Eric SJ

Attention all traders:

From now on, before making any trades, especially in the morning, I will post an intraday trend analysis for the assets we're targeting, usually BTC, ETH, and SOL.

The trading direction and analysis for the day will primarily follow this note, and the trend must remain consistent unless I explicitly indicate a direction change.

If there's a violation, I'll be sending out red envelopes for each infraction!

Charlie Munger once said: Go fishing where the fish are.

It’s been exactly one month since ARC launched.

Longyi $ARGUS isn’t even at 10M yet, and a few days ago it almost fell below 5M.

If you’re still paying too much attention to ARC right now,

then I don’t want to talk to you!

However, TSMC's $TSM HPC growth has been rapid for some time; this is not a new development that began in the third quarter.

In each of the first two quarters, HPC grew by at least 20% quarter over quarter. What makes this especially impressive is that it has continued to grow by around 20% per quarter even as it has come to account for nearly 70% of total revenue.

Is HPC contributing such a large share of incremental revenue because TSMC is selling more wafers, or because these orders generate more revenue to begin with?

Considering the gap between wafer shipment growth and revenue growth discussed earlier, I’m inclined to think the latter is playing an increasingly important role in this growth cycle.

The revenue mix of advanced process nodes this quarter also offers some supporting evidence.

For example, wafers made using the 2nm process accounted for just 3% of revenue last quarter, but that share rose to 8% this quarter. The share for 3nm also increased, from 30% to 33%.

Together, the two now account for 41% of wafer revenue, up 8 percentage points from last quarter.

Looking at the data from the past few quarters, one possible interpretation is that TSMC may no longer need wafer shipments to grow at the same rate in order to generate higher revenue.

Growth is highly concentrated in high-performance computing and related demand for advanced manufacturing.

And because advanced manufacturing demand is the main driver of growth, TSMC's results alone cannot be taken as proof that the semiconductor industry as a whole is recovering in tandem.

That is also why many market observers say the semiconductor industry has yet to recover.

After all, for manufacturers that mainly produce chips using mature process nodes, customer demand and market conditions may be an entirely different story.

TSMC's latest earnings data is out
$TSM

Net profit has hit a new record for six consecutive quarters

But

we can see that in the third quarter, TSMC's wafer shipments grew just 9.6% year over year—the lowest level in these six quarters

The figures that contrast with this are not only net profit, but also revenue growth

As shown in my chart, the red and blue lines have moved almost in tandem over the past few quarters. This quarter, however, they diverged far more sharply than in recent quarters

(1) Year-over-year wafer shipment growth slowed further, from 16.6% to 9.6%, while revenue growth in U.S. dollars rose from 33.7% to 41.2%
(2) The gap between them was 31.6 percentage points—the widest divergence in the past six quarters

If this were simply a broad recovery in the semiconductor industry, bringing in more chip orders,

we would expect to see wafer shipment volumes rise in tandem

But revenue is now growing far faster than shipment volumes. This suggests that simply shipping more wafers is not enough to explain this growth, nor does it support the conclusion that the semiconductor industry is broadly recovering

A more likely explanation is a shift in the mix of products customers are buying

So, what type of customer is driving this shift?

The answer is high-performance computing (HPC)

HPC is a business category TSMC defines by end application. It includes AI accelerators, server processors, and other chips related to high-performance computing

These chips typically have demanding performance, power consumption, and manufacturing requirements. Many products require advanced process nodes, and some also need complex advanced packaging

In the same quarter last year, HPC accounted for 57% of TSMC's total revenue. In its latest earnings data, that figure had risen to 68%

Based on a rough estimate using quarterly revenue and business mix, HPC revenue grew approximately 80% year over year in the third quarter

Even more strikingly, HPC growth can explain nearly 90% of TSMC's year-over-year revenue increase this quarter

In other words, other businesses are certainly generating revenue too, but the growth driving TSMC is now heavily concentrated in orders related to high-performance computing

This also explains why TSMC's revenue and wafer shipment volumes are becoming increasingly out of proportion

Is the endgame of Web3 just stablecoins???

The market as a whole warmed up somewhat in Q3

Under normal circumstances, market sentiment should have improved as well

But Hyperliquid's core revenue didn't grow accordingly—instead, it fell nearly 4% quarter over quarter

What drove growth for the quarter was reserve income from USDC, something many people overlooked

This is also one of the clearest revenue-sharing deals from the USDC channel apart from Coinbase

And last quarter, this income was only counted for a little over a month, not the full 90 days

The reserve income currently attributable to Hyperliquid averages around $530,000 a day

If the scale of USDC and the effective interest rate stay at current levels, a static estimate suggests it could generate nearly $48 million in revenue in Q4

For a single quarter, that would amount to roughly a quarter of its perpetuals trading fees...

Put that way, it really feels like the endgame of Web3 is stablecoins

In the past, Hyperliquid could only collect fees when users traded

Now, as long as billions of dollars in USDC remain on the platform, the cash and Treasury reserves backing it will keep generating income and buying back $HYPE

There’s currently about $136,000 remaining in the buyback execution contract for $PONS

I’ve been wondering whether, in protocols that are actually executing buybacks, a more flexible buyback design might be better than mechanical automatic buybacks, @MEADGod

For example, when looking at historical trends, if the current market is identified as being in a downtrend, execute a larger proportion of buybacks.

Conversely, if it’s identified as being in an uptrend, reduce that proportion.

For example, PONS buybacks are currently around 0.46e per hour. This is the hourly baseline for the current cycle.

(1) If the trend formed by the past 20 hourly candles is identified as a downtrend, the buyback amount would be 0.6e/hour (1.3 times the baseline amount).

(2) In an uptrend, the buyback amount would be adjusted to 0.32e/hour (0.7 times the baseline amount).

(3) In a neutral market, the buyback amount would stay at the baseline.

I had AI make a simple chart to help everyone understand.

The key here isn’t to have the protocol allocate more than the budget, but to change how the same budget is allocated across different price ranges.

The amount spent less during an uptrend goes into a separate buyback pool, and then more is drawn from this pool during a downtrend. This way, buybacks can continue as usual, while at lower prices the same amount of money can buy back more PONS and carry out more burns.

Using only the past 20 hourly candles is just an example. The mechanism itself is easy to implement—I could build one with Codex, since I implemented similar trend detection for my market-monitoring bot.

A more sensible approach would be to analyze the price trend of $PONS over the past week before executing each cycle.

There is one issue with this design: if the market remains in a downtrend for a prolonged period, the funds will be depleted faster.

So it would be best to use this flexible mechanism during an uptrend, while keeping the original 30% of the pool in reserve.

It could also be implemented at this stage by adding a condition to the contract: before executing each cycle, check the balance of the reserve pool. If the balance is less than 0.3 × the total funds for the cycle, the buyback for the current cycle defaults to the baseline amount.

➠ In other words, suppose 10E is available for buybacks this cycle. If the reserve pool balance carried over from the previous cycle is less than 3E, then the total buyback amount for this cycle defaults to 10E rather than 13E.

The team's trading tactics have changed since yesterday. Maybe more people are watching, and they're closing positions faster.

Everyone knows something's off—it's basically playing with fire.

You ask if you can get in? They can too.
You ask if I'd dare put money in? Absolutely not.

Ever since I posted about it yesterday, this team has been putting out trades more frequently, too.

Most likely the trading really is wash trading; otherwise, with this kind of high-frequency activity and turnover,

there's no way it wouldn't be able to generate buzz around Robinhood Chain 😮‍💨

I just came across a friend sucking up in the comments under a post by $ASTER

I checked the price, and it’s still around 0.7?

The only difference from a stablecoin might be that it depegged for a while at the beginning of the year

Sharing an on-chain observation

Last night, several sizable tokens launched on Robinhood Chain in quick succession.

Their peak market caps ranged from $6M to $30M, and they got there pretty fast. They had a few things in common:

1. They were all launched through Uniswap
2. None were original IPs
3. Their trading tactics were almost identical

They all launched with over 100K in liquidity. Early wallets held a large share of the supply, while also steadily distributing large amounts.

As their market caps grew, liquidity stayed between 100K and 300K.

It felt really shady—like I’d stumbled into a money-laundering operation.

Just when you thought they were about to dump, inexplicable funds would come in and push the price higher.

Search their CAs on 𝕏 and you’ll find that Memes in this market-cap range barely have any community buzz.

0x370dd1181288392843bf52d998eb85c735d91688
0xde09eec6f12311d98d982b07c2e0f9360dd002f9
0xe3e3df8bfb891d269a78ee33a456d36fb7344ec1

After noticing that the first and second had similar tactics, I didn’t hesitate much when the third one appeared—I went in to give it a try.

When I woke up, sure enough, it hadn’t dumped. I sold everything and left it at that; just a small test.

After that, I couldn’t make sense of what was going on.

Sharing this for discussion.

So far, 84 L2 projects have officially announced that they are shutting down.

More than 200 others haven’t made an announcement.

But even more have “died” quietly, without a word.

Of those 200-plus L2s, perhaps only half are still active in the market.

Fewer than one in ten are widely known.

Fewer than one in a hundred are still attracting attention in the secondary market.

In the past, Ethereum showed the industry that L2s were the way forward → VCs and project teams joined forces → airdrop campaigns built community expectations → further fundraising → exchange listings.

Now, building an L2 no longer automatically means it’s the right move. VCs have shifted their focus to AI, project teams are finding it increasingly difficult to make money in Web3, and the number of people willing to participate in airdrop campaigns has plummeted.

So some projects are shutting down, while others are changing direction.

Seed-stage projects used to have a path to listing on top-tier exchanges. But after a full market cycle, what’s clear is that two leading exchanges have sharply reduced their listing frequency.

As a result, more projects that can’t establish a viable business model or attract customers—and bring in outside capital—before Series A will end up in the same cycle: running out of funds, then shutting down.

From what I understand, the OTC market is also extremely sluggish right now. Previously, project teams could use it to exit more quickly. Now, the pool of capital available through this exit route has shrunk dramatically.

When exchanges collapsed in the past, we could clearly feel that “winter was coming.”

This wave of shutdowns and delistings is a different kind of painful silence—the loudest sound is the one that makes no sound.

$OUSD has been live for two days. Its size is now approaching $670 million, making it a top-20 player by stablecoin market capitalization.

Originally, it claimed partnership with more than 140 companies. In the latest published list, the number has grown to 200+.

It doesn’t seem like it’s just marketing. I’ve seen people say that in the end it will only see limited adoption, similar to $PYUSD from PayPal.

I actually don’t think so.

Stablecoin adoption and distribution use cases are the first principles of the business model.

Even though I believe there are still some “extra numbers” in its partner list,

even if only a small portion is real, it will still create broader distribution scenarios (after all, many of them already control payment institutions with large customer and merchant bases).

On top of that, $PYUSD is just one product under PayPal, whereas the core business of Open Standard is to drive adoption of OUSD.

The development path ahead, based on what we can see now, is simply to keep expanding distribution use cases.

For these channels, after integrating OUSD, they can continue providing their existing services for receiving payments, exchanging, and making payments.

Meanwhile, the OUSD balances brought by their customers and the related usage activity can also give the channels additional revenue shares.

This revenue mainly comes from the interest generated by the U.S. dollar reserves behind the stablecoin.

(OUSD’s design is that, after deducting management fees, this portion of the收益 is distributed to the partner companies that help drive adoption.)

They can choose to keep this revenue themselves, or they can take part of it to subsidize customers to bring in more capital.

Referencing Circle’s business model, this revenue is constrained by interest rates. Even if the number of payment transactions increases, reserve interest will not grow proportionally.

So the appeal will differ for different channels.

But at least, the business motivation for promoting it is valid.

So far, Stripe has set OUSD on Tempo as the default stablecoin configuration and integrated it into the product’s checkout/payment flow.

For Stripe’s customers, it may be nothing more than enabling a payment service they were already using.

In the future, large-scale stablecoin adoption will happen when many users don’t even realize they’ve already started using stablecoins.

Earnings beat expectations but still couldn’t sway the market??

The current Micron earnings situation is basically a level-hell difficulty

$MU — Shares fell more than 3% in the first hour

And the earnings released after the close yesterday were above market expectations

On one side, memory demand is being described in increasingly exaggerated terms

On the other side, the market widely doubts the durability of that momentum

And this skepticism has been present in the market for a while

Just when you think this time’s upside surprise is already “locked in,” the market says, “Not yet—there’s still room to be even more surprising than expectations.”

When you bet on the next “even more than even more” upside surprise, the market has already priced in (“over-discounted”) this “even more than” part ahead of time

I’ve seen market voices say, “Earnings beat expectations but still couldn’t impress the market.”

Based on the numbers, this earnings report’s revenue and EPS:

➠[Q4 revenue was $54.23 billion, versus market expectations of about $51.07 billion; adjusted EPS was $3.342, versus market expectations of $3.161]

The midpoint for next-quarter revenue guidance is $61.5 billion, whereas the market had only expected around $57 billion. The EPS guidance is also higher than market expectations

Within the guidance, the less encouraging figure is the gross margin guidance, which declines slightly quarter over quarter

➠[Micron’s Q4 adjusted gross margin was 87%, but next quarter’s guidance is 86.25%]

In other words, while revenue grows sharply, gross margin actually dips slightly quarter over quarter

However, management specifically explained on the earnings call that the Q1 gross margin decline is largely due to Q4 increasing employee incentives

So at least according to management’s explanation, this quarter-over-quarter gross margin drop is not mainly caused by demand or pricing hitting a peak

And if you look at the capital expenditures line item, there’s also no obvious cause

This looks pretty similar to the general expectations from prior quarters: capacity expansion, continuous expansion

So endlessly asking the same question doesn’t really matter

If they stop expanding capacity, the market might actually be even less happy

Therefore, if you absolutely have to find a fundamental change inside the report that could explain a 3% drop, it feels a bit too far-fetched

And it’s hard to say the market simply won’t buy this deal......

There’s a difference between on-chain and the secondary market: for the same order of magnitude—millions to ten-millions-level assets—

For Robinhood Chain, you basically can’t see what people call so-called support and resistance levels.

A 20M position can return to 5M in thirty minutes, and a 2M asset can also be ramped to 10M within thirty minutes.

But for assets in the millions to tens-of-millions range, on-chain is much, much better than the secondary market.

Because the latter most likely has a large trapped-fund supply—coming down from hundreds of millions to over a billion.

So on-chain vs the secondary market for selecting assets—the answer is very clear to me.

At the level of ten-million and below, the on-chain risk-reward is a bit higher.
But if it’s a market cap in the hundreds of millions, the secondary market is more稳妥.

What size of capital should go into what kind of pool to “fish”?

This long medical-industry report of mine

If I had to recommend just one company, I would recommend Intuitive Surgical $ISRG.US

Why?

Because it is almost like the [Apple of the medical field]

Robotics + healthcare is a market with both high certainty and ample room for imagination

According to estimates from third-party reports

[Intuitive Surgical’s share in the global surgical robotics market is close to 60%, and in sub-segments its share is even close to about 80%]

Its core product is a surgical system known as [da Vinci]

This is not a robot that performs fully autonomous surgeries—it assists surgeons during operations

Its business model is similar to Apple’s: it sells expensive equipment to customers first, and then it keeps charging around that equipment through consumables and after-sales service

So, Intuitive Surgical actually has three layers of revenue:

(1) Selling or leasing surgical robots

(2) As the number of procedures increases, it continues selling instruments and accessories

(3) Providing maintenance services to hospitals that already have the installed equipment

Among these, the parts that can generate long-term repeat revenue are the latter two

➠ In the most recent quarter, Intuitive Surgical achieved $2.892 billion in revenue, including $1.735 billion from instruments and accessories and $472 million from service revenue

That means instruments, accessories, and services together contribute about 76% of revenue, while revenue from selling robots alone accounts for only about 24%

As shown in Figure 1, it reflects Intuitive Surgical’s installation growth trend over the past six quarters and changes in its fee structure

The combined share of average revenue for both items remains above 75%, and installation volume growth is also steadily expanding

And what these later charging steps reflect is the workflow that customers build around the same system

That workflow itself is the moat—and an exceptionally clear one. Even if later competitors build a robot with parameters close to da Vinci, they can’t directly replace da Vinci

The entire business model is driven by a double-helix mechanism:

The more robots installed, the more doctors adopt the da Vinci system, and the higher the replacement cost later
The more robots installed, the more after-sales services and accessory consumables are needed

From the current market view, the W-bottom pattern formed on the daily chart is at a key resistance level, waiting for a confirmed breakout

That’s all for now. The above

What trashy and counterfeit market trends are starting on the market right now

Can this really be called an End-of-the-World war machine?

A real End-of-the-World war machine—I only recognize these three $DOT $FIL $ICP

If I have to add one more, it would be the legendary $EOS from days gone by (now it has been converted into $A)

In the last two rounds, every counterfeit coin faces these four ancient true gods—it’s like a minor vs a major

They all have a few common traits:
（1）All were once in the top ten by market cap
（2）A large number of holders
（3）Top-tier narratives
（4）God-tier lineups

$HYPE Surprisingly, it got listed on Binance spot

Now that it’s come to this, $ASTER only has [Spiral ascent and take profit. Make everyone follow.] left as a way to break the deadlock


Otherwise, tomorrow $LIT will also be listed