Cruise橘子哥

#股票财报季
Google's Q3 market consensus is for revenue of approximately $127–128 billion and EPS of approximately $2.94–$3.02.
I predict another modest beat, with revenue around $128 billion and EPS around $3.05–$3.10. There are three reasons:

First, revenue is relatively stable. Last quarter, revenue was $119.8 billion, up 24.2% year over year. Search revenue grew about 16.8%, suggesting AI has not yet significantly disrupted search advertising. As long as Search continues to grow at a double-digit rate and YouTube and subscriptions maintain their momentum, reaching around $128 billion should not be difficult.

Second, Cloud is the main source of potential upside. Last quarter, Cloud revenue was $24.77 billion, up about 82% year over year, with operating profit of $8.8 billion. If Cloud revenue approaches $28 billion this quarter and its margin holds at around 35%, both revenue and profit could exceed expectations. Conversely, if Cloud comes in below $27 billion, the market may start worrying about AI demand and capacity.

Third, we shouldn't be blindly optimistic about profits. Capital expenditures were $44.9 billion last quarter, and free cash flow was already negative $5.86 billion. This spending won't all be expensed in the current quarter, but it will gradually weigh on profits through depreciation. Fortunately, the market isn't extrapolating last quarter's unusually high EPS of $9.11; instead, it has brought normal quarterly expectations back to around $3. So a modest EPS beat shouldn't be too difficult.

Threshold’s T is at the top of the gainers list today, trading at $0.0055, up 20% in 24 hours. At one point, it was more than 40% above its current price, then gave back those gains all the way down, closing near the bottom of its move. Volume was 169 times the norm, with around $17 million traded. I’ve seen this pattern too many times: buyers rush in on the spike, then sellers take over on the way down. Neither side looks like long-term money.

The project is an old name: it came out of the merger of Keep and NuCypher, and works on Bitcoin cross-chain applications like tBTC and node staking. It’s not a new story, and over the past 30 days, it’s actually down a little.

Token ownership is highly concentrated, with the top 10 addresses holding 75%. On Binance, just 61 KYC-verified addresses hold the token, with an average purchase price of $0.012—more than double the current price. In other words, those holders are still underwater. The on-chain pool is tiny, almost negligible, so most of the volume in this move came from exchanges. It’s hard to tell who’s driving it. What’s clear is a long upper wick and a crowd of holders stuck above the current price.

$MEGA Mega has been bouncing around a lot lately. It feels like it’s about to take off.

It used to be a star, but now it always feels pretty lackluster. The tokens released by the NFT project will have their final tranche unlocked in a few days, which will bring the token releases to an end. Back then, one SBT bought for 1 ETH got you 50,000 Mega, and half was released at launch. Even though the price has fallen a lot, those holders are still way up overall—they haven’t lost money~
If only I hadn’t been so convinced of it back then and had started shorting at launch—that would’ve been so satisfying. These days, as long as a coin isn’t being manipulated, even these no-utility coins can make you a freaking fortune.

$PREDICT

This is great news: Predict has launched on Binance Pre-Market Futures. It’s currently priced at 1.52, with a market cap of around 1.5 billion. In this market, it’s undoubtedly the hottest project right now.

Getting its TGE in before Polymarket, backed by Binance, and the world’s first prediction market integrated into Binance—the rise of Predict as the leading prediction market is now a done deal. Congratulations to Predict, and to all the Predict ambassadors earning at least seven figures a year. What a great time! @predictdotfun-华语

AVA is now at $0.413, up 11% in 24 hours. It’s climbed for a week straight and looks set to retest last month’s high.

Volume hasn’t kept pace. Trading volume is $4 million, only half the usual amount, and the order book is so thin that a few trades can push the price up. It’s hard to tell who’s driving the move, but it doesn’t look like big investors are quietly accumulating.

AVA is the token for Travala, a hotel and flight booking platform. You can earn cashback when you pay for a stay, and the market has always been small. On-chain liquidity is thin, and the top 10 addresses hold 70% of the supply. If someone starts selling, there won’t be many buyers to absorb it.

Binance holders have an average cost basis of $0.60, well above the current price, so they’re still underwater. The contract also retains the ability to mint more tokens—something that shouldn’t be ignored.

ZEC is up 3.4%, trading just above 1,200, with a market cap of 20 billion. While everything else is deep in the red, it’s still in the green and looking pretty resilient. But its daily trading volume is only a little over 4 million. Against a 20-billion market cap, that’s so little volume that at first glance I thought I’d misread the numbers.

Privacy coins have already had a rally this year, and the biggest wave of hype is behind us. The price is edging up mostly because sell orders are scarce. With just over 4 million in volume against a 20-billion market cap, the market is paper-thin.

The iPhone Duo had an awkward first day on sale. It launched on October 23, with deliveries pushed back to late January 2027, while the resale premium on Dewu shrank by nearly half in a week. That same day, photos began circulating on social media of the inner screen cracking along the fold, and Apple’s official repair quote topped nearly 8,000 yuan. People scrambled to buy it, but scalpers were the first to bail. Apple’s stock barely budged that day, and its market cap briefly surpassed Nvidia’s intraday.

$BILL The tokens staked at the end of the month are about to unlock. Could all those passively staked tokens coming out cause a massive sell-off?

I have 400,000 of them. At $2,000, just thinking about it feels pretty nice. I wonder if we’ll get the chance, haha. I haven’t seen any major news lately either.

$PREDICT —who knows how many people playing Predict this time have reached A7, A8, or A9 with a single token?
This must be the best time ever for the ambassadors. I’m so jealous—it's been ages since we’ve seen such a huge wealth effect.

The crypto world really is full of opportunities.

#韩国推迟加密货币所得税
This would be the fourth time South Korea has postponed its crypto income tax. It was originally set to take effect next January. This week, the Blue House said it “respects young people’s paths to building assets” and signaled it was considering a delay, while Park Min-gyu’s amendment goes further, explicitly pushing the start date to 2029.

The tax itself is straightforward: profits from crypto transfers over the course of a year are taxed separately at 22%, after deducting 2.5 million won.

The problem is the glaring contrast with another policy: profits made by retail investors trading stocks have long been tax-free, while crypto gains aren’t exempt by even a penny.
And international tax information exchange won’t come online until 2027. Starting the tax now would mean catching only people whose money is still on domestic exchanges.

So, unusually, both sides of the political aisle are aligned this time: put the tax on hold for now. The rate is secondary. The Digital Asset Basic Act is still stalled in the National Assembly, and there’s no clear definition of what digital assets are, yet they want to send out tax bills first.
The share prices of Upbit and Bithumb’s parent companies have been languishing near their yearly lows. After all these delays, no one can say whether people will actually have to pay next year.

Do you know what else Google is called?

Alpha? Hahaha, didn’t see that coming. Someone say the full name. 😄

#股票财报季
Amazon will report its Q3 earnings after the U.S. market closes on October 29.
The market expects revenue of about $202.1 billion and EPS of around $2.01.
My prediction: revenue will come in slightly above expectations, and EPS could also top $2.01, though profit may be less dependable than revenue. There are three reasons:

First, the revenue outlook is relatively reassuring.
The company’s Q3 guidance is $197 billion to $202 billion. Market expectations are already slightly above the high end of that range, but Amazon has often beaten its own guidance. Q2 revenue reached $200.6 billion, up about 19.6% year over year, and AWS growth accelerated again. As long as the consumer business doesn’t weaken significantly, reaching $202 billion to $204 billion in Q3 shouldn’t be difficult.

Second, AWS is the most important thing to watch this time.
Q2 AWS revenue was about $42.2 billion, up nearly 37% year over year, with operating income of about $16.6 billion. The market expects AWS growth in Q3 could approach 40%. If revenue reaches $46 billion to $47 billion, that should be enough to support a revenue beat. Investors are buying Amazon mainly for AWS and the renewed acceleration in demand for AI cloud services.

Third, be cautious about profit.
Q2 capital expenditures reached $54.2 billion, operating cash flow was $45.4 billion, and free cash flow was negative $8.8 billion. In Q1, free cash flow was even more negative, at $18.2 billion. These AI data-center investments won’t all be expensed in the current quarter, but depreciation is already increasing: Q2 depreciation and amortization was about $20 billion, up from about $16.8 billion in the same period last year. So even if revenue grows quickly, operating income may not grow at the same pace.
The good news is that Q3 operating income last year included about $2.5 billion in FTC settlement costs and about $1.8 billion in severance costs, which won’t recur this time. Last year’s Q3 operating income was only $17.4 billion, while the company’s guidance for this quarter is $22.5 billion to $26.5 billion. Even after accounting for AI investment and depreciation, there’s room for a significant year-over-year improvement in profit.
However, Amazon’s EPS will also be affected by changes in the fair value of investments such as Anthropic, so don’t focus on EPS alone.
So, my final prediction is: revenue of $202.5 billion to $204 billion and EPS of $2.05 to $2.15, a slight beat.

Another thing everyone cares about is the share price. Ultimately, three things will determine where it goes:
Whether AWS can sustain growth of around 40%, whether capital expenditure guidance for Q4 is raised again, and when free cash flow will turn positive again.

This is my personal analysis and is not investment advice.

Threshold’s T closed at 0.00634, up 37% in 24 hours. Trading volume was $12 million—121 times its usual level—with a market cap just over $80 million. That kind of volume in a coin nobody usually talks about really stands out.

It focuses on cross-chain Bitcoin, through the tBTC project, and T is used for staking and governance. It surged during the day, then gave back 20%. Aggressive buying still hasn’t outweighed selling, and overhead sell pressure has persisted during the rally. It looks like someone lit a fire, but there wasn’t enough follow-through.

Token holdings are highly concentrated: the top 10 addresses hold 78%. On-chain liquidity pools contain only a little over $10,000, and almost all trading takes place on exchanges. Among Binance KYC users, only around 60 hold the token. Their average selling price is 0.0124, nearly twice the current price, while those who bought near the earlier highs are still underwater.

Meta spent $34.9 billion on capital expenditures in the third quarter, and free cash flow turned negative at $2.1 billion—for the first time since its 2012 IPO. Revenue and profit both beat expectations, yet the stock still fell nearly 7% in a single day after the earnings report. To keep building data centers, the company announced it would issue non-voting shares, so shareholders don’t even get a say.

LUMIA is at eight cents, up 7% in a day, with a market cap under $20 million. A typical micro-cap, tagged as DeFi and RWA.

At this size, a 7% move doesn’t tell you much. On-chain liquidity is thin; one sizeable buy order can pull the candle up, and it’s hard to tell who’s buying. The weekly chart is still falling, so this move looks more like a crawl up from the bottom.

The token distribution, though, is clear. The top ten addresses hold 97%. On Binance, there are only a dozen or so holders, with an average cost of 27 cents. At the current price of eight cents, they’re still down more than 70%. The contract still has minting privileges—a detail more worth watching than the price gain.

IOST is currently priced at $0.0008, up 6.4% in 24 hours. One of the public blockchains from the 2018 cohort, it now has a market cap of around $29 million.

The gain isn’t especially strong, but the volume says more: $3 million traded, just 30% of the usual level, with a 0.1% turnover rate—meaning the tokens have barely changed hands. This kind of rise looks more like there simply aren’t many sellers at this price, rather than strong buying pressure.

A month ago, it was trading considerably higher. It’s still nearly 30% below its monthly high, so this rebound has only lifted it slightly out of the hole. With the order book this thin, buy orders worth just a few million dollars can produce a couple of bullish candles. But if holders try to sell, they may find just as few buyers on the other side.

Amazon plans to pour $220 billion into data centers next year. Jassy laid out the numbers clearly in his a16z interview: that’s how much money they’re spending, and they still can’t keep up with customer demand.

Capacity is already booked out three years ahead. AWS has sold out its capacity for this year, most of 2027 is spoken for, and customers have already signed commitments for what little is left in 2028. Commitments, not expressions of interest—it’s like putting down a deposit now to reserve data center space several years in advance.

Yet chip stocks have been falling all this time on the assumption that there’s too much computing capacity. On one hand, there’s a queue stretching into 2028; on the other, stock prices are acting as if nobody wants to use it. One of these two stories will have to give. Which side is right depends first on delivery: if the data center racks can’t be built, even a mountain of orders is just a number on paper.

Memory chip stocks moved in opposite directions on Friday: Micron rose more than 2%, and SK hynix gained over 4% in South Korea, while SanDisk and Western Digital fell 2–3%. Same sector, but some were being snapped up while others were being dumped. Micron’s earnings report showed that demand for high-bandwidth memory remains robust. Just a week earlier, chip stocks had taken a hit from concerns about “excess computing capacity.”

Amazon CEO Andy Jassy was blunt in an interview with a16z on Friday: AWS has already sold out most of its compute capacity for next year, will be nearly sold out the year after, and has already booked a large chunk of its capacity for 2028. It’s spending more than $200 billion on capital expenditures this year, and still can’t meet a lot of customer demand.

The contrast is hard to miss. Since October, the narrative of a glut in computing capacity has wiped more than $2 trillion off semiconductor stocks, while the S&P 500 and Nasdaq have fallen for two straight weeks. On one hand, stock prices are acting as if nobody needs the capacity; on the other, the companies selling it say customers are waiting in line until three years from now. Put those side by side, and I honestly don’t know which one to believe. Jassy offered a partial answer: very few companies have actually deployed AI inference in production systems. He estimates it’ll take five to seven years for that to become widespread, because inference is still too expensive. The orders are real, but demand is concentrated in a narrow slice of the market—heavy at both ends, with a gap in the middle. He calls it a barbell shape.

Amazon reports its third-quarter results after the market closes this Thursday. AWS revenue was $42.2 billion last quarter, up 36.7% year over year. On the earnings call, whether the company steps up capital spending again or reins it in will be more interesting than the results themselves.