Oli.
Oli.
🍓Web3投研 🍑人工智能 🚀《干翻狗庄》系列工具作者
5Following
279followers
Feed
Livestream
Feed
Among SoftBank's losses this time, what concerns me most is the equipment prepayments already made that cannot be easily recovered.
The company posted a net loss of about ¥1.12 trillion from July to September, including approximately $3.7 billion in losses from adjustments to the U.S. data center project, involving arrangements such as locking in equipment and manufacturing capacity in advance. Financing costs also rose to about ¥462 billion.
This makes the risks of AI infrastructure very concrete. When demand grows, people worry about not being able to buy equipment and are willing to pay in advance to lock in capacity; after plans change, commitments originally made to ensure progress may limit flexibility.
SoftBank mentioned that future purchases could receive rebates of up to about $2.1 billion. However, future purchase discounts and cash refunds today obviously involve different cash flow arrangements and cannot be directly considered as losses already recovered.
I am not pessimistic about AI demand just because of one financial report, but the contract terms during the construction cycle do deserve closer scrutiny than before. Whether cancellations or rescheduling are possible and at what cost could all affect the final returns.
Currently, the industry talks most about computing power and power shortages, but the efficiency of capital use should receive the same attention. Securing equipment solves one problem, but whether the project can be put into production as planned and whether there will be sufficient demand after production still need to be fulfilled. The larger the scale, the more obvious the cost of a wrong early judgment becomes.
#软银Q2净亏1.12万亿日元，AI投资收益承压
Anthropic's IPO has finally moved from various valuation rumors to a stage where the numbers can be calculated based on shares.
The company announced on November 10 the start of its roadshow, planning to issue 825 million Class A common shares, with an expected offering price of $70 to $78 per share. Simply calculating within this range, the total issuance amount is approximately $57.8 billion to $64.4 billion, before deducting related fees and excluding the overallotment option.
I think having formal terms is more useful than continuing to debate whether it is worth $1 trillion or $2 trillion. Investors can now judge what they are actually buying by combining the equity structure and the use of funds.
But the price range is only the starting point of the roadshow. The final offering price, subscription demand, and trading performance after listing still need to be awaited. Popular companies entering the public market are not automatically exempt from price constraints.
Especially since current interest rate expectations are still changing, future profits need to be discounted to today's value. When investors' required rate of return changes, valuations may be significantly adjusted. This part does not always synchronize with how good the product is.
I still recognize the commercial demand brought by AI and hope this roadshow can explain growth, investment, and long-term profits more concretely. The company needs sufficient funds to continue building, and subscribers need reasonable returns. Only when both sides accept the same set of conditions can pricing be truly completed. A loud listing valuation cannot replace this step.
#AnthropicIPO进展：路演在即，估值预期存分歧
NVIDIA's market value shrank by about $190 billion in a single day, which looks shocking. But for holders, a more practical question is: how much more risk discount needs to be added to the valuation they were previously willing to pay.
This approximately $190 billion market value change corresponds to the drop on November 9. A decrease in market value does not mean the company’s accounts lost the same amount of cash, but it indicates that the marginal trading price has changed, and all holdings must be revalued at the new price.
The board investigation is not yet complete, and related allegations cannot be prematurely taken as conclusions. What can be confirmed now is that investors face an uncertainty that is difficult to incorporate into income models.
Chip orders and profits can be forecasted, but how long the investigation will last and what subsequent impacts it will bring are currently hard to calculate accurately. Some choose to wait, some reduce their positions first, and others believe the price has already reflected enough. The divergence itself may increase volatility.
I don’t want to rely on guessing the storyline to judge this market. What’s more useful is to distinguish between disclosed facts, issues still under investigation, and how much information gap one is willing to bear.
The company’s business is strong but may experience valuation contraction. A large short-term drop does not prove that the risk has been fully cleared. Upcoming earnings reports and investigation updates need to be tracked separately; any good news on either side may not fully answer the questions on the other.
#英伟达调查风波持续，市值单日缩水逾1900亿美元
New proposals have emerged again in the US-Iran talks. I will first confirm: what exactly have both sides agreed on, or are mediators still conveying opinions?
On November 10, a senior Iranian security official stated that there are currently no negotiations with the United States. After the US engaged with Qatar, it still has not accepted the existing Iranian ceasefire proposal. At least based on these public statements, there is still a distance from a mutually recognized arrangement.
This does not mean diplomacy is ineffective. Contacts may continue through mediators, but circulating proposals, indirect communication, and formally reaching an agreement should not be considered the same progress.
However, crude oil prices tend to react in advance. As long as the market believes the probability of large-scale conflict is decreasing, it may first discount some risk premium. Prices leading execution is normal and also means subsequent actual actions are needed to support it.
What I want to see more is whether there is a unified stance on passage conditions, how the blockade will be adjusted, and when related measures will begin to be implemented. Each of these affects supply restoration and cannot be inferred solely from a statement like "talks are going well."
This round of fluctuating news is already testing patience. Just as judgments were made based on peace expectations, new statements cause hesitation. My attitude is to acknowledge the opportunity for easing while leaving room for execution difficulties. Before an agreement is finalized, the most expensive assumption in trading is often "this time it will definitely succeed."
#美伊和谈传出新方案，霍尔木兹通航能否恢复？
The S&P 500's strongest rise in seven months happened on November 9. The next day, with CPI exceeding expectations, the market had to re-digest interest rate pressures. Watching these two days together is more interesting than looking at just one day.
Monday's rise was supported by factors like falling oil prices and tech stock recovery; the Nasdaq rose about 2.74%, and the S&P about 2.35%. But the improved risk sentiment did not preemptively solve the inflation problem; the data released the next day still changed capital calculations.
This is also why I have become more cautious about big rally days. A single bullish candle can show strong buying on that day but hardly proves that earnings and financing conditions will improve over the coming months. Prices may recover quickly, but fundamental changes may not keep pace.
Especially during this period, the market has to handle AI investment returns, geopolitical risks, and interest rates simultaneously. Easing in one area can bring a noticeable rebound; renewed pressure in another can also take back some gains.
I am willing to acknowledge Monday's recovery and will continue to see if it can withstand subsequent news. Real substantial improvement is when the market can still gradually find buyers even after unfavorable data.
Rather than guessing which day the bottom has been reached, I want to know what my judgment really depends on. If holding positions only feels comfortable when oil prices fall, inflation cools, and interest rates ease all at once, then the assumptions might be too full.
#美股三大指数集体上涨，标普500创7个月最大单日涨幅
The most disappointing aspect of this CPI is that although the overall increase slowed compared to last month, core inflation did not come down as the market expected.
October's overall CPI rose 0.4% month-over-month, higher than the expected 0.3%; core rose 0.3%, also above the expected 0.2%. Food prices increased by 0.5%, housing by 0.3%, with the latter contributing more than 30% of the overall monthly increase. Price pressures are still evident in everyday expenses.
I think this is trickier than simply discussing energy prices. Oil prices can quickly fall with diplomatic news, but changes in housing and service prices usually don't happen that fast. Even though oil prices have recently dropped, that won't reverse the price increases that already occurred in October.
Previously strong nonfarm payrolls gave the Federal Reserve room to continue addressing inflation, but now with CPI not signaling the expected easing to the market, December policy expectations naturally become harder to relax.
However, in trading, we still need to distinguish: hotter inflation and how much the Federal Reserve will ultimately raise rates are two different things. Officials' opinions, subsequent data, and financial conditions will all influence decisions; you can't directly translate one reading into a definite outcome.
For BTC and tech stocks, the most painful thing right now is that funding costs are not coming down anytime soon. Long-term growth can still be optimistic, but the room for optimistic assumptions in valuations probably needs to be re-examined. Holdings need to withstand a longer period of high interest rates.
#美国10月整体与核心CPI双超预期
🟢 Oli Daily Brief｜2026.11.11
BTC has fallen below $74,000 again, ETH has retreated to around $2,050, and SOL has further dipped to $83. After the release of the US October CPI, the crypto market continues to face significant pressure. But what is most noteworthy today is not just the continued price decline, but that the market's capital structure is sending several warning signals. The latest full trading day saw a net outflow of $201 million from the US BTC spot ETF and a net outflow of $184.6 million from the ETH spot ETF. Even more notable is that yesterday the BTC ETF temporarily showed a tentative net inflow, but after missing data was filled in, the final result turned into a clear net outflow. Meanwhile, global stablecoin supply has decreased by 4.56% over the past 30 days, Solana's on-chain perpetual contract weekly trading volume increased by 65.49%, but on-chain stablecoin supply continues to decline. Increasing trading activity does not necessarily mean new funds are entering. On the other hand, Visa has added Monad stablecoin settlement options, institutions like Coinbase are involved in regulatory disputes over US stock perpetual contracts, and ETH withdrawals from staking queues have also noticeably increased. These events collectively indicate that industry development, regulatory changes, and market capital flows are moving in different directions. The real question today is: when macro data releases lead to continued price pressure, ETF and stablecoin funds have yet to improve, but derivatives trading volume is clearly expanding, is the market accumulating strength for a rebound, or is it still in a high-volatility defensive phase? 📉 Market: BTC falls
This time when Vitalik talks about ETH as a store of value, I think the most worth discussing is whether fee burning can provide a reliable valuation floor.
His point is very specific: if on-chain activity and ETH price rise and fall together, then when the market weakens, activity and burning may also decrease simultaneously. Using the fees from prosperous times to estimate a long-term floor is not as solid as it seems.
He still supports retaining EIP-1559 and does not advocate canceling burning; the key is that there is no need to push the network toward higher fees and more active intervention just to increase burning in the long term.
I can understand this part. From a user perspective, cheaper transactions are usually an improvement; it shouldn't be considered bad news just because holders want a bit more burning and low fees.
But the positioning of value storage also needs more concrete support. Network security, censorship resistance, and how much financial activity ETH can facilitate all affect whether people are willing to hold long term. Simply changing the narrative name won't automatically support the price.
I am willing to shift focus from how much ETH is burned daily to what long-term demand the network actually accumulates. At the same time, issuance and burning are still worth tracking, but a single metric shouldn't bear the entire valuation burden. This discussion finally touches on harder and more important issues.
#Vitalik重提ETH价值存储定位，手续费销毁还是核心吗？
The 401(k)-related rules have entered the final review stage. What I am more curious about is whether ordinary participants will be able to understand what they have actually bought in the future.
The previous proposal emphasized the process by which managers evaluate investment options, involving factors such as liquidity, valuation, and complexity. The final text is still under review, and specific arrangements are yet to be announced. Even if the rules move forward, retirement products need to clearly explain these issues.
Underlying crypto assets can be traded around the clock, but once packaged into retirement products, participants may face a different set of valuation and subscription/redemption arrangements. When the net asset value is calculated and how transactions are handled during severe market fluctuations are all related to the actual experience.
I have expectations for this direction, but also some concerns. The inflow of long-term funds can promote industry improvements and services; however, if product structures become increasingly complex and users only see the phrase "gain crypto exposure," they may easily overlook what risks they are actually taking on.
With the CLARITY Act stalled and market structure rules still not fully determined, it is even more necessary for products to clearly explain responsibilities and operational boundaries.
What would truly make me feel progress is if a retirement saver without any crypto experience can also understand where the returns come from, where losses might occur, and how to exit. The industry has earned the opportunity to be considered, and going forward, it will depend on clear disclosures and stable services to make participants willing to stay.
#美国加密监管：《CLARITY法案》推进受阻，401(k)规则进入最终审查
In the discussion about Anthropic's valuation, I want to look at one more question: if customers use more, can revenue grow proportionally?
The prospectus reveals that as of the end of September, more than 11,900 customers contributed over $100,000 in revenue in the past 12 months, showing that enterprise adoption is indeed rapid. But the document also mentions that price cuts on the new model and contract discounts offset about $600 million of quarter-over-quarter revenue growth in Q3.
This detail is very important. Increased usage proves product demand, but price changes affect how much of that demand actually turns into revenue. Even if customers rely more on the product, the company may not be able to keep all the efficiency gains for itself.
I think this is closer to what IPO investors really want to calculate than just looking at user numbers: how much future growth comes from new customers, how much from existing customers expanding usage, and how much price concession is needed to achieve that.
Price cuts may also attract more usage and cannot be simply judged as a bad thing. The key is whether the service cost per task can decrease simultaneously, and whether the economic returns left to the company ultimately improve.
I acknowledge Claude's progress in commercial adoption and understand why investors still want to press down the valuation. There is a gap between a good product and what the stock is worth, bridged by costs, competition, and returns. Calculating these clearly before discussing how much one is willing to pay is more meaningful than arguing over a trillion-dollar figure.
#Anthropic仍计划11月IPO，投资者下调估值预期
