reserves are being tapped to find solutions.
The U.S. Department of Energy announced plans to lend out about one-third of nearly one million barrels of emergency heating oil reserves to ease price pressure. This move can help local supply, but it also shows that the issues between crude oil, diesel, and heating oil can no longer be understood by just watching a single international oil price.
Having crude oil doesn’t mean refineries can immediately produce more of the needed refined products; even when refined products are produced, they still need to be transported to places truly facing shortages. Any disruption in this chain can cause the final bills to remain unpleasant.
My attitude toward releasing or lending reserves is that it’s good to get some relief, of course, but don’t take it as proof that supply has returned to normal.
Reserves are meant for emergencies. After they are used, what’s more important is whether subsequent production and transportation can catch up. If normal supply continues to lag, the market will still worry about how long this relief can last.
For investors, this also means that different performances may appear within the energy sector. Crude oil prices and refining profits may not move in sync, and changes in transportation costs may also redistribute earnings.
What’s most frustrating is that negotiation news makes people feel things are about to ease, yet consumers’ bills keep rising. Now I’m more concerned about when normal supply can make up for these temporary measures.
#中东能源风险升温，国际油价维持高位
The most noteworthy change in this round of memory chips is that customers have started to lock in supply by paying cash in advance.
Micron previously disclosed that under long-term supply agreements, customers' financial commitments have increased to about $32 billion, most of which are cash prepayments. This is much more concrete than just saying "AI demand is strong"; at least customers are willing to pay actual costs to secure future deliveries.
For manufacturers, this helps arrange capacity and also alleviates financial pressure during expansion. I am somewhat positive about this change because customers paying in advance carries more weight than merely expressing a long-term purchase intention.
But prepayments should not be directly regarded as profit. The company still needs to deliver according to contracts, and expansion, yield, and product performance must keep pace. The money received today corresponds to work that must be completed in the future.
Competition will also continue. With current supply tightness, customers may prioritize securing goods; when more capacity is released, price, performance, and delivery stability will again influence choices.
Therefore, I prefer to see which companies can turn the current shortage into long-term cooperation, rather than just raising prices while the market is good.
The memory industry’s previous cycles have repeatedly reminded us: when profits are good, expansion is the easiest trap to fall into. This time, with customer cash commitments supporting it, there is indeed an additional basis, but ultimately it still depends on the capacity invested and whether there will be good business to follow in a few years.
#存储芯片股反弹，AI扩产与竞争并存
This time the ETF buying has returned, but BTC has not yet shown a clear strength. I actually think this is more worth studying than simply looking at the inflow amount.
According to Farside's current data, on October 21, the net inflow of US BTC spot ETFs was about $525 million, and the buying has already spread to multiple products. Unlike before when it mainly relied on IBIT contributions, this time the participation range is broader.
The question is, with increased buying, why hasn't the price simultaneously shown a satisfactory performance?
One possible explanation is that there are still enough people willing to sell on the upside. Those who bought earlier are waiting to break even, short-term funds cash out on rebounds, and the new demand is first absorbed by these chips. Which type of sellers dominate still requires more trading and on-chain data confirmation.
I won't say ETF funds are meaningless just because the price hasn't risen. Without this part of the buying, the market might have faced greater pressure. But we also can't assume that a nice net inflow means the rise is just a matter of time.
What is more worth observing now is whether the selling gradually decreases after continuous buying. If buying is maintained and the price slowly rises, then it indicates that supply and demand are starting to change.
Seeing several hundred million dollars flowing in is indeed exciting. But the market never promised that funds coming in would make all holders profit on the same day. Sometimes what it does first is to catch the positions others are eager to exit.
#ETF买盘持续回流，BTC为何仍未走强？
If SpaceX and Tesla really merge, the toughest challenge might be getting less enthusiastic shareholders to vote in favor.
Musk has confirmed that both companies have set up independent special committees of directors to evaluate the potential merger. He also mentioned that Tesla needs to reach a support threshold of about half of all shares, not just a majority of those voting. The final procedure still depends on the formal transaction documents.
This distinction is very important. The strong online support for the merger does not directly translate into enough affirmative votes. Shareholders not participating could also make the deal harder to advance.
So, for now, I won’t take discussion popularity as the probability of completion, nor will I assume the deal is already settled just because the committees have been established. There needs to be a concrete plan next so shareholders know exactly what they are supporting.
I understand the appeal of combining cars, satellite communications, and AI, but I also worry that the more businesses there are, the harder it is for ordinary shareholders to understand why they hold the company.
Some buy Tesla because they believe in cars and autonomous driving; some buy SpaceX because they are optimistic about Starlink and computing power businesses. After the merger, people who originally chose different businesses will have to accept the same asset portfolio together.
When it comes to the voting stage, I hope everyone has sufficiently clear materials, rather than just approving all conditions in one go because they like Musk.
#SpaceX业绩大增，与特斯拉合并能否推进？
A detail easily overlooked in this October rate decision: no new economic forecasts or dot plot.
This means the market loses a direct reference to count "how many more rate hikes" remain, making attention more likely to focus on the wording of the statement and press conference. Even a single word change can be magnified back and forth.
I think this will make "pause in rate hikes" harder to trade.
If the outcome meets expectations, everyone will quickly turn to the next meeting. How much adjustment space the Fed is willing to keep and whether the description of inflation risks changes may influence subsequent pricing more than the rate staying unchanged.
But don’t interpret every sentence as a promise. Officials express their judgment at the time; if data changes later, policy may change accordingly. The most vulnerable are those who hear a definitive answer in one sentence and then hold positions based on that answer.
This round, I want to leave some room for judgment. After the meeting, first see how the market adjusts its expectations for future rates, then see if that adjustment can be sustained.
As the rate decision approaches, information will become denser and the tone more certain. But no dot plot doesn’t mean anyone can draw one on behalf of the Fed. For BTC, don’t buy into months of easing just because you hear the word "pause."
#10月FOMC临近，美联储暂停加息还有哪些变数？
The item that should not be casually dismissed in Meta's latest earnings report is the legal expenses.
Q3 revenue was about $64.6 billion, up 26% year-over-year, but expenses grew even faster, including approximately $9.89 billion in legal settlement-related costs. Excluding these expenses certainly helps to observe daily operations; however, the economic cost ultimately borne by shareholders does not disappear just because the profit metric changes.
I am not pessimistic about Meta's core business profitability. Advertising revenue can still grow, indicating it has the capacity to continue supporting investments. But if legal disputes periodically bring a large bill, it is worth factoring into long-term operational assessments, rather than just calling it "special" every earnings report.
On the other hand, AI infrastructure spending continues. Q3 operating cash flow was about $32.8 billion, capital expenditures and finance lease principal payments were about $34.9 billion, resulting in negative free cash flow.
These figures make me feel that Meta now needs to manage several very costly matters simultaneously. The advertising business must keep growing, new businesses need to prove returns, and existing legal issues must be handled.
I can accept the company investing for the future, and I can accept a quarter with poor cash flow. But I hope shareholders hear the full bill. The stronger the operational capability, the less necessary it is to only highlight the most attractive profit figures.
#MetaQ3营收超预期，法律费用与AI支出承压
What relieved me most about Tesla this time is that it finally generated positive free cash flow again. But this surprise needs to be examined in detail.
In Q3, free cash flow was about $1.5 billion, and the company simultaneously lowered its full-year capital expenditure forecast from about $25 billion to about $20 billion. Operations brought in more cash, and the pace of investment spending also changed, so the improvement cannot be entirely understood as car sales suddenly becoming especially profitable.
Spending less itself is not something to criticize. If projects can be arranged more efficiently, there is no need to spend the entire budget just to prove confidence in the future. Shareholders' money is not for showing determination.
What I care about is whether the reduced investment comes from efficiency improvements or if some projects have been postponed. If delivery times and long-term goals can be maintained, this adjustment is commendable; if it’s just moving today's payments to tomorrow, future funding needs still have to be accounted for.
The rebound in automotive gross margin and positive cash flow at least make the operating situation more comfortable than previously feared. But net profit is still under pressure, reminding everyone that the recovery is not yet complete.
Now I am willing to raise my expectations a bit, but I hope management clarifies the specific arrangements for cutting expenses. Keeping cash on hand is good, but it still needs to be proven that saving this money will not delay the business.
#特斯拉Q3汽车毛利率回升，自由现金流转正
