夏木KRIS
Feed
Feed
Pinned
In the era of information transparency, is it really harder for financial crises to occur?
A major reason why financial crises were prone to brewing in the past was the slow spread of information. Underdeveloped communication and low market transparency mean information is often controlled by a few. Many risks can accumulate within the system for years, yet most market participants remain unaware. When the problem truly emerges, the market often has no time to react, panic spreads rapidly, and it eventually evolves into a systemic crisis. Today's environment is completely different. Smartphones, the internet, and real-time information enable information to be transmitted almost instantly worldwide. Macro data, policy changes, and market price fluctuations can all be captured by the market within seconds. Many local risks have not yet been amplified by the market in advance. From this perspective, traditional financial crises that rely on information lag and accumulate gradually have indeed become harder to form. But things are not that simple. More information does not mean more understanding. When information spreads extremely fast, the market is also more easily surrounded by noise. True and false news are mixed and emotions spread rapidly, often amplifying market reactions. Short-term panic, concentrated sell-offs, and instant liquidity contraction are actually more common in modern markets. The early days of the pandemic in 2020 were a typical example. The U.S. stock market triggered four circuit breakers in just one month. It's not that the economic fundamentals suddenly collapsed within weeks, but rather that the market quickly developed a panic of consolidation under the intense information shock. When everyone reacts in the same way, market volatility is amplified to the extreme. A deeper problem actually stems from the complexity of the financial system itself. 2008 Financial Crisis
$KAITO dropped to $0.327, and I officially bottom-fished.
A few days ago, I was waiting for $KAITO to keep dropping. Today, the price reached around $0.327, so I officially entered to bottom-fish here.
Since the high at the end of July, $KAITO has dropped more than 70% in a short time, and market sentiment has basically been crushed to freezing point. Coupled with the recent token unlocking pressure, there are clearly fewer and fewer willing to buy.
But this position actually starts to match my buying point.
The project itself is still progressing, with Kaito Studio, Capital Launchpad, and staking products all still running. I'm willing to bet on a recovery rally at this price.
Of course, $0.327 may not be the lowest point, so I will still keep some funds reserved.
The first batch at $0.327 has already entered. If there are deeper dips later, I will continue to buy.
$KAITO
Snapshot at 18 Aug 2026, 21:26
Xiaomi is releasing its earnings report today
What I really want to see is not the phones
Xiaomi is about to release its Q2 earnings report today, with the market currently expecting revenue of about ¥117.85 billion. Last quarter, Xiaomi's revenue was only ¥99.1 billion, below market expectations, so there is considerable pressure on this report.
What I personally focus on most is the automotive sector.
In Q2 last year, Xiaomi delivered 81,302 cars, and losses from new businesses like automotive and AI have already narrowed to ¥300 million. The automotive scale is now much larger than back then, so what I most want to see this time is when the automotive business will truly stabilize and become profitable.
Also, on the phone side, we need to see how much the storage price increase has eaten into profits.
So for Xiaomi's earnings report today, I mainly look at three things: automotive profits, phone gross margin, and the upcoming delivery guidance.
$SNDK $XIAOMI $SNDK
#财报观察员：小米即将发布财报，你更看好哪条业务线？
The 30-year US Treasury yield surged to 5.31%
This is a level unseen since the 2007 financial crisis
US Treasuries have started to sell off again recently.
The 30-year Treasury yield recently spiked to 5.31%, reaching the highest point since 2007, and is getting closer to the pre-crisis peak of 5.44%.
What’s particularly troublesome this time is that short-term rates haven’t surged in sync; the pressure is mainly on long-term bonds.
The US fiscal deficit, long-term government bond supply, and AI companies borrowing heavily to build data centers are all competing with the US government for long-term funds. Recently, rising oil prices have also brought back long-term inflation expectations.
So now the market is simultaneously discussing future rate cuts, while the 30-year Treasury yield hits a 19-year high.
This actually indicates that the market’s concern is no longer just about the Federal Reserve, but about what interest rate will be required to justify holding US long-term debt for decades to come.
$BTC $ETH $SOL
#30年期美债收益率创2007年以来新高
Gold has climbed back above $4400, and I continue to be bullish.
Gold has recently pulled back.
On August 17, spot gold rose about 1% to $4420, having touched a high of $4434 a few days ago, and now it has returned above $4400.
What’s interesting about the current environment is that U.S. Treasury yields have surged to a 19-year high, which should put significant pressure on gold, but gold just isn’t falling much.
The reason is simple: tensions in the Middle East are heating up again, Brent crude oil has returned above $90, and safe-haven funds are flowing back.
So I’m still holding gold as before.
I started going long around $4000, and my view hasn’t changed. If $4400 can truly hold and the previous high of $4434 is broken, I’m still looking toward $4700 next.
$XAU $XAUT $CL #黄金站上4430美元，期权资金转向看涨
Snapshot at 18 Aug 2026, 11:31
Recently, I've been waiting for a position in $KAITO. The price has now dropped to around $0.33, compared to the peak of $1.37 at the end of July, a maximum drop of over 70% in less than a month.
I think this is a good entry point.
Of course, there is still a risk: on August 20th, 32.6 million tokens of KAITO will be unlocked, worth about $11.48 million, which is about 7.63% of the circulating supply, so a further short-term drop cannot be ruled out.
But because of the unlocking expectations, whale selling, and the overall weakness in the altcoin market, the price has already dropped significantly in advance.
So I won’t buy all at once; I’ll start building a position around $0.33, and if the price drops further after the unlock, I will continue to buy in batches.
At a level where it has dropped more than 70% and the market has lost interest, I’m actually starting to get interested.
$KAITO
Harvard holds a heavy position in $SPCX
Holding $2.2 billion, directly becoming the largest single holding
After the latest 13F was released, the institutional holdings of $SPCX began to be revealed. One that surprised me was the Harvard University endowment.
As of the end of June, Harvard Management Company held SpaceX stock valued at approximately $2.2 billion, making it the largest single holding in its publicly disclosed US stock portfolio. Harvard's total disclosed US stock assets in this 13F amount to about $4.3 billion, meaning SpaceX alone accounts for a very large proportion.
And Harvard is not the only one disclosed this time.
Alphabet holds about 551.2 million shares of $SPCX, valued at $94.2 billion at the end of June; NVIDIA also holds about 122.8 million shares, valued at approximately $21 billion at that time.
Tiger Global's latest 13F also shows SpaceX holdings for the first time, while reducing some large tech stock positions.
So I will continue to pay attention to $SPCX recently. How institutions allocate after listing is just beginning to be gradually revealed to the market. Harvard directly placed it as the largest holding in its public stock portfolio, so this position is no longer a small test.
#SPCX持股结构曝光，哈佛13F重仓
AMD recently completed the largest US dollar bond issuance in the company's history, raising $4.75 billion in one go.
This issuance is divided into four bonds maturing in 2029, 2031, 2033, and 2036, with interest rates ranging from 4.6% to 5.5%. AMD's official statement only mentioned that the funds will be used for "general corporate purposes," and some may be used to repay existing debt, but no more specific plans have been disclosed.
However, the timing is quite noteworthy.
AMD is not actually short on cash, currently holding about $13.1 billion in cash, while capital expenditures have significantly increased this year, with AI chips, data centers, and supply chains all demanding more funding. Choosing to borrow $4.75 billion at once now seems more like preparing funds ahead as AI investments continue to expand.
Moreover, it's not just AMD raising money recently; the entire AI industry is aggressively seeking capital.
The AI competition has increasingly become a capital war. Anyone wanting to keep up with NVIDIA must not only be able to produce chips but also have deep pockets.
$NVDA $AMD $SNDK
#AMD完成历史最大美元债发行：融资47.5亿美元
The Hormuz agreement still hasn't been signed
Crude oil is now just waiting for the final result
Recently, oil prices have been stuck here, and the key issue remains the Strait of Hormuz.
Negotiations between Iran and Oman have indeed made progress, and they have now discussed specific issues such as channel management, but the final agreement has not yet been officially finalized. The market has already pre-traded part of the expectation that the "strait will return to normal," so oil prices haven't continued to surge recently.
But I think we shouldn't be too relaxed at this point.
The market currently estimates that Brent crude oil may still carry about a 10.7% Hormuz risk premium. If the agreement is really signed, this premium might continue to be squeezed out; but if negotiations suddenly break down, oil prices could easily spike quickly again.
So recently, crude oil is actually easy to understand—everyone is waiting to see whether that final agreement will be signed or not.
$CL $XAU $XAUT #霍尔木兹协议待落地，原油风险等待定价
US stock earnings reports are so strong, why does Wall Street only dare to see around 8000 points?
This round of US stock earnings reports is actually stronger than many expected.
Among the S&P 500 companies that have reported earnings so far, about 85% have exceeded market expectations, and overall profits for the second quarter have grown by about 30%. As a result, JPMorgan has raised its 2026 S&P 500 EPS forecast from $350 to $365.
The problem is, companies are making more and more money, but the index itself has already risen quite high.
JPMorgan recently raised its year-end target from 7800 points to 8000 points, but at that time the S&P 500 was already near 7758 points, meaning Wall Street believes the remaining upside is actually only about 3%. Citibank is a bit more optimistic, currently targeting 8100 points.
So the most interesting thing about the US stock market now is that earnings are still being revised upward, but valuations are no longer cheap.
#标普盈利超预期，华尔街为何仅看7894点
$GOOGL $MU $SPCX
AI Bet Backfires
Jane Street Loses $15 Billion in One Month
This is really outrageous
Jane Street, a top Wall Street proprietary trading firm, lost about $15 billion in July alone, marking the company's first monthly loss in about 10 years.
One major pitfall was the AI fund Situational Awareness managed by former OpenAI researcher Leopold Aschenbrenner.
Jane Street initially invested about $2.5 billion, and this investment once surged to nearly $10 billion, but as the fund's highly leveraged AI trading collapsed, its value ultimately shrank to about $3 to $3.5 billion. Jane Street's other AI and Asian stock positions also suffered losses, resulting in a total monthly loss of $15 billion.
What’s even more shocking is that Jane Street’s revenue for the first seven months of this year had already exceeded $40 billion.
Even a trading firm of this caliber can lose $15 billion in one month during the AI market volatility, showing how extreme this AI-driven fluctuation has become.
#AI押注受挫，华尔街交易巨头月亏150亿美元 $SNDK $NVDA $AAPL