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!

However, there are still about 8 billion people in the world who do not have Bitcoin. The group holding $BTC accounts for only about 4.4% of the total population.

According to AI estimates, the number of people who hold gold is about 18%—30% of the total population.

Therefore:

The number of gold holders is approximately 4.1—6.8 times that of Bitcoin holders, and the market value is about 24.4 times that of BTC.

Empty, why not empty—what reason not to be empty?

When you reach the resistance level and then wait for emptiness at intermediate positions, the risk-reward ratio on intraday trades looks really bad, and there’s no certainty either.

$BTC —if aiming downward conservatively, it can still be a good 500 points.

I highly recommend everyone to pay attention to Cloudflare ($NET.US ). This company’s market value is currently in the trillions.

I only realized how potentially amazing this company could be after I recently built my own website.

If you’re like me right now—where you hardly ever proactively open a specific website, and instead just let AI fetch things for you—then you’ll understand what I’m about to say.

Because Cloudflare has embedded one of its business scenarios into the middle layer between AI and websites.

➠ In the past: I have a need → open a website → the website returns data

➠ Now it’s becoming: I have a need → ask the AI → the AI accesses the website/API → scrapes and organizes the data → returns it to me

What it’s doing isn’t actually creating an entirely new business. I took the time to break it down over the past few days, and I found that almost all of its product launches are built around this propagation path:

➠ customers → Cloudflare → origin server/app ← Cloudflare ← requests from external users or Agents

Even the future business logic is the same—you could say it’s evolving from “carrying the internet” to “carrying the internet empowered by AI.”

As of 2026 Q2, about 23% of websites worldwide already use Cloudflare as a reverse proxy (Figure 1).

That means Cloudflare already sits at the request entry point for a sizable portion of websites.

And from the data of the past seven quarters, it has already validated the growth potential of the original business cycle in advance (Figure 2):

➠ quarterly revenue increased from $460 million to $696 million;

➠ year-over-year growth rate rose from 27% to 36%.

➠ DBNR recovered from 111% to 120%.

➠ the number of enterprise customers with annual revenue exceeding $100k increased from 3,497 to 4,698.

This original business was honestly not that “sexy.” It was more like an old story.

The new story (the link where AI requests access to websites) is what I personally feel has the biggest room for growth.

Because one person can’t, at the same time, continuously open a dozen websites and repeatedly call dozens of APIs, all day long, executing tasks. But that’s exactly the kind of behavior that AI can do—and does.

According to Cloudflare’s own network observation metrics, non-human traffic has exceeded 50% for the first time.

Not all of it is AI Agents—there are also traditional scrapers, automation tools, and malicious bots—but the direction is already very clear: machines are becoming an important visitor to the internet.

It wasn’t until deals like TMD were made that I realized the previous line had a typo—it should be $BTC 6万4, with a sell short order 😅

$BTC 6w5 sell short pending order

If it gets filled, settlement must be completed before the CPI release in the evening

$SOL temporarily encountered resistance in the 15-minute cycle

Within the day, I’ll do a wave-only trade and watch for a pullback to see if we can break below this trendline and reverse direction following it

If I’m lucky, I can have it fully delivered before bedtime tonight. The stop loss is simply set at the previous high 76.66

Once I’ve placed the bet on it, if I didn’t place it then I can push the loss and get it to the next day

Over $400 billion left behind by Buffett has started to be spent

What does this tell the market?

This may be the most worth taking seriously Berkshire Hathaway’s earnings report over the past three years $BRK.B.US

Because this time, the data shows that Berkshire has once again become a net buyer of stocks, ending its prior streak of 14 consecutive quarters of being a net seller of stocks.

In the past few years, Berkshire has been selling everything and not repurchasing.

Then the cash kept piling up, and ultimately reached nearly $400 billion.

And the latest earnings report shows that in Q2 Berkshire bought about $23.5 billion of stocks and sold about $3.7 billion, resulting in a net stock buy of nearly $19.8 billion in the quarter.

I think the most important thing to highlight is the share repurchase action.

➠ Berkshire repurchased about $4.8 billion of its own shares in the first half, and the vast majority occurred in Q2 (about $4.5 billion).

Here’s some background you need.

Starting in 2026, Greg Abel officially took over as Berkshire’s CEO, while Buffett continues as Chairman. Berkshire’s current repurchase policy also clearly puts the capital allocation decision in the hands of the CEO, and requires consultation with the Chairman.

In other words, after both parties discuss it, if the repurchase price is below their conservative estimate of intrinsic value, they will proceed with the repurchase.

So at the very least, in Q2 it tells us one thing: the top management at Berkshire believes the price is undervalued and worth buying.

The interesting part is that Berkshire itself is a huge asset portfolio.

Repurchasing Berkshire’s own shares is, in essence, not fundamentally different from buying Apple and related portfolio holdings.

Because Berkshire isn’t a single business company; it’s a bundle of capital allocation made up of insurance float, operating businesses, and a stock investment portfolio.

This is more direct than any interview where someone says, “I’m confident about the market ahead.”

_____________________________________

However, Berkshire’s net stock purchases in Q2 were only about 5% of its broad cash position at period end.

So if you conclude that the bearish trend has ended based on just this one quarter, it would be too early.

All we can say is that over the past few years, this cash has been more like a card laid on the table.

Berkshire chose not to play its card—until Q2, when it played part of it.

But the card still hasn’t been played out.

At least it has started playing—that’s a good sign. You can observe it; I won’t go into more details from the rest of the earnings data.

A little-known fact: Pump Fun’s DEX business, PumpSwap, is also one of the largest DEXs on Solana.

This means Pump Fun occupies two key positions at the same time:

The largest launchpad + the largest trading venue

And the biggest source of Pump Fun’s fees is actually not the launchpad, but PumpSwap.

After reorganizing the data in Figure 2, it can be broken down into three parts:

（1）Bonding Curve fee volume: $78.14 million

= $62.06 million in protocol fees + $10.86 million in creator fees + $5.22 million in rebates

This is the core of Pump Fun’s token issuance stage, and it’s also currently the largest portion of the protocol’s cash inflows.

（2）PumpSwap fee volume: $124 million

= $23.40 million in protocol fees + $25.64 million in LP fees + $74.87 million in creator fees

This is the part that the market is most likely to overlook.

Once the Token completes the Bonding Curve, it doesn’t end there—it moves into PumpSwap for further trading, carrying the token through the second stage of its lifecycle.

（3）Other business: $10.11 million

Smaller scale, not expanded here.

Note that DefiLlama tracks the total fee volume generated by the protocol; some of it is allocated to participants such as LPs and creators.

So what you truly need to look at is the money the protocol keeps.

In the Q2 fee volume of about $212 million, Pump Fun ultimately keeps about $91.55 million in gross profit.

Of that:

➠ PumpSwap captures $23.40 million
➠ Bonding Curve captures $62.06 million

In other words:

PumpSwap has a larger fee volume, but the Bonding Curve has stronger value-capture ability.

Now let’s look at the trend over the past few quarters (Figure 3):

The Bonding Curve clearly depends on the Meme market’s heat and Solana’s on-chain activity, so it’s more volatile.

The PumpSwap growth curve is smoother, which also indicates that Pump Fun’s revenue structure is changing:

In the past, it earned money from the token issuance stage;

now it’s starting to capture revenue from the token’s circulating/liquidity stage.

Bonding Curve earns from issuance; PumpSwap earns from the lifecycle.

Finally, take a look at this:

$PUMP is currently in a pullback phase within an uptrend. Around 0.0022, it has not yet touched any key support for the moment.

That’s all for now~

Although liquidity is poor on the weekend, currently on the $BTC board the price has formed a triangle relay pattern and closed it out.

In my opinion, from this position at the upper edge, you can try taking a directional trade for an intraday cycle—go short.

The stop-loss is also easy to place. Since it’s the weekend, the 6w5 integer level won’t be easily reached; just place the stop above 65066. For extra safety, 65120 is also fine.

Why put it so close? The reason is still the weekend—if the weekend can push up into this level, then it’s game over.

Suddenly realized

“Circle $CRCL and companies like it—stablecoin issuers—feel a bit like someone standing in the era of internet media, yet still choosing to go all-in on old-fashioned print media.”

“No matter how chaotic it gets on-chain, it still eats its share of profit from reserve yields based on scale.”

“In a future rate-cut cycle, Circle may be in trouble.”

Add some data:

Circle’s second-quarter USDC scale and usage are growing rapidly:

➠ End-of-period circulation was $73.3 billion, up 19%; quarterly on-chain transaction volume was $148 billion, up 151%

➠ But Circle’s total revenue and reserve income were only $701 million, up 7%, below expectations

Putting it together:

USDC growth is network-scale growth, while Circle’s growth is mainly interest-income growth—there’s no strong positive correlation between the two.

Because stablecoin issuers don’t take fees from every on-chain transaction, the gas fees generated by USDC transfers mainly go to the respective chains.

So that $148 billion in transaction volume reflects an increase in USDC’s usage and settlement role, but it doesn’t directly translate into Circle’s revenue growth.

What truly determines Circle’s profit is reserve yield.

Average USDC circulation increased about 25% year over year, but reserve income grew only 5%. The core reason: the reserve yield fell by 66 basis points year over year.

Scale is growing, but unit returns are declining.

This also reinforces the point:

“No matter how chaotic it gets on-chain, it still eats its share of profit from reserve yields based on scale.”

Stablecoins are an on-chain product, but the issuer’s income statement is highly dependent on real-world interest rates.

If the economy enters a rate-cut cycle, the same scale of USDC would generate noticeably less revenue.

And currently, reserve income makes up about 95% of Circle’s total revenue—other businesses aren’t yet enough to become the second growth driver.

Circle has a massive pool of US dollar reserves, but it hasn’t built a value-capture system that’s tied to stablecoin network growth.

So it urgently needs a second growth curve.

Otherwise, a stablecoin-issuing company is like someone in the era of internet media, yet still choosing to go all-in on old-fashioned print media.

This time, $BTC 6w5 is probably going to break upward

When you zoom in on the amplification period, it doesn't seem quite right

This round $BTC fell back to the upper edge of the channel and is considered very standard.

Give an opportunity to get on the train; the risk-reward ratio is also very good. For stop-loss, simply place it at the channel’s middle line.

Recently my analysis style has changed a lot. I don’t want to talk too much about this or that—straight to the point.

It’s going up.

This is a long setup after a pullback to support.

Big pancake, stop causing trouble with me anymore—get up there quickly, I want to sleep.

$BTC reached the initial target. I didn’t even move the line; the profit-to-loss ratio is 2-to-1

If we can break through this channel, then we’ll need to get back to 65,000.

Get it up!

各单位注意各单位注意

$BTC 要上去了，依旧日内单

小周期往下插了一下收了个pinbar出来，沿着支撑趋势线估摸得上去了

500点