Oli.
Oli.
🍓Web3投研 🍑人工智能 🚀《干翻狗庄》系列工具作者
5Following
257followers
Feed
Livestream
Feed
In Tesla's latest earnings report, there is one less flashy improvement: Services and other businesses achieved a gross profit of about $734 million in Q3.
When everyone discusses automotive gross margins and cash flow, attention easily focuses on new cars. But after the cars are delivered, businesses like charging, insurance, and parts continue to impact the company's revenue.
I think this part is worth following slowly. As the vehicle fleet expands, if related services can be made more efficient, the company has the opportunity to gain more returns from the already established customer relationships.
Of course, this category includes various businesses, so you can't attribute all gross profit to a single item, nor can you directly value it as software business. Different services bear very different personnel, facility, and risk costs.
More importantly, service revenue growth is best built on improved experience. Shareholders certainly want to make money, but if users only feel that repairs are more expensive and fees are higher, the long-term relationship may not improve.
My expectations for this business are quite simple: charging is more convenient, parts and repairs are smoother, and the company can reasonably make a profit.
It may not be as exciting as robots and autonomous driving, but it is already contributing results in the financial statements. Now I prefer to pay more attention to these concrete improvements, at least without having to pin all returns on plans several years from now.
#特斯拉Q3汽车毛利率回升，自由现金流转正
OpenAI's financing valuation has been lowered, and my first reaction is that investors are finally starting to negotiate seriously.
Media reports say the company's new round of financing target valuation has been adjusted from $1.4 trillion to about $1.25 trillion, with a planned financing scale of about $30 billion. What is being discussed now are the targets and negotiation terms, which cannot be directly taken as completed financing results.
Moreover, this does not mean the valuation has dropped below the previous round. Compared to the previous financing level, the new target is still higher. The "lowering" in the headline needs to be understood in comparison to what.
What I think is truly worth noting in this news is that the price the company wants and the price investors are willing to pay are starting to show a more obvious gap.
AI products can continue to improve, and revenue can also grow, but shareholders still need to calculate how much return will remain after continuous investment. Leading technology cannot automatically answer what to do if the price is too high.
The fiercer the financing competition, the more each company needs to prove that their funding arrangements are convincing. Because investors' money can go to other AI companies or choose not to invest for the time being.
I don't think a single price adjustment proves the AI boom is over; on the contrary, I think bargaining is normal. What is more worrying is if the market only accepts valuations going up for a long time and no one is willing to ask about the terms. At least now someone has started to take out the calculator, which may not be a bad thing for those preparing to participate later.
#OpenAI估值下调，AI巨头融资竞争加剧
Meta is preparing to shift more resources from Reality Labs to AI, and I think this is more worth watching than simply increasing the budget.
The company expects Reality Labs' losses to narrow to about $15 billion next year and to reallocate some investments to other key areas, mainly AI. This is the management's plan, but whether it can be realized depends on future results.
I am willing to acknowledge that it has started to adjust resource allocation. If old projects do not deliver expected returns, there is no need to keep spending at the original pace just to stick to past judgments.
But don't automatically assume that shifting to AI means money will be spent more effectively. Even if the budget is redirected, new returns still need to be justified.
Especially since there are differences within AI itself. Investments to improve ad recommendations can show effects relatively quickly through existing business; training stronger models and developing new products may take longer. You can't prove profitability for all AI projects just because ad revenue is growing.
My attitude toward Meta will be more nuanced because of this: which projects have already improved operations, which are still in the investment phase, and it would be best to disclose them separately.
Shareholders can accept the company making mistakes, but they hope adjustments will really be made afterward. What I want to see now is whether the reduced losses in old businesses have been realized and what results the new investments have produced. If in the end both directions just keep getting more money, then the reassurance brought by this resource adjustment won't last long.
#MetaQ3营收超预期，巨额AI投入引发担忧
SpaceX's performance is very good, but the stock price does not necessarily move solely based on profits and merger rumors. When early shareholders can sell also depends on other factors.
The recent phased unlocking has allowed some previously restricted shares to enter the tradable range. Unlocking does not mean these holders will immediately sell all their shares, but the potential supply the market faces has indeed changed.
This is why I don't simply interpret the temporary pressure on the stock price as everyone rejecting the financial report. Some early investors, even if they remain optimistic about the company, may need to raise funds or adjust their portfolios.
Conversely, the buying interest brought by merger speculation may not last forever. If the official plan stalls, short-term trading funds may have much less patience than long-term shareholders.
Currently, this stock is influenced simultaneously by operational performance, trading expectations, and changes in share circulation. Explaining all price fluctuations by just one factor is overly simplistic.
I am interested in SpaceX's business progress but will consider the company's operations and market supply and demand separately. The former determines how much long-term value can be created, while the latter may significantly affect the price during certain periods.
Especially when the financial report has just been released and merger discussions are heated, people tend to feel that with so many good news, the stock price should rise. However, as more people become able to sell the stock, the market still needs time to find a new transaction price. This process is not necessarily comfortable.
#SpaceX业绩大增，与特斯拉合并能否推进？
ETF single-day outflows exceeding $600 million are certainly ugly. But focusing only on this number can easily skew the view of the week's fund movements.
According to Farside data, from October 19 to 21, the total net inflow of US BTC spot ETFs was about $1.075 billion; on the 22nd, there was an outflow of about $636 million. The latter withdrew a large portion of the earlier buying, but did not erase all the inflows from the first three days.
This at least indicates that the current situation looks more like funds adjusting back and forth, and it’s not yet justified to conclude that all institutions are exiting based on just one day. As of the query time, some major products had not yet reported on the 23rd, so don’t rush to calculate the full week’s results.
What makes me uneasy about this round of market activity is how quickly buying and selling switch. One day it looks like funds are recovering, and the next day there’s a large redemption, which easily makes people raise their expectations and then immediately doubt themselves.
What’s more worth observing now is whether subsequent redemptions will continue and whether prices will keep weakening accordingly. Continuous outflows and a single concentrated adjustment require different judgments.
Of course, buying in the first three days is not a reason to feel confident holding now. Funds can change at any time; no one promises they will stay once they come in.
I don’t want to rush to call the bottom for the market yet. Let’s wait for the data to be complete and then observe the cumulative changes over a few days. One day’s data is enough to stir emotions but usually not enough to clearly explain the trend.
#BTC现货ETF资金反复，市场买盘能否企稳？
For Robinhood's next earnings report, I want to look at Gold members first, then trading volume.
In Q2, Gold subscription users reached 4.8 million. More than how many contracts were traded, what concerns me more is whether the business can sustain.
Trading activity fluctuates with the market and hot events, but whether members are willing to renew depends on the platform continuously providing useful services. If users are willing to stay even when they don't trade much, Robinhood has a chance to gain more stable customer relationships.
But membership growth also comes with costs. To attract subscriptions, what discounts the platform offers, how much users actually use, and whether related costs increase accordingly all affect how much profit this business ultimately makes.
So you can't just multiply the number of members by the subscription price and call it new profit. Customer acquisition and service expenses must also be considered.
My expectation for Robinhood is that it can gradually get users to keep more of their daily financial activities on the platform. That way, even if a certain trading product cools down, other demands will support the business.
However, becoming a platform that users use long-term is different in difficulty from attracting everyone to open accounts during a hot market period.
When the earnings report is released on October 27, if membership, net deposits, and customer retention all improve, I will be more willing to believe the growth is sustainable. Trading screenshots look good, but on those less lively days, users still willing to pay is what reassures me more.
#Robinhood交易业务增长能否延续？
SK Hynix's Q3 report, what I most want to hear about is the mass production yield of HBM4, not just how many orders were secured.
The company is scheduled to announce its Q3 results on October 28. Whether HBM4 can boost quarterly profits cannot yet be concluded in advance.
New products are priced high and naturally attractive, but how many qualified products can be produced during manufacturing also affects costs. Especially with complex stacking and packaging, after securing orders, stable delivery is required to truly demonstrate profitability.
SK Hynix has previously emphasized the importance of yield, quality, and supply capability. This time, I am more eager to see whether these advantages continue after the new generation product ramps up.
If revenue growth is accompanied by cost control, that growth will be more convincing; if higher costs are incurred to meet delivery deadlines, profit improvement may be slower than the market expects.
I am optimistic about HBM demand, but it is no longer easy to be directly impressed by the words "next-generation product." Competitors are also advancing technology, and customers will ultimately compare performance, price, and supply stability.
What truly makes me willing to raise my evaluation is the company's ability to repeatedly deliver complex products on time and with quality. Everyone gets excited when parameters are released, but after the financial report comes out, we find out how well the factory is performing.
#SK海力士Q3财报在即，HBM4能否提振盈利？
When oil prices rise, the most easily confused are risk expectations and actual supply losses that have already occurred.
News such as the evacuation of Saudi Aramco headquarters does make the market nervous, but personnel evacuation, facility threats, and actual production declines correspond to different levels of impact. You can't just see a safety measure and casually calculate how many barrels of oil the world has lost.
I think the difficulty in trading oil prices now is precisely because the market must price in risks in advance, while the real-world impact takes time to confirm.
If you wait until supply losses are fully verified to react, prices may have already moved; but if you calculate all threats as the worst-case scenario, once the situation eases, the premium is easily given back.
So this time I want to distinguish two things: the actual affected links and how long recovery takes. Short-term interruptions and long-term damage have very different implications for supply.
Also, high oil prices do not benefit all energy-related companies equally. The pressures of procurement, transportation, and cash flow also fall on other companies in the industry chain.
I remain vigilant about supply risks but do not want to make judgments chasing the scariest headlines. The more frequent the news, the more you have to resist the urge to turn "possible" into "already happened." Otherwise, if you chase the worst-case scenario today and get tossed around by a clarification tomorrow, it’s exhausting.
#中东原油供应风险再起，油价持续走强
Storage manufacturers making more and more money doesn't necessarily mean that phones and computers are suddenly selling much better.
Looking back at Micron's previous earnings materials, there's a very specific phenomenon: revenue growth in mobile and client businesses is mainly driven by price increases, while shipment volume decline actually offset some of the growth.
This detail quite well explains the current market situation. Strong AI demand not only increases data center orders but also affects how manufacturers allocate capacity. After some supply shifts toward high-value businesses, other markets, even with average demand, may face higher purchasing prices.
So my view on storage stocks doesn't simply stop at whether "AI is hot or not." It also depends on whether revenue growth relies on volume or price.
Profit improvement from price increases can happen quickly but also raises customer costs. Customers might reduce purchases, adjust product configurations, or look for other suppliers. A nice price increase doesn't mean demand will never be affected.
I'm still interested in this market cycle but don't want to directly apply current profit margins to several years later. Especially when stock prices already include high expectations, even a slight slowdown in price growth can cause noticeable volatility.
The most exciting thing about the storage business is that when good times come, reports improve particularly fast. I hope the next earnings report will show that beyond price, real shipments and product mix also support growth.
#AI需求支撑存储市场，芯片股行情反复
The most uncomfortable aspect of this PMI is that economic growth has slowed slightly, but companies are raising prices even faster.
The preliminary US PMI for October shows that business activity is still expanding, but the increase in sales prices for goods and services has reached the fastest pace since August 2022. For those waiting for the Federal Reserve to pause rate hikes, this combination is really unfriendly.
Slower growth might have provided a reason to pause rate hikes; but if companies can continue to raise prices, the Fed will find it hard to be at ease. Especially after costs rise, whether companies choose to absorb them or pass them on to customers has completely different implications for future inflation.
What worries me more now is that the market is overly focused on whether the economy is cooling down, while underestimating the possibility that price pressures will continue to transmit.
This will also affect stock assessments. Companies that can raise prices may have profit support; those lacking bargaining power will suffer more when costs rise. The same economic data can lead to very different results in different companies' financial reports.
For BTC, expectations of a pause in rate hikes remain, but the policy space afterward may not necessarily widen because of this. Don’t jump to schedule rate cuts just because growth indicators have fallen a bit.
Before next week’s meeting, I will pay more attention to the price components. It’s certainly better for the economy to hold up than to suddenly contract; but if prices stubbornly refuse to come down, the easy days the market is hoping for may still be a while away.
#10月FOMC临近，美联储暂停加息还有哪些变数？
