区块博士

Before the U.S. market opened, Micron, SanDisk, and Western Digital all fell by about 4%, while Intel fell by 3.2%. In the previous trading session, the U.S. memory sector already dropped by more than 7%. SK hynix ADRs, Micron, and SanDisk continued to hit fresh lows in this round.

On its first day, Changxin Memory (Shanghai) surged 466%, effectively writing competitive expectations directly into global valuations. My take is simple: strong demand doesn’t mean the stock price won’t fall. Once a new variable appears on the supply side, the most expensive expectations get booked first.

#U.S. memory stocks further expand losses

Chip selloffs are spreading across markets: SK hynix ultimately fell 14.65%, Japan’s Kioxia dropped 18%, and Nvidia fell 5% in the prior trading session. On the same AI trading chain, shares are sold from New York to Seoul, then from Seoul to Tokyo.

The market is worried about two things: whether financing for AI infrastructure can be sustained, and whether China’s competition in memory chips is accelerating. Orders may not disappear within a day, but valuations can be replaced with an entirely new set of algorithms within a day.

#SK hynix falls 11% chip stocks see a selloff

Intel Q2 revenue reached $16.13 billion, beating the market’s estimate of $14.33 billion; adjusted EPS was $0.42, exactly double the expected $0.21. After the earnings release, the stock rose by about 9.9% at one point during after-hours trading.

But the next day, the price closed down 7.89% to $92.32. The numbers won, but the position didn’t. What I care about most is this contrast: even when results beat expectations, it couldn’t hold onto buy-side demand—what the market is trading isn’t this earnings report anymore.

#Intel after-hours up 9%

SpaceX fell from $225.64 to $113.50 during intraday trading on June 16, a pullback of nearly 49.7%. Its market cap has evaporated by more than $1.2 trillion from its peak. At the high point, the total market value was about $2.66 trillion—valuation gravity has finally kicked in.

Next, there are up to 911.5 million shares of restricted stock that can be gradually sold. The company and its rockets haven’t suddenly slowed down— the stock is simply starting to account for supply.

#SpaceX market cap evaporates $1.2 trillion

KOSPI fell 10.84%, SK Hynix fell 14.65%, and Samsung Electronics fell 13.39%. This is no longer a normal pullback—during trading, the exchange triggered a circuit breaker. Even after trading resumed, selling pressure continues to expand.

Bears will say AI valuations should cool off; bulls will say HBM orders have not disappeared. Both sides can be true, but until the liquidation chain is over, fundamentals can only line up. Friends, look first at leverage being cleared before talking about what’s cheap.

#SK Hynix down 13%, Korean stocks KOSPI down 10%

From July 1 to July 27, Korean retail investors net bought US stocks worth $3.59 billion, exceeding 5 trillion won—5.5 times the entire month of June. Just the 3x bullish leveraged semiconductor ETF SOXL alone pulled in $1.759 billion, nearly half of the total.

This isn’t de-risking; it’s just moving leverage from Seoul to New York. I think the most dangerous misjudgment is mistaking a market switch for a strategy switch.

#Korean retail investors net bought US stocks worth over 5 trillion won

South Korea revised its rules on April 28, and starting May 27, allows single-stock 2x leveraged ETFs for Samsung Electronics and SK hynix. Two months later, lawmaker Kim Eun-hye has already been exploring pathways for state compensation for investor losses.

I don’t think investment losses automatically mean the state should foot the bill. But regulators first opened the leverage gate, and then faced circuit breakers and liquidations—this account can’t be settled by having retail investors sign off on it alone. Whether compensation is paid is one thing, but the policy process must be thoroughly investigated.

#Korean lawmaker seeks state compensation for losses on leveraged ETFs

The 60-day correlation coefficient between the KOSPI and the Nasdaq-100 has risen to about 0.50, the highest since 2021. You think you’re diversifying by buying both Korean and U.S. markets, but in the end, both markets are betting on AI chips.

In early July, the weights of Samsung Electronics and SK Hynix in the KOSPI have risen to about 28.07% and 25.56%, respectively. Geographic diversification remains, but factor diversification is gone. If correlations continue to rise, stop-losses across markets could be triggered at the same time.

#Korean stocks and Nasdaq correlation hits a new high since 2021

Brent falls below $88, with WTI near $82. In the prior trading session, the Brent October contract closed down 6.3% to $85.87, as geopolitical premium was rapidly squeezed within 24 hours.

Longs are watching for supply risks that have not disappeared, while shorts are trading expectations tied to U.S.-Iran talks. For me, there is only one variable: whether actual shipments resume. The headline news is not shipping oil.

#CrudeOil continues its downtrend

Asian session's low hit 63,065 USD, then returned to around 63,500 USD. The bounce is about 435 USD, which does not mean the risk is cleared. The same day still fell by about 3.1%, and from July 23 to 24, U.S. spot ETFs saw total net outflows of more than $465 million.

What I care more about is that BTC is currently tracking tech stocks' risk appetite. If the Fed’s outlook remains more hawkish, the 63,000 USD level will likely be tested repeatedly; if ETFs see a renewed net inflow, then the rebound will have a second layer of support.

#比特币自亚洲盘低点回升 $BTC

Many people understand BTCFi as moving BTC to another chain, but Babylon Trustless Bitcoin Vaults (TBV) takes a different approach: BTC stays locked in Bitcoin’s Taproot scripts, verification of the collateral status happens on the Ethereum side, and then it integrates with Aave v4 to borrow stablecoins. Babylon’s official documentation is very clear: each vault is a separate UTXO, not a shared pool of funds, and it can’t be re-staked or used for anything else.

What I find most interesting about TBV is that it breaks down “cross-chain trust” into verifiable scripts, transactions, and proofs, while users still participate in their own BTC release path. The costs should be made clear up front too: lending has liquidation risk, redemption has time locks, and in the current public environment this is still limited to signet and the Sepolia testnet—test assets have no monetary value. Friends, understand the trust assumptions first, then decide whether to participate. @BabylonLabs_io $BABY #baby

Many people talk about BTCFi, and their first instinct is still to wrap BTC into a token, then hand it over to a bridge or a custodian. Babylon Trustless Bitcoin Vaults (TBV) takes a different route: BTC remains on the Bitcoin network, locked by Taproot scripts that are created with user participation. On the Ethereum side, the collateral status is verified through cryptographic proofs. The first public testnet integration is Aave v4. To me, the value of TBV isn’t about turning BTC into another asset—it’s about putting native BTC self-custody and DeFi lending/borrowing into the same set of verifiable processes. This is still the testnet; the funds have no monetary value, but the risks must still be clearly understood. Friends, do you value no bridging more, or liquidity efficiency? @BabylonLabs_io $BABY #baby

CLARITY latest draft adds new self-custody protections: digital assets cannot be deemed lost, abandoned, unclaimed, or subject to state government seizure solely because they are inactive on-chain or have entered dormancy.

According to the draft content relayed via Binance Square, federal rules take precedence over state-level unclaimed property laws.

This rule protects asset ownership—not price fluctuations. If you self-custody your wallet, you also bear the responsibility.

#CLARITY draft prohibits treating self-custodied digital assets as dormant

U.S. stock index futures rose on July 27. Nasdaq 100 futures gained more than 1%, while S&P 500 index futures rose 0.8%.

This week has big-company earnings, as well as an FOMC meeting from July 28 to 29. The market first trades the decline in oil prices to cool inflation, then waits for earnings and interest rates to set the direction.

I’ll put technology earnings from companies like Microsoft and Meta, along with the Federal Reserve’s language, into the same table to look at them together—just looking at futures price gains isn’t enough.

#U.S. Stock Index Futures Climb on Major Earnings and the Fed Meeting

Changxin Technology’s stock opened at 49.50 yuan on its first day after listing, up 471.59% from its issue price of 8.66 yuan, raising about 57.919 billion yuan.

These figures have pushed sentiment for A-share new listings to a very high level, while also lifting the pressure on the float, turnover rate, and valuation.

The percentage gain is the result, not the answer. What truly needs to be tested is whether DRAM market conditions can validate the market’s imagination of a valuation on the order of 3 trillion yuan.

#Changxin Storage’s shares surged 472% on its first day of listing

South Korea’s KOSPI turns lower during the session, with chip stocks once again becoming a drag. Samsung Electronics and SK Hynix together account for more than half of the index’s weight, naturally amplifying semiconductor sentiment.

According to Reuters’ earlier records, on July 8 Samsung fell 6.3% and SK Hynix dropped 5.7%, prompting the market to begin asking whether AI-related capital spending can remain sustainable.

I will be watching chip-stock trading activity and earnings expectations. It’s not hard for the index to fall—the challenge is figuring out when the bellwethers will stop downward revisions.

#South Korea KOSPI turns lower; chip stocks drag

Spot gold prices rose more than 1% during the day, with Reuters data at one point showing $4,103.99 per ounce.

The pullback in oil prices eased inflation concerns, and a weaker US dollar added further support to gold. At the same time, investors are waiting for the FOMC meeting on July 28–29.

Gold’s rise does not necessarily mean broad market risk aversion; for now, it looks more like a bet on cooling oil prices and a repricing of the interest-rate path.

#Gold price rise

WTI crude oil futures fell as much as nearly 8% intraday. At 14:17 on China’s First Finance (First Financial), the quote was recorded at $83.86 per barrel, down 6.1%.

In the same selloff, the maximum intraday decline and the figures shown in media clips differ—this is a normal time lag in the futures market.

As oil prices retreated from around $100 back into the $80 range, inflation expectations may loosen first, but earnings forecasts for energy stocks will also be pressured at the same time.

#WTI crude oil futures down 8%

Changxin Technology’s STAR Market IPO is expected to raise RMB 57.919 billion, with an issue price of RMB 8.66 per share, and an initial issuance of approximately 6.688 billion shares.

This money will go into DRAM capacity, technology upgrades, and R&D. The numbers are huge, but the storage industry’s cycle moves fast as well.

What I care about more is whether the capacity can be delivered after the fund-raising—this isn’t just today’s fundraising news. Credit being received is only the starting point; shipments and profits are the acceptance criteria.

#Changxin Storage STAR Market IPO Fundraising RMB 57.9 billion

Since 2026, 10 S&P 500 component stocks have already fallen by more than 40%. Intuit’s drawdown is about 55.27%. AI concerns are shifting from valuation discussions to mark-to-market P&L.

According to a KuCoin summary, the market isn’t worried about whether AI is useful—it’s worried whether heavy investments can be turned into cash flows in time.

Friends, the next round of divergence will be more direct: companies with revenue validation will stay, while story-only companies will pay tuition first.

#AI Concerns Cause 10 S&P 500 Component Stocks to Drop More Than 40%