福禄寿炒币版

The missiles sounded—and crude oil surged!
Iran fired ballistic missiles at a U.S. military base in Jordan
Houthi forces attacked a Saudi oil tanker in the Red Sea
Meanwhile, the U.S. carried out precise strikes against armed groups in Iraq supported by Iran
The situation in the Middle East has escalated again
The market’s most direct reaction is that crude oil keeps strengthening
In the short term, rising oil prices will once again lift inflation expectations
It will also weigh on expectations for the Fed to cut rates
That’s not good news for growth stocks—especially the AI sector
But what ultimately determines the AI rally
is the next earnings reports from a few technology giants
Let’s look at SK Hynix’s earnings
A few key figures:
Revenue: KRW 79.3 trillion, up 257% year over year, setting a record high
Operating profit: KRW 60.5 trillion, up 557% year over year, also a record high
Net profit: KRW 93.9 trillion, up more than 13 times. The difference of about KRW 33 trillion between operating profit and net profit mainly comes from gains from Kioxia’s (Kioxia) investment.
More worth noting are the signals released by management
AI demand hasn’t changed—AI infrastructure buildout is still ongoing
Long-term orders continue to rise
Capital expenditures keep expanding
Overall, I think this earnings report is still quite solid
So why is SK Hynix’s stock still falling?
The market is no longer trading the results—it’s trading expectations
Over the past few years, SK Hynix has climbed by more than ten times
Institutional investors have already accumulated huge unrealized gains
When the stock price has priced in years of future growth in advance
even if earnings hit new highs
it may not be able to push valuations higher anymore
The recent pullback
looks more like profit-taking and valuation digestion
rather than deterioration in fundamentals
Tomorrow, after the U.S. market close, Microsoft, Meta, and Qualcomm will report earnings
What will truly affect the next phase of the AI sector’s outlook
is not whose profits are higher
but whether a few tech giants are still willing to keep investing tens of billions of dollars to build AI infrastructure
If capital expenditures continue to increase,
it means AI demand is still strong
If they start to contract,
the market will need to reassess this round of the AI rally!

I bought Corning for 220, and it dropped to 120. I’ve experienced the halving of small-cap, fake coins before. This is the first time I’ve experienced a high-quality US stock getting halved too. Luckily, back then it was only a small trial and I only built an initial position. My hunch was really right: I never averaged down. When it fell to the 110 halving level, no matter what, I have to buy another 500,000 RMB worth of spot. The 110 level has already fallen below a $100 billion market cap. If 110 is the halving point and I add another 500,000, the opportunity comes from the drop—controlling position size really is crucial! Thanks to the fact that my spot positions in this year’s Q1 endured a halving and made me engrain proper position sizing into my bones. $GLWB

Tomorrow SK Hynix will release its Q2 earnings report! Q2 performance is highly likely to beat expectations, HBM demand remains very strong, and SK Hynix’s HBM market share is still #1 globally. Will future growth slow down? Past glory doesn’t determine the future stock price—what will SK Hynix’s stock price look like tomorrow? It can’t still fall and trigger a trading halt, right!

Changxin Technology’s listing officially brings China’s DRAM into the global capital market’s pricing system.

On the same day, the Korean KOSPI triggered a circuit breaker during intraday trading; storage stocks such as SK Hynix and Samsung Electronics fell sharply. AI supply-chain stocks in the U.S., including Corning, SanDisk, and Micron, also weakened in tandem.

Many people attribute the cause to Changxin’s listing, but it’s not that simple. Changxin Technology’s listing was merely the fuse.

Changxin currently focuses mainly on DRAM, and in the short term it has not yet developed the capability for large-scale HBM mass production.

As HBM—high-end DRAM with the highest profits and the highest technological barriers in the AI era—SK Hynix is still the global leader today.

SK Hynix’s true core competitive strength has not changed in the short term.

The main reason is that the storage sector has surged too much over the past year and its valuation is too high; once there is even a hint of trouble, profit-taking positions get liquidated all at once.

In addition, the market is re-evaluating the future global DRAM competitive landscape, domestic semiconductor companies continue to break through, and the Federal Reserve has maintained high interest rates for the long term, with expectations of another rate hike still present in September. Liquidity remains tight, and all these factors jointly amplify this wave of selling pressure.

AI is the greatest revolution for humanity. Opportunities are created by declines. Build positions in batches, prepare for investment cycles of five or ten years, and seize the wealth redistribution brought by the AI revolution.

ChangXin Technology (688825) hit a midday high of 55.03 yuan on its first day of trading; at one point, its total market value reached about 3.65 trillion yuan. Note! It’s RMB, not USD—converted, that’s roughly US$54 billion. Some people blurt out “more than US$30 billion”—are they in the wrong scene? ChangXin Technology is a Chinese company listed on the A-share market, so the unit is of course RMB. If it were truly worth more than US$30 billion, then it wouldn’t be the A-share market value king—it would be the undisputed #1 across the entire world. Don’t stay in US stocks too long and even forget to switch back currencies. $CXMT

A new week has begun—and it’s going to be lively!

Iran and the U.S. have exercised restraint to renew a ceasefire, reigniting hopes for negotiations. Brent crude has fallen below 90 as a result, and at least this week, risk markets may be able to breathe a bit easier.

Today, Hefei Changxin Technology listed on the A-share market. Changxin is a leading domestic DRAM player, and one of the largest IPOs in the history of the STAR Market. Its issue price is 8.66 yuan, valuing it at about 580 billion yuan upon listing.

On Wednesday, SK Hynix released its Q2 financial report. I believe the importance of this report is even on par with Nvidia—it’s one of the most important signposts for this round of the AI rally.

On Thursday, the U.S. PCE data: if core PCE month-over-month comes in higher than expected, the market may further price in that high rates will be maintained for longer. U.S. Treasury yields and the U.S. dollar could strengthen, putting pressure on tech stocks, BTC, and gold. If core PCE month-over-month comes in lower than expected, the market will instead trade the improvement in liquidity again—good news for both AI tech stocks and crypto assets.

The PCE tells the market how inflation is doing. Then, on the same day, the Fed’s FOMC interest rate decision is essentially telling you what the Fed is prepared to do.

Meta, Microsoft, Qualcomm, and ARM will all release their 2026 Q2 earnings reports after the U.S. stock market closes on July 29. Along with SK Hynix, they will jointly determine the direction of global AI tech stocks and risk assets for the next quarter.

After this week ends, there will be more data forecasts for the likely trend in Q3 and Q4 risk markets. AI is the future, not a bubble—at least so far, there hasn’t been a bubble yet!

In the early period after the establishment of the Han Dynasty, it went through the chaos that broke out toward the end of the Qin era and the Chu–Han contention.

The state treasury was empty, the people were impoverished, the regional lords were powerful, and the northern Xiongnu continued to raid southward.

Facing internal troubles and external threats, Emperor Wen of Han did not choose to confront them head-on; instead, he continued the policy of heqin (marriage alliances) and recuperation to restore the economy and let the country recover.

He used decades to develop the economy, stabilize internal affairs, and build up national strength.

Only after Emperor Jing of Han completed the handoff did the Former Han truly grow strong.

It was only when Emperor Wu of Han ascended to the throne that there was enough confidence to launch a comprehensive counteroffensive against the Xiongnu, ultimately completely reversing the balance of offense and defense.

America’s biggest challenge now may not be Iran, but rather high debt, high deficits, inflationary pressure, and deep political divisions at home.

At a time like this, if the United States were to fully escalate a war for short-term political gains, it would very likely further strain the finances and the economy.

Trump should bide his time now—stabilize the domestic economy first, push the return of manufacturing, control the fiscal deficit, win the midterm elections, and lay a solid foundation at home.

Once the American economy and political environment are more stable, then concentrate efforts on solving the Iran issue.

In real major-country competition, it’s not about who throws the first punch; it’s about who has a stronger foundation and greater staying power.

Temporary strategic restraint is not weakness—it means waiting for a moment to strike with a higher chance of success and lower costs.

The expectation for the CLARITY bill to pass is still cooling down

The bill is very unlikely to complete a vote before the August recess

Trump’s roughly $1.4 billion in crypto gains have become a new focus of controversy

Democrats believe existing moral constraints are not sufficient

Intense negotiations are still underway over issues such as enforcement authority, indirect-ownership oversight, and the effective period of provisions

The banking industry is also worried that stablecoin yield terms will accelerate deposit outflows, and continues to apply pressure

Not passing this year may not necessarily be a bad thing

The biggest contradiction right now isn’t regulation—it’s liquidity

The Federal Reserve is still keeping interest rates high

Global liquidity has not truly improved

Even if CLARITY passes this time

it’s still hard to kick off a new bull market with a single piece of legislation

Assuming liquidity starts to improve next year

the Federal Reserve enters an interest-rate-cut cycle

CLARITY is officially implemented

Regulatory certainty and liquidity resonating together

Institutional capital’s willingness to step in will be stronger

Creating a powerful boost for the entire crypto market

Will tomorrow’s Korean stock AI industry chain usher in a frenzy of collective celebration?

The signals Huang In-hoon has released these past few days are more important than many earnings reports.

In addition to announcing that NVIDIA and SK Group’s future cooperation will exceed 500 billion USD, and that they have continuously locked in SK hynix’s HBM procurement for years.

Anthropic has also reached long-term partnerships with Samsung Electronics and SK hynix. Global AI giants have almost simultaneously brought Korea’s memory industry into their core supply chains.

Huang also said that over the next decade, the global semiconductor industry’s scale could expand to as much as 10 times today. Real compute demand will no longer come only from humans, but from AI agents, robots, and other intelligent endpoints.

Basic infrastructure such as HBM, advanced packaging, data centers, and power will remain in chronic short supply for the long term.

I also believe AI is not just a typical technology cycle, but the greatest productivity revolution in human history.

In the short term, valuation bubbles may appear, but bubbles will eventually fade—the technology will remain.

When AI evolves from serving humans to serving billions of AI agents and robots, the way society produces will undergo a fundamental transformation.

In the future, what will truly change the world is AI’s full integration into every sector. What AI brings is not just one round of industrial upgrading, but a revolution that moves from quantitative change to qualitative change.

And what we need to do is to seize the opportunities brought by this greatest revolution!

The significance of logging into Futu through BNB is not just that it provides another market entry point—it’s an even deeper step into the asset view of traditional financial investors. As more and more brokerages begin to showcase, research, and even support crypto assets, BNB, as one of the world’s top crypto assets, is also accelerating its integration into mainstream asset allocation systems.

A bigger trend is the two-way convergence of TradFi and Crypto. Through products such as Stocks and bStocks, Binance brings traditional assets into the on-chain ecosystem, enabling users to both hold traditional assets and, according to their needs, convert them into on-chain assets. In the future, a single account will be able to connect to more markets and manage more assets—without the need to constantly switch between brokerages, banks, exchanges, and on-chain wallets.

By connecting the widest range of assets and pooling the most liquidity, Binance’s new-generation financial infrastructure is rapidly taking shape.

South Korea’s storage industry has, almost on the same day, received two separate collaborations from AI giants. On one side, Nvidia is expanding its partnership with SK Group—supporting the construction of 2GW AI data centers and pushing forward long-term cooperation on next-generation HBM. On the other side, Anthropic has reached long-term supply agreements with Samsung Electronics and SK hynix, along with strategic investments. Meanwhile, Intel has raised its Q3 earnings guidance, and Qualcomm has announced a double-digit price increase. The signals being released across the entire industrial chain are becoming increasingly consistent: investment in AI infrastructure continues to accelerate.

In the short term, the capital market is still worried that AI capex is too high, putting pressure on technology stocks. But the answers provided by the industry are completely the opposite—long-term orders keep increasing, companies continue to expand production, and earnings guidance continues to be upgraded. Market sentiment may fluctuate, but real orders won’t lie. AI infrastructure construction is still at the early stage of a long-term cycle.

NVIDIA announced an AI cooperation plan with South Korea’s SK Group with total investment exceeding $500 billion. The two sides will launch long-term collaboration in areas such as AI data centers and next-generation HBM high-bandwidth memory. Among them, SK Telecom will build 2GW AI data centers, using NVIDIA’s Vera Rubin chips and SK hynix’s HBM4. The first project is expected to go into operation in 2027.

In the past, NVIDIA’s purchase of HBM was purely a procurement relationship. This time, by establishing long-term cooperation with SK hynix, both sides are starting to jointly define the next generation of HBM4. As AI training keeps getting larger and GPU performance improves further, it still can’t run without HBM. GPU shortages can be alleviated by expanding production, but HBM requires a longer construction cycle and more complex packaging processes—so HBM is the true bottleneck. What NVIDIA is doing now is securing the most critical components for the years ahead.

What SK hynix gains is certainty. The biggest risk in the AI industry is not technology, but demand. If customers buy today and don’t buy tomorrow, companies won’t dare to expand production. But with NVIDIA directly tying itself to SK hynix, it effectively tells the market in advance: HBM demand in the coming years is real. This gives SK hynix more confidence to increase capital expenditures, further strengthening its leading position in the HBM market.

AI competition has already moved from GPUs to the entire compute ecosystem, and many people are still focused on GPUs. The real AI competition has become: GPU + HBM + advanced packaging + data centers + power + networking. The 2GW AI data center initiative is the best example. NVIDIA is now getting involved in building data centers, chip planning, storage layout, and even ecosystem coordination. In the future, what’s being sold won’t be just a single GPU, but a complete AI factory (AI Factory). This is also the direction Huang Renxun has long emphasized. The entire industrial chain is shifting from single-point breakthroughs to a phase of systematic buildout.

When James, an NBA superstar, joined the Lakers just as he came into the league, he kept using James wearing Lakers jersey as his avatar for eight years. Now James has joined the 76ers. I’ve just changed my avatar to a cartoon-style Sun Wukong the Monkey King fighting Buddha card version that I generated with AI. I wish I can achieve enlightenment like the Great Sage too.

I increasingly feel that among altcoins that can still outperform BTC in the future, only about 5% to 1% might be able to do so. Most altcoins will gradually lose liquidity due to a lack of funding, users, and real value. The only ones with a real chance will be concentrated in a few niches, such as AI Agents, RWA, payments, and truly DeFi that is supported by real income and cash flow.

Even if, in the future, the U.S. Federal Reserve starts cutting rates, it will likely only move from restrictive interest rates back to neutral rates—not something like the limitless QE of 2020. The era when you could make money just by pushing liquidity and mindlessly buying altcoins probably won’t return. In the future, the crypto market may only be able to sustainably attract capital for a tiny number of high-quality projects, and a full-scale altcoin season will be a thing of the past.

Most importantly, global capital has more and more options. AI tech stocks, gold, crude oil, U.S. Treasuries, stablecoin yields—even much crypto capital has started allocating to AI U.S. equities. Capital doesn’t just rotate within the crypto space anymore; it’s looking for yield among global assets.

Trump believes the only language Iran understands is force. He is frustrated that the Iran war has dragged on without resolution, but says there is no better option, so they have to keep fighting. Why not cripple Iran once and for all—long pain is worse than short pain. The trouble was caused by Trump; let the entire risk-asset market carry the pot together.

The escalating situation in the Middle East has already translated into clear market pricing. As U.S. Air Force B-1B strategic bombers strike targets in Iran, the operational scope and intensity have entered a new phase. Oil prices then accelerated higher, and Brent broke through the $100 mark in step.

For the market, inflation expectations will be repriced. Energy prices will directly push up July’s CPI, while by late month the PCE and the Fed’s policy room will tighten in tandem; rate expectations swing back to a more hawkish stance and feed through to the valuations of tech stocks and crypto assets. Going forward, for every $1 that Brent rises, risk assets face higher inflation and tighter liquidity.

As the situation in Iran escalates, the Houthis have resumed attacks on merchant ships, and Brent crude has broken through $100. The rapid rise in oil prices has quickly fueled market concerns about inflation. US 2-year Treasury yields have climbed to around 4.34%, 10-year yields to above 4.7%, and the 30-year yield has even reached 5.19%, all hitting the highest levels for this period. The 2-year yield reflects expectations for Federal Reserve policy; its continued rise indicates that the market has started to reprice the possibility that the Fed may take more hawkish action sooner.

War escalation → oil price increases → inflation rebound → the Fed maintains high interest rates or even hikes → Treasury yields rise → risk-asset valuations fall. For tech stocks, higher rates mean a higher discount rate for future profits; for crypto assets such as BTC, tighter liquidity is also bad news.

If oil prices are only spiking temporarily, the market will quickly return its focus to AI and corporate earnings. But if oil prices stay above $100 for an extended period, even edging closer to $120, then in the coming months the key focus of global market trading will no longer be AI, but inflation and liquidity. The US stock market, BTC, and even risk assets as a whole will face another round of valuation pressure.

Financial reports are the market’s most direct barometer. After the data from Google and Tesla were released, their stock prices fell in tandem, indicating that the market is no longer satisfied with revenue growth and has started asking when AI investments will translate into profits and cash flow.

The next few sessions are even more crucial: on July 30, Meta, Microsoft, and Qualcomm will submit their results in a concentrated release, testing AI applications, cloud computing power, and on-device AI; on July 31, it will be Apple, Amazon, and Coinbase—covering consumer electronics, cloud computing, and the crypto market, respectively; from August 4 to 5, Palantir and AMD will need to answer whether AI software orders can catch up to their valuations, and just how strong demand is for AI chips beyond Nvidia.

Finally, it will be Nvidia in the early hours of August 27, essentially the grand finale of this AI earnings season. Over the following month, what the market will trade won’t just be whether results beat or miss expectations, but which AI path has already started making money—and which company is still propping up its valuation with capital expenditures. It’s going to be fascinating.