每周編輯精選 Weekly Editor's Picks（0620-0626）

郝方舟

Odaily资深作者

@OdailyChina

2026-06-27 02:23

本文約3619字，閱讀全文需要約6分鐘

優質深度分析文章及一週熱點惡補。

AI總結

展開

核心觀點：本週精選文章深入剖析了日圓貶值困境、比特幣ETF套利驅動、SpaceX股價承壓、美光半導體產業轉型及預測市場交易本質，揭示了當前宏觀、投資與Web3領域的核心矛盾與結構性變化。
關鍵要素：
1. 日本央行面臨「加息則財政崩潰，不加息則通膨惡化」的兩難，干預匯率效果有限，市場不信其能繼續加息。
2. 比特幣ETF資金流主要受隱藏的套利交易驅動，而非長期信仰，該交易正持續離場且已持續兩年。
3. SpaceX股價下跌主因是AI概念股回調與散戶買盤枯竭，其太空敘事仍需更多階段性證明以驗證資本開支。
4. 美光已簽署16份長期戰略協議覆蓋20% DRAM與三分之一的NAND出貨量，行業正從「先擴產再找需求」轉向「先鎖單再擴產」。
5. 預測市場Polymarket套利機會有限，核心依靠資訊差與倉位分散，需警惕文字陷阱與尾部風險。
6. 以太坊有機會成為中立基礎設施，因為巨頭拒絕在對手地盤建設，中立層成為唯一選擇。

The information flow moves too fast, making it easy for in-depth analytical pieces to be drowned out by hot topics. The "Weekly Editor's Pick" column sifts through the vast sea of information to bring you valuable content, filtering out the noise, retaining insights, and providing inspiration.

Macro Landscape

Yen Nears 40-Year Low: BOJ Has Raised Rates to 1%, So Why Can't It Stop the Fall?

This time, both the defenses of rate hikes and intervention have simultaneously failed.

There is a growing consensus within the Bank of Japan that 1% is not enough. But the market doesn’t believe it will dare to keep raising rates. "Not changing domestic monetary policy while intervening in the exchange rate is like stepping on the brake while keeping your right foot firmly on the accelerator—best case, passengers get jostled a bit; worst case, the brake pads burn out." —— which is exactly the current situation.

The BOJ is in a dilemma: without raising rates, the yen continues to fall, worsening imported inflation; raising rates sends government debt interest payments soaring, making fiscal policy unsustainable.

The three variables that will determine the future direction: whether the Fed will actually raise rates again, the actual outcome of US-Iran negotiations, and the BOJ's stance at its July meeting.

Investment & Entrepreneurship

Rented Conviction: How Much of the Bitcoin ETF Flow is Real Money?

Week by week, ETF flows are primarily driven by a hidden arbitrage trade, not conviction. Weekly Bitcoin price movements can barely explain the capital flows at all. ETF flows overestimate the "volatility" of conviction, not its "level." This trade is now exiting and has been ongoing for two years.

SPCX Drops Below $150 IPO Price Pre-Market, But Don't Rush to Buy the Dip

The $20 billion bond financing was just the trigger for the decline; SpaceX's share price is also under pressure from the broader sell-off in global AI concept stocks. This is compounded by exhausted retail buying and the negative impact of equity unlock events being priced in early.

Why is NVIDIA's Bond Issue Fine, But SpaceX's Leads to a Plunge?

The market is reassessing SpaceX's future capital expenditure pressures. In contrast, NVIDIA's concurrent bond issue saw strong demand, providing a counter-reference: the AI narrative has entered a phase of revenue and profit validation, whereas SpaceX's space narrative still requires more tangible proofs of concept.

Web3 & AI

Super Spiral Explosion: Micron's Earnings Rekindle the Semiconductor Long Bull Market

Compared to the market's repeated discussions of HBM over the past year, the more noteworthy aspect of this earnings report is that the impact of AI is starting to spread across the entire memory supply chain. Looking at the business structure, almost all of Micron's core business segments are growing simultaneously.

Micron has now signed 16 long-term Strategic Customer Agreements (SCAs) covering data center, consumer electronics, and automotive clients. Most of these agreements have a term of up to 5 years, with some automotive client agreements lasting 3 years, extending coverage as far as the end of 2030. These agreements already cover approximately 20% of DRAM shipments and about one-third of NAND shipments.

For decades, the industry's operational logic was always "expand capacity first, then wait for demand to absorb it." Now, Micron is gradually shifting to a different model: lock in orders first, then expand capacity.

The current capacity expansion activity looks more like executing already-secured orders, rather than the traditional gamble on cyclical expansion based on demand forecasts.

Institutions Rake in 100x Profits: Has Zhipu AI's Stock Price Peaked?

Both institutions and employees are the big winners in this wealth creation cycle. The density of wealth creation ranks among the very highest in the history of Chinese tech company IPOs.

Zhipu AI's most solid revenue stream is from localized deployments. As long as government and enterprise AI budgets persist, Zhipu's revenue ceiling shouldn't be too low. The shareholder structure also backs this business model. The second wave of sentiment was ignited by GLM-5.2 being rediscovered within English-language tech circles. Dissemination is crucial for Zhipu. Chinese investors see Zhipu AI as a Tsinghua-affiliated, state-backed, government-enterprise-deployed, scarce Hong Kong-listed AI stock. The English-speaking tech community, looking at GLM-5.2, sees a different question: Can it partially replace Claude and GPT? Can it be deployed locally? Is it open-source? Is the cost low enough?

The most certain batch of lock-up expirations on July 8th involves approximately 25.68 million shares from cornerstone investors. It's not a hundred-billion-level shock, but it's enough to change the supply-demand dynamics of the chips.

Zhipu AI now needs to prove two things most urgently. First, can the developer buzz generated by GLM-5.2 translate into real revenue? Second, after July 8th, can the market absorb the newly available sellable chips, transitioning the stock price driver from "low float momentum" to "fundamentals-driven"? If both are achieved, Zhipu's high valuation at least has room for the narrative to continue.

The Endgame for AI is Light: A 10x Stock Industry Chain Map Most People Overlook

When the industry upgrades from 800G to 1.6T, and eventually towards 3.2T, the first to capture excess returns are often not the hottest star companies in the spotlight, but the indispensable suppliers that all giants must rely on, like Corning, Amphenol, and Ciena, as well as the upstream materials and testing segments.

Prediction Markets

From Signal Monitoring to Strategy Copying: How Does PPP Lower the Barrier to Entry for Polymarket Trading?

Exclusive Interview with PPP: The World Cup Ignites Prediction Markets, How to Find 'Replicable Smart Money'?

Whether You Know Soccer or Not, Betting on a Draw is the Best Strategy for This World Cup?

As of June 22nd, out of 40 World Cup group stage matches, 13 ended in a draw. If you bet $1000 on a draw for each match, with a total investment of $40,000, the settlement from the winning matches would be approximately $81,914. After deducting the total investment, the net profit would be about $41,914, a return rate of nearly 105%. Looking purely at trading profits, the most profitable script for the group stage wasn't necessarily a big win by a powerhouse, but the recurring 1-1, 0-0, and 2-2 draws.

Coding's Betting Panel Made Money, But Polymarket Really Isn't a Great Place for 'Arbitrage'

Previously, on-chain arbitrage mostly had clear rules, and price discrepancies could be locked in. But Polymarket is different; it really tests your logical understanding of different shifts in sentiment regarding a bet (this is hard to express accurately in writing). For example, regarding political and economic dynamics related to East Asia, Chinese-speaking users might indeed have some information asymmetry advantage. This is worth exploring, but it doesn't guarantee you will win.

Ultimately, Polymarket doesn't settle based on "your understanding of reality"; it settles according to market rules and specified data sources (UMA's manipulation issues have also been prevalent).

Furthermore, just because you think something is a sure thing in a Chinese language context doesn't mean its definition is the same under English rules. Especially in the rule settings for each bet, there are often some textual pitfalls.

So, based on my actual experience, PM doesn't have that many arbitrage opportunities. It mainly relies on information asymmetry and position diversification. Even high-conviction bets can encounter black swans. For Polymarket, don't treat it as a stable income tool, don't equate high win rate with good trades, don't ignore tail risks, and don't do pseudo-diversification. Polymarket is actually an excellent training ground for judgment.

CeFi & DeFi

Wall Street's New Move: US Stock ETF Dividends Automatically Invested in Bitcoin

Franklin Templeton has filed with the US SEC to launch two new Bitcoin DRIP ETFs, which will automatically reinvest stock dividends into Bitcoin.

In the product design, Bitcoin acts as a long-term gain factor within the US stock bubble. A 5% Bitcoin asset allocation also acts as insurance for the portfolio. If the AI bubble bursts and global capital flows back into safe-haven assets, Bitcoin might also see a price increase.

Ethereum & Scaling

Ethereum is Retracing the Path of the Internet and Linux: Everyone Refuses to Submit, and in the End, the Neutral Party Wins It All

Stripe wants everyone to use Tempo, JPMorgan wants to push its own chain, Circle wants to promote Arc – giants will never build on each other's turf. This is precisely Ethereum's opportunity: when everyone refuses to yield to a company's infrastructure, the only choice is a neutral layer that no one controls.

Week in Review: Hot Topics Catch-Up

Policy & Macro Markets

Japan and South Korea stock markets see "Black Tuesday": Korea circuit breaker, Japan's Nikkei plunges, AI boom faces interim correction;

SK Hynix plans to raise over $29 billion via a Nasdaq listing;

SpaceX (SPCX.O) to join the Nasdaq 100 Index on July 7, 2026;

Views & Voices

"White-haired stock guru" Serenity outlines NAV discount opportunities for multiple Asian tech stocks, focusing on targets like Wistron; Missing the AI super-cycle could be costly, photonics, storage, and Neoclouds worth watching;

Bernstein: Raises Micron Technology (MU.O) price target from $510 to $1300;

Analysis: Apple's price hike sounds alarm on AI costs, market reassesses AI stock logic;

Analysis: Rumors of OpenAI IPO delay hit market sentiment, US stock futures fall, tech and chip sectors lead declines;

Adam Back: Strategy won't go to zero, shorting STRC is baseless;

Analysis: ETH is severely oversold, the $1070-$1370 range could be a heavy accumulation zone;