Powerpei

I joined Binance Square in November last year.
Back then, I saw Yingge promoting the Creator Platform, so I hopped on.

I have to say, Yingge is kind of my benefactor.
I should say she’s a benefactor for a lot of small retail investors. Everyone who knows her knows she’s very capable and also easygoing (and the key is, she’s also beautiful).

The deepest impression I have is when I ran into some small problems and went to ask Yingge for advice—I was afraid I wouldn’t do it well. Yingge told me: Just be yourself.

After that, I met Cy, who also gave me a lot of help and guidance.

I didn’t think too much. I just adjusted the content I usually posted to align more with the task direction.

Bit by bit, I managed to get a few first-place rankings; in total, I think it added up to nearly $10,000.

To be honest, that number isn’t that big, but the advantage is the low barrier to entry.
You don’t need 100,000 followers. You don’t need to grind data every day. If your writing is good, you still have a chance to get on the leaderboard.

Even now, I still update posts every day. Occasionally, when I see creator tasks that fit, I’ll join in.

Not grinding, but it’s pretty interesting.

Compared to writing on X for half a day just to get a few likes, Square at least gives you positive feedback—you write something and you can actually earn money.

#币安广场

PONS is really making waves lately

CoinGecko’s search interest ranks #1 over the past 3 hours; it’s doubled over 24 hours, nearly 6x over 7 days, with a new price high and market cap reaching $200 million

It’s a token launch platform on the Robinhood Chain

Users can create a fixed-supply token in one click using their own wallet, while also deploying a liquidity pool with locked liquidity. The platform doesn’t touch your funds

What I find interesting is the deflationary design
A portion of the trading fees goes back to the protocol, and most of it is used for automatic buybacks and to burn PONS. So far, it has burned nearly 30% of the total supply

The platform has launched hundreds of thousands of tokens in total, and trading volume is also not small
This indicates that usage is real
It’s not inflated by airdrops; there are actual people creating tokens there and trading them

But I’ll say this: the deflation mechanism looks good only if trading volume can hold up. Right now, hype is at the top—fees are high, burns are fast, the price is rising, and hype is even higher
That’s a positive feedback loop. The problem is that when things go the other way, it’s just as fast. Once trading volume drops, the burn rate slows, and the narrative won’t stand

The window for meme token launch platforms is always short
Pumpfun got popular; now PONS is taking the baton. Whether it can retain users depends on whether the platform has anything left once the hype fades

Crypto markets are extremely volatile. The above is market observation, not investment advice.
$PUMP

Kanye released a coin called YZY on Solana in August last year. The branding was Yeezy-branded payments ecosystem—paired with Ye Pay and the YZY Card.

On the day it launched, it surged to a valuation of several billion dollars, and then it collapsed all the way down. By the end of the year, the price was only around $0.0008.

Then, in 2026, it rebounded.

As of August 29, YTD gains are roughly 42,730%. CoinLore ranks YZY #1 for 2026 performance. The current price is between $0.29 and $0.35, and the circulating market cap is roughly between $90 million and just over $100 million.

The numbers are terrifying. But I need to explain how this surge actually happened.

From $0.0008 to $0.3—the percentage is obviously outrageous—but the starting point was a price that was almost zero. A low-base rebound and ongoing fundamental growth are two different things.

A few facts: total supply is 1 billion coins, circulating supply is about 300 million, and the remaining ~70% is still locked in Yeezy Investments LLC’s vesting/entitlement plan.

On August 16, about 120 million coins were just unlocked—that’s 12% of the total supply.

The pattern for meme “starnames” never really changes: when it’s rising, it makes you feel like a genius; when it’s falling, it makes you question your life. YZY already played this out on its very first day after launch.

NVIDIA Q2 Revenue Reaches 96.2 Billion, Up More Than 100% Year Over Year
Guidance Is Even More Aggressive

For FY2028, growth of 70%—before Wall Street would only dare to project 45%

The stock surged 8.74% on the day; single-day market cap added $442 billion

My take: The most important part of this earnings report isn’t the numbers themselves—it answers a question that’s been hanging for two months: Has AI compute demand already peaked?

The answer is no.
And it’s accelerating.

The market was briefly led astray by DeepSeek—after reasoning efficiency improves, training demand would supposedly collapse.

But the 70% guidance tells you one thing: Big tech is still feverishly stockpiling GPUs, and order visibility is far higher than what Wall Street models assumed. Efficiency gains haven’t reduced demand; instead, they make more scenarios cost-effective. Demand is being unlocked—not replaced.

My own positioning logic hasn’t changed:
NVIDIA isn’t cheap, but it’s the toll gate across the entire AI infrastructure pipeline.

As long as downstream players keep expanding capacity, it keeps collecting tolls.
The thing that has truly made me hesitate has never been valuation—it’s only been the day when the customer starts cutting capex.
This earnings report tells me that day hasn’t arrived yet.

Take one step further: If NVIDIA says 70%, should upstream HBM, CoWoS packaging, and power infrastructure growth expectations also be revised upward? Those have not yet been priced in sufficiently.

The above is for personal information breakdown only and does not constitute any investment advice. DYOR $NVDA

This price action is pretty wild

It went from a market cap of a few thousand dollars all the way up to a few tens of millions. It got cut roughly in half several times in the middle, then rallied back again. Recently, from the lows it rebounded again with increased volume; the current price is about $0.058, up over 40% in 24h, and the market cap is back around 58 million

The previous high was about $0.077 (market cap close to 80 million). It’s still ranging within this box. Trading is active, with about 48,000 holder addresses

The attention brought by the film going viral is still there. But fundamentally, it’s a sentiment-driven market—up and down like this is normal. Don’t chase the price.
$牛来

I always felt that the way bStocks used to be played was too monotonous: you buy, wait for it to rise, and then sell. That’s pretty much no different from buying tokenized stocks directly on a normal exchange.

The composability that on-chain assets should have hasn’t really been used—this time, with the @Binance Wallet launch of the bStocks DeFi Center, the gap has finally been filled.

First: provide liquidity, earn trading fees
Put bStocks into the liquidity pool. When others buy and sell, the trading fees generated are shared with you.

Traditional LP operations require you to prepare the corresponding assets in advance and manually set the ratios; concentrated liquidity pools also need you to adjust the asset proportions based on price ranges, which is a hassle. Now, with the V3 pools that support Zap, it’s much easier: you can deposit single-coin or any ratio of bStocks + stablecoins, and the system automatically sets the proportions. If you want to exit, you can even withdraw only one type of coin.

Honestly, I really like this design. Before, holding bStocks meant you could do almost nothing besides wait for it to go up. Now, at least during a sideways market, you’re still getting fees.

Second: collateralized lending
Hold NVDAB (Nvidia bStocks) and SPCXB (SpaceX bStocks). If you don’t want to sell but need USDT to do something else, you can now pledge bStocks and borrow against it.

When the stock goes up, you still earn—while you also have an additional amount of usable funds in hand.

This logic is called stock collateralization in traditional stock markets. Back then, you had to go through procedures with a broker. Now you can do it directly in your wallet, which is the efficiency improvement that— in my view—on-chain truly brings.

The lending list will show the collateral, borrowable assets, protocol, borrow interest rate, and total borrowed amount. It supports quick viewing of borrowable assets and which bStocks can be used as collateral.

Entry point: Binance Wallet → Stocks section → bStocks Savings & Lending → LP or Loan tags

My real take: between these two features, I’m more bullish on the lending side.

The reason is: with liquidity provision, there’s the issue of impermanent loss.

If your bStocks price is volatile, when you withdraw it, you might end up with less than if you had just held directly. For people holding high-volatility assets like Nvidia and SpaceX, the returns might not be worth it.

The lending side is more straightforward.

You maintain your exposure to the stocks while freeing up a portion of liquidity. If you’re right, you benefit on both sides.

Of course, the risk is here too: if bStocks drops below the liquidation threshold, the system will forcibly sell to repay the debt.

This isn’t a small matter—before borrowing, think carefully about how much drawdown your position can withstand.
DYOR.
#BStocks

On August 19, Moderna’s stock surged 177% in a single day.
From $63 straight up to $174. Its market value grew by nearly $45 billion in one day. On the day, the shorts had a paper loss of $5.5 billion.

A vaccine stock—
not AI, not chips, but a cancer vaccine.
——
Moderna and Merck collaborated on a personalized skin cancer vaccine, and the Phase 3 clinical trial succeeded.
In plain language: after cutting out the tumor, this vaccine significantly lowers the chances of cancer coming back and spreading.
Over the past two or three years, the market felt the mRNA pandemic dividend was over—Moderna was basically being treated like a forgotten corpse.
Then, one round of cancer data directly slapped that narrative in the face.
· · · ·
The next day it gave back more than 20%; the profit-taking crowd ran.
But on August 25, Barclays raised its target price, and it jumped another 14%, closing around 158.
It’s not a one-day drop and you’re dead. Someone is seriously re-pricing the story.
——
My take:
The AI narrative is too crowded. The market is extremely hungry for a “real breakthrough that isn’t AI.”
Anything that can prove “technology can still truly save lives” will be priced insanely.
I won’t chase it just because it spiked one day, and I won’t dismiss it just because it pulled back the next day. The scientific progress is real, but the pricing may be running ahead of commercialization. $MRNA

The hardest part isn’t finding trading opportunities.

It’s holding on when nobody around agrees with you and the charts have been dead for weeks. @osf_rekt said it perfectly.

I’ve been through it. I buy what I believe is undervalued, watch the timeline rotate through five different narratives, and my position just sits there, flat and lifeless.

Every day you doubt yourself. Whenever something else pumps, it feels like proof that you were wrong.

But that’s exactly the filtering mechanism. Most people can’t stand boredom. They need action. They need the screen to move.

So they sell too early, chase the next thing—over and over again.

Belief-based trading isn’t glamorous. There’s no screenshot of “I caught a 5-minute candlestick.” There are only weeks of silence, and then one day the thesis comes true—you realize the edge was never about the entry.

It was about the patience afterward.

“Find what you’re good at, and do as much of it as possible—not chase what you’re not good at.”

That’s the whole game. Respect.

XRP just poked the top around 1.55 on the 15-minute chart, and now it’s pulling back to 1.4587, down 2.6%

Looking at this volume spike, to be honest, I’m a bit uneasy
The short-term moving averages are pressing down hard, the sell pressure is still there, and a reversal isn’t that easy

Worst of all is that Charles message—he somehow again brought up the CLAIMIT bill, planning to re-label XRP as a security
With this double squeeze from both geopolitics and regulation, market sentiment is stretched so tight it feels like a string

First, let’s see whether the 1.45 support level can hold
If it can’t, then in the short term it will most likely still need to go lower to find a bottom

But if you’re holding spot and don’t feel like selling right away, you can open the Prediction section in the Binance Web3 Wallet—since the market is currently choppy, you might as well use these swings to hedge a bit within the ecosystem

What everyone’s playing with—predicting predicted market trends—is itself a kind of defensive strategy, and you can even earn some WLFI rewards along the way.

Right now the market is basically a game of who can be more cautious—don’t open high-leverage positions while everything is just swinging wildly; otherwise you’ll end up taking the loss.
Let’s see if 1.45 support proves itself—hold steady and we’ll talk again.
$XRP

Last week, U.S. Treasury yields suddenly spiked, startling the whole market

This week (Aug 24–28) is even more intense
With AI earnings reports and the Fed’s stance—both laid bare in the same week

I’ll go over a few key points

//

On Wednesday night, Nvidia (NVDA) releases its earnings report after the close

This is the single most important event of the week—no question

Wall Street expected revenue: $91–95 billion, nearly doubling year over year

But the numbers themselves aren’t the main point

What everyone is really nervous about are three things:

➢ How smooth the chip transition is—Blackwell moving to the next generation, Vera Rubin
➢ Whether data-center orders are still booming
➢ Whether gross margin can keep holding around 75%
(Explanation: gross margin is how much gross profit you have left out of every 100 dollars in sales. 75% is already extremely high—the market worries it could drop.)

If Nvidia’s guidance for the next quarter beats expectations → people will feel that the AI investment cycle isn’t anywhere near over, and they’ll keep buying

If the guidance is conservative—even a little → highly valued tech stocks could all fall together

On the same day, Salesforce and CrowdStrike also report results

In other words, the entire AI software sector is being tested at once: Are you making real money with AI or not?

//

Macroeconomic data is also packed into this week
Core PCE and GDP revisions come out back-to-back

You can think of PCE as the inflation gauge the Fed cares about most

If this number comes in too high → it suggests inflation hasn’t cooled enough, interest rates may stay elevated for longer, and the stock market will face pressure

If it unexpectedly falls → tech stocks may finally catch their breath

//

On Thursday: the Jackson Hole annual symposium

Fed Chair Kevin Warsh is scheduled to speak

This event happens once a year. Global central bank governors all attend, and it’s one of the most important occasions when the Fed signals its intentions

With the 30-year Treasury yield just brushing the highest levels in nearly 20 years, the market really wants to know: What does the Fed think about this yield spike? Will it take action?

Every word Warsh says will be parsed line by line

//

In between, there are also consumer confidence, new home sales, and earnings from several retailers

These are more like background indicators
But they answer one question: Are ordinary people still spending?

//

So the core conflict this week is:

> In AI: the story is still accelerating, and Nvidia is the proving ground

> In interest rates: persistently high yields are compressing valuations

> For tech stocks this time, they have to answer two questions at once—whether growth is still there, and whether the cost of that growth gets eaten away by higher rates

A dog sticking a pipe pops up on the Robinhood Chain. pipedog ( $PIPEDOG )
Official quote: “this pipedog / he has pipe / he has question”
Pure Meme, community-consensus driven, no complex storyline.

//

Total supply is about 12.34 billion coins, fully circulating, with no unlock pressure.
Launched on July 28. The next day it touched an ATH above 0.005, then it later dropped by more than 60%.
It’s now around 0.0022, with a market cap of about $27 million.

//

What Meme coins are betting on is whether attention—and whether the chain itself—can bring traffic over.
I don’t hold any position; I’m just observing.
Stuff like this can rise by a few multiples, or fall to zero—there’s no in-between.
If you want to play—use spare money, keep a small position, and don’t get carried away.
Data as of August 23, 2026; refer to live market quotes for the latest.

ZEC breaks 800

This isn’t just a modest rise—over the past 24 hours it surged more than 30%, briefly touching around 859.

The 2018 high was roughly at 800, and after that it never came back.
This time it went straight through.

//

The catalyst is very clear.

Grayscale is pushing its Zcash Trust to an existing spot ETF proposal, planned for the NYSE Arca under the code ZCSH.

DCG-related entities are still discussing acquiring about 200,000 ZEC via a trust—at the then-prevailing price, that’s on the order of $160 million.

Institutional allocation expectations jumped immediately.

//

Another thread is the privacy narrative.

Zcash uses zk-SNARKs—zero-knowledge proofs—offering optional privacy.

Shielded pool usage is rising. The network has just completed the Ironwood upgrade, and with the AI era, people are becoming increasingly sensitive to data privacy.

With these factors stacking together, sentiment turned bullish.

Winklevoss’s side has laid out Zcash mining capacity through a Cypherpunk strategy. They hold a significant share, and the supply side is also tightening.

//

My take:

The direction is correct.

The privacy-coin track has been dormant for too long. In this sector, ZEC’s technical maturity really is the highest, and the institutional signals are also the most concentrated.

But in a 24-hour window, the swing exceeds 40%. Futures trading volume has reached the multi-billion-dollar range, with a very high level of leverage participation.

With this kind of move, the risk of chasing is something I don’t need to spell out.
Next, what needs watching is whether the ETF can truly get finalized, and whether regulators’ stance toward privacy coins will shift.

The current price is roughly fluctuating in the 800–820 range. Market cap has risen to around $13–14 billion, putting it among the leaders in the crypto market.$ZEC

The past few days’ performance of mRNA has been quite interesting

On August 19, Moderna and Merck (MSD) released interim Phase 3 results for their mRNA cancer vaccine

With a personalized vaccine plus Keytruda, it significantly outperformed Keytruda alone in melanoma

The trial enrolled 1,137 participants, and both the primary and secondary endpoints were met

Previous Phase 2b five-year data: recurrence risk reduced by 49%, and risk of distant metastasis reduced by 59%

On the day, the stock price jumped from 63 to 174, up 177%
The next day it fell 23%, closing at 133

//
Scientifically, this is a real breakthrough
For the first time, mRNA has proven itself in a large-scale Phase 3 cancer trial
It’s no longer just a COVID vaccine
But on the trading side, I won’t chase after it has already run up; even so, after the drop, MRNA is still 63% above the average analyst target price of 82.62

In the last quarter, the company had a net loss of 782 million, with revenue of 145 million
Personalized manufacturing, costs, approvals—every step takes time
Out of the 177% surge, how much was driven by sentiment and short-covering? The 23% drop the next day already provides part of the answer

//

Moderna indeed has an additional new story supported by data
But the current price has already priced in a lot of the optimism in advance
Don’t chase, and don’t deny it either
Wait for the full data—no rush
The above is purely my personal observations and not any investment advice. Dyor
$MRNA.US

SEC this week released a new rule proposal specifically targeting a registration exemption for crypto investment contracts.

Two paths:

The first is a one-time exemption, allowing issuances of up to $5 million within four years. The second allows issuances of up to $75 million every 12 months.

Both require the issuer to make principle-based, narrative disclosures to investors. The second threshold is higher—you must provide financial statements and continue with ongoing reporting.

In plain terms: the SEC is no longer blanketly saying “all token offerings are illegal,” but instead has provided a compliance pathway.
You can issue—but you have to disclose, be transparent, and follow the rules.

//

0xSammy mentioned that this couldn’t have better timing for $UMIA. I looked into the logic:

What UMiA is doing is capital formation for token-native projects plus a governance technology stack.

Simply put, they help projects issue tokens and raise funds in a compliant way, and build governance. They’ve already submitted Blockworks B1 transparency disclosures.

If these SEC exemption rules really take effect, then for infrastructure like Umia—“helping projects go through compliant issuance processes”—demand would ramp up directly.

Before there was a compliance channel, everyone could only operate in grey areas.

Now that there are clear rules, projects that want to do things the right way will need tools and services to help them meet disclosure requirements.

A few elements are indeed aligned: regulators are opening a window + the infrastructure is already being built + the token has visibly bounced today.
Whether it’s worth following, you decide. Early-stage projects + policy still in the proposal stage means there are plenty of uncertainties. DYOR.
#SEC

Binance Wallet’s prediction market has reached TI playoffs. For tomorrow’s Iron Wing vs Spirit BO3, I’m planning to place a bet.

I’m picking Iron Wing.

//I

Iron Wing (former Tundra roster) went from the group stage all the way up through the playoffs, beating Nigma 2-0, then taking down GamerLegion. They’re only getting better as the games go on.

Spirit, on the other hand. They started 3-0 perfectly, then went on to lose back-to-back against Vision and Nigma, and nearly missed the playoffs. Yatoro has nothing to say for himself, but right now it feels like he’s carrying the whole team by himself.

In a BO3, current form matters more than pedigree. In lane matchups, Iron Wing doesn’t have to take a loss either—BZM is likely to outplay Larl. 33’s in great shape, and Pure can compete in both early and late game. Market odds are 52% vs 48%, almost a fifty-fifty. If Iron Wing is truly stronger, there’s room in this line.

————

How it works
: Choose a Dota 2 match that qualifies, lock in at least 20 shares in a single entry, and complete it before the match starts.

Guess correctly to earn points. The grand final round has a prize pool of 75,000 USDT + 4,000,000 PP.

The entry is on the Binance Wallet homepage in the Dota 2 banner. If you haven’t played before, opening it gives you a 1U体验金 (bonus). Don’t hesitate.

This round: $300. Iron Wing wins.

Bet the trend, not miracles. If you lose, accept it—if you win, enjoy it. Don’t get carried away. DROY

Last night, the Philadelphia Semiconductor Index fell by nearly 5%, closing at 11,992. Chip stocks were hit across the board, and the Nasdaq also dropped by 1.33%.

Two things came at the same time:
The 30-year U.S. Treasury yield spiked, once hitting a near 19-year high.
Long-term yields moved higher, meaning the market is repricing borrowing costs.
For a capital-intensive industry like semiconductors, the higher the interest rate, the more valuation pressure there is.

On the other side, tensions between the U.S. and Iran remained stuck around the Strait of Hormuz, pushing oil prices up. Higher oil prices → inflation expectations heat up → rate-cut expectations get pressured → growth stocks face continued strain. With both lines tightening at once, semiconductors are the first to get hit.

————

A 5% drop in a day isn’t unusual—PHLX Semiconductor has always been volatile. But the logic behind it is worth paying attention to: this isn’t a company’s earnings blow-up; it’s the macro environment changing.
Rates and geopolitics are both applying pressure. If long-term yields keep moving higher, the valuation “center” for semiconductors may need to adjust again.
In the short term, it’s a question of whether the selloff has flushed enough emotion. In the medium term, it’s about the direction of interest rates and how geopolitics evolves. Don’t rush to bottom-fish, and don’t panic either.
These are my personal observations; data should be verified against official disclosures.

In the AI sector, most tokens prop up their prices with narrative, and Venice is one of the few that speaks with real revenue.

On August 17, the founder, Erik Voorhees, announced that Venice.ai’s annualized revenue has surpassed $100 million.
A month ago, the figure was still $70 million.
After the news broke, $VVV surged 8%-10% within 24 hours, and trading volume increased noticeably.

What Venice.ai is doing isn’t complicated: it offers open-source AI services—text, image, and code generation—but the core selling point is privacy.
No user data is stored, and there’s no censorship.
It runs on the Base chain, founded by Erik Voorhees (that old OG from ShapeShift).

In tokenomics, there’s one design I think is pretty solid: platform revenue is directly used to buy back and burn VVV. More than 40% of the supply has already been burned, emissions are still steadily decreasing, and the goal is net deflation.
Staking lets you earn rewards, unlock Pro features, and also mint DIEM (daily AI inference allowance).

The price is currently around $12–13.

———

Monthly revenue jumped from $70 million to $100 million in one month. This kind of growth is hard data by AI token standards. It’s not painting a dream—it’s real people paying for its services.

The deflationary mechanism also directly ties revenue growth to token supply contraction.

Of course, crypto is volatile, and whether this growth can continue is another question. Personal observation—DYOR.

USD1 This week I got something really important:
A banking license.

On August 14, the U.S. Office of the Comptroller of the Currency (OCC) granted World Liberty Trust Company’s National Trust Bank license an initial conditional approval.

In plain English: USD1 can now issue and redeem on its own and manage its reserves. No need to rely entirely on third-party custody anymore (previously, most of this was handled by BitGo). Get the license and do it yourself.

This isn’t a typical commercial bank. It doesn’t take deposits or make loans. It’s a federally regulated entity dedicated to trusts and digital asset custody.

Reserves include cash, short-term U.S. Treasuries, and money market funds—things like that.

//

Why do I think this step matters more than the previous exchange listing or getting some celebrity endorsement?

In the stablecoin business, it ultimately comes down to one thing: whether institutions dare to put big money in.
What do institutions look at before they move money in?
Licensing, the regulatory framework, reserve transparency, and the redemption mechanism.

No matter how strong your political narrative is, without these, big capital won’t touch you.

Right now, USD1’s market cap is around the $4 billion range—about fourth in the stablecoin space. It’s deployed across multiple chains (Ethereum, BNB Chain, Solana, Tron).
Before this, it relied on the Trump label and topic-driven traffic to surge.

But after getting this license, it now has the legitimacy to compete on the same level as USDC.
Circle already has a New York trust license, and now World Liberty has a federal trust license too.

The race moves from “who has the toughest background” to “who’s more fully in a proper compliance framework.”

————

Of course, this is still only an initial conditional approval for now.
It must meet a minimum $20 million in Tier 1 capital and a series of compliance requirements. Even before final approval, it could still be revoked or changed.

We’re not at the finish line yet, but the direction is already very clear.

Earlier, I wrote about USD1 pulling in Cb’s people to become CBO. Now it has also obtained a federal license.
Looking at the two steps together, this project really is moving from political meme status toward becoming proper financial infrastructure.
Whether it will go all the way is another question. But at least each step is moving in the right direction.

The above content is for information disclosure only and does not constitute any investment advice. DYOR.
#USD1

Everyone is asking: Who will AI disrupt?

I recently turned the question around: Who is an AI company that can’t be defeated?

The answer is a class of companies called HALO.

Heavy Assets, Low Obsolescence — heavy assets, low replacement rates.

Mines, power generation equipment, offshore engineering vessel fleets, power grids, heavy machinery, logistics vehicle fleets, outdoor billboards.

The moat isn’t code and patents—it’s steel, land, ships, and factories.

You can’t use big models to mine minerals, and you can’t use GPT to lay undersea power cables.

And these companies are even the ones taking the money from AI, in reverse.

Do data centers need to be built? Then you have to buy backup generators, cooling systems, and upgrade the power grid.

If mining infrastructure needs to expand, then you have to buy heavy machinery.
They aren’t the enemies of AI—they’re the suppliers of AI infrastructure.

————

Two names I find interesting but not mainstream enough:

Finning International (FTT)

One of the world’s largest Caterpillar dealers. Data center primary and backup power demand directly becomes its orders. Its power business backlog is close to C$1.2 billion, while it also locks in the copper mine and oil & gas expansion cycle.

Subsea 7 (SUBC)

An offshore engineering contractor. It has a scarce fleet of cable-laying ships, with a backlog of $13.6 billion. If underwater data centers really take off—
seawater cooling, subsea power, and fiber-optic cable installation—
it’s all its work.

Others that are often included in this framework: Cummins, Eaton, Vertiv, Southern Copper, Lamar Advertising.

Goldman Sachs and BlackRock have already treated HALO as an official allocation direction this year. Roundhill even launched a dedicated ETF (ticker LOHA).

————

This kind of business doesn’t tell stories—it’s all about orders and backlog.
Valuations are a notch lower than tech stocks, and they get less attention.
But the combo of being hard for algorithms to eliminate and still getting paid from AI infrastructure makes it scarce in today’s market environment crowded with AI software.

Heavy-asset businesses have strong cycles, and volatility won’t be small.
Personal observation—data is based on official disclosures; DYOR.

#美股 #AI