Powerpei
Powerpei
I will share some of my experiences here, welcome to follow me, I will also reply to build the OKX planet together
1.6KFollowing
1.4Kfollowers
Feed
Feed
The market has finally caught its breath these past few days
Looking at the screen full of green bars
Various groups started flooding the chat again, shouting that the bull is returning quickly
Everyone is watching how much blood their altcoins have recovered
But I am watching another floodgate of capital outflow
Yesterday I came across a PR report from Bitget official:
They do forex and gold CFDs, and their daily trading volume surprisingly reached 8 billion USD
Let's not discuss how much fluff is in this official announcement data
But at the moment when the crypto market is warming up
Such huge capital not rushing into memecoins
Instead, crowding into trading traditional assets, this itself is very counterintuitive
If you carefully consider the capital flow behind this
Whales holding massive amounts of USDT
If they want to hedge a bit in US stocks or gold during this macro volatility
Using traditional brokers’ T+2 settlement and cumbersome capital checks
They would practically be asked to strip down to their underwear
Exchanges understand these big funds’ pain points very well
They directly use a unified margin model
Forcibly bringing the traditional finance desk into crypto accounts
No need to withdraw funds, no need to exchange currencies, just use U to open positions directly
My personal view is that crypto giants are taking advantage of liquidity recovery
Frenzily stealing market share from traditional brokers
Whether you play these cross-border derivatives or not, one trend is clear:
Web3 funds are no longer satisfied with fighting in their small pools
Using the extreme efficiency of crypto channels
To inversely devour the huge liquidity of traditional finance, this is the ultimate big casino that these smart money players are laying out.
Note: The above content is for informational reference only and does not constitute any investment advice!!
AMD's Q1 earnings report last night was like throwing a fire into a haystack
Revenue broke through $10.3 billion, with data center business surging 57%
This extremely violent liquidity not only forcibly pushed NVIDIA back to a $5 trillion market cap
The cross-market spillover effect is astonishingly fast
Today, Hong Kong stocks like SMIC and Hua Hong Semiconductor directly soared
Southbound funds swept over 910 billion in buying volume
Confirming that capital is frantically searching for semiconductor undervalued spots across the entire network
Seeing the screen full of surges, many people emotionally shout that the bull market is back
If we strip away this noise and look at the underlying data
The foundation of this rally is actually very solid.
Macroscopically, the ceasefire news in the Middle East directly hammered oil prices down by more than ten percent
The shadow of inflation rebound has been temporarily forcibly dispelled
Looking at the micro industry side, capital expenditure in the AI inference era has become an open card
Now the big players are competing not just on GPUs
Severe capacity mismatch between CPUs and HBM memory is triggering real supply-demand panic in the supply chain.
Capital is pouring down this hardest industrial logic chain, leaving no room for shorts.
But the purer this logic is, and when everyone thinks "buying means making money"
The risk level on the market is even higher
When fundamental benefits are quickly overdrawn in a very short time
Crowded trading channels inevitably plant the seeds for a sharp high-level shakeout
A trend of this magnitude is indeed worth using a small position to ride the premium
But heavy buying at this point is purely like putting your neck on the chopping block
My current move is to only keep the core leaders' base positions
Firmly hold over 30% cash exposure
Let the frenzy money be earned by others
When this wave of sentiment is exhausted and a real structural correction is hammered out
The cash in hand will be the only proof to hold chips at the bottom.
#HongKongStocks #USStocks #AIsemiconductors
Powerpei
The May Day holiday just ended, and I’m finally coming to my senses after the traffic jam on the highway.
As soon as I opened my computer, I was slapped in the face by the Hong Kong stock market's opening surge.
On Twitter, people are still arguing about the AI singularity every day.
The grand narratives of giant valuations are hanging in the air.
But today, the money in the market is extremely real.
It directly ran to vote for the most struggling manufacturing sector.
➤ Xiaomi surged nearly 7 points, and semiconductor stocks like SMIC and Hua Hong also took off.
The logic is very straightforward: after being battered by geopolitical risks and regulatory washouts several times, smart money has long stopped looking at PPTs.
If you can deliver 30,000 cars in a month, if you have real chip orders, the funds will flow to you.
But to be honest, looking at these bullish candles, I didn’t feel excited; instead, I felt an extreme sense of defeat.
Because I realized that as a person, I am becoming more and more like a useless player in this game.
——————
I just went to check the backend logs of my self-developed Web3 monitoring tool.
In the seconds before the spot market moved, my AI Agent had already captured abnormal capital flows from the Perp order books of Hyperliquid and Aster’s tokenized stocks.
It directly issued a warning in natural language.
While I was hesitating in front of the screen about whether to chase the high, still weighing the situation in the Middle East in my mind,
the machine had already calculated the slippage without emotion and made the risk control decision.
This is the most despairing gap.
You think you are competing with the big players,
but in reality, you are fighting against code that responds tirelessly in milliseconds.
In a volatile market, those few seconds of early reaction from the machine save you the principal you would have lost.
I’ve completely given up on the idea of "manual trading".
These days, I don’t want to do anything else; I just plan to focus on iterating my AI Agent’s cross-market monitoring and risk control modules.
➢➢➢
The trend has already changed.
Tools serve people, and people are in turn being eliminated by tools.
For those still relying on their hands to draw support lines and chase hot stocks based on market feel, good luck to you.
The money you lost on slippage will ultimately turn into electricity fees for the algorithms.
The clumsy methods in the past of relying on deploying offline hardware and scanning biometric features for identity verification
Everyone knows why they haven't gained traction over the past few years
Handing over irreversible biometric information to a hardware black box
This inhumane promotion approach is doomed to struggle
In the identity verification business, the ultimate winner will most likely be the low-barrier pure software path
This explains why Billions Network ($BILL), which just launched its token on May 4th, has an unusually high funding heat
---
I thoroughly reviewed the fundamentals of this project
>The teams behind Polygon ID and Disco abandoned hardware
>They directly use mobile devices combined with ZK (zero-knowledge proofs) for the identity layer
In this year’s environment where Deepfake is rampant and AI Agents will soon outnumber real people online
They are actually collecting a very scarce trust tax
➤ The core business logic is very straightforward:
Grant machines property rights
Allow every bot online to be traceable back to a real boss
While cutting off the risk of underlying privacy leaks.
Top-tier essential narrative, backed by Polychain and 2.4 million real verified users, the credentials are impressive
➢ From a trading perspective, the current circulating market cap of $BILL is around $150 million, with an FDV of $650 million
---
Many people in the group have already started calculating odds based on a $1 billion market cap target
Frankly, combined with the relatively slow unlocking schedule and the current AI trust demand
It indeed has the framework to touch that red line
However, the early presale phase experienced adjustments in lock-up rules
Indicating that the capital game behind it is very real
This kind of initial chip structure collision requires a wave of real market turnover to solidify the foundation for an upward push.
---
You can hardly treat it as a blind all-in gamble
This is purely an infrastructure asset that requires strict calculation of risk and reward
➢ Its value capture mechanism (staking for reputation, payment verification fee discounts) is logically closed-loop
But whether this system can convert into sustained buying pressure in the secondary market
Heavily depends on the willingness of third-party apps to actually integrate and pay
If no one genuinely calls it, this economic model is just a piece of worthless paper
The initial token launch window naturally amplifies volatility
Instead of blindly chasing highs at the peak of emotions, it’s better to slow down the pace a bit
The sector logic is sound, but I personally prefer to wait for the initial chips to settle
→ Meanwhile, observe its real on-chain call volume later
Set a comfortable mental position with an alert, and let the fundamentals verify the price.
Note: The above content is for informational reference only and does not constitute any investment advice, DROY!
It seems Justin Sun has really hit a hard wall this time.
I just glanced at the public court documents, and @worldlibertyfi has directly sued Justin Sun in Florida for defamation.
This drama is a bit overwhelming for me.
Usually, conflicts in the crypto space at most involve posting long tweets or hosting a Space to exchange family greetings.
But WLFI didn’t waste any words and slapped a lawyer’s letter right in his face.
The cause of the issue is actually extremely simple.
At the end of last year, Justin Sun made a large purchase of WLFI tokens, and the sales terms clearly stated that transfers were prohibited at this stage.
But he stubbornly tried to operate anyway, precisely triggering the freeze mechanism in the smart contract, and the funds were directly locked.
Accepting the bet and its consequences should be common sense.
Yet he immediately used his influencer status to launch social attacks, even publicly threatening to crash the token price.
He kept shouting that the project team left a backdoor to commit fraud.
This is highly ironic.
Everyone constantly talks about “Code is Law.”
But when the code ruthlessly enforces the rules and freezes a whale’s funds,
the whale’s first reaction is still to use their media privilege to overturn the table.
WLFI’s counterattack this time serves as a wake-up call for the entire community.
These people are backed by traditional political resources and old money capital.
They have no interest in playing the “decentralized community governance” game or any of those half-hearted tricks.
I honestly don’t care who wins or loses this lawsuit in the end.
What I’m really looking forward to is the evidence exchange (Discovery) during the court trial.
When both sides are forced to lay all their cards on the table, it will definitely reveal on-chain information that’s ten times more exciting than what we have now.
Going to make a cup of coffee and wait for trial updates.
#WLFI #DeFi
Powerpei reposted
BNB Chain's Data and Layout in the AI Agent Track
Recently, I saw a set of data: on January 1, 2026, the number of AI Agents on BNB Chain was 337. Four months later, this number is 75,957.
No matter how you view this ecosystem, a growth rate of 50,048% is definitely worth noting.
According to data from 8004scan, BNB Chain currently ranks first with 75,938 registered Agents, surpassing Ethereum and Base, accounting for more than one-third of the ERC-8004 deployment share.
What are they doing?
From the tech stack perspective, BNB Chain has launched several tools for AI developers in recent months:
bnbagent-sdk: a Python SDK that enables Agents to have wallets and order-taking capabilities (based on ERC-8004 + APEX)
bnbchain-mcp: allows AI tools like Cursor and Claude to operate on-chain data using natural language
bnbchain-skills: plug-and-play profiles that teach AI Agents how to use BNB Chain
In short, they are lowering the threshold for AI developers to access on-chain infrastructure.
You don’t need to understand crypto; just pip install and you can add payment and identity capabilities to your Agent.
Why now?
The timing is indeed critical. OpenAI is promoting the Apps ecosystem, Anthropic is pushing the MCP protocol, the ERC-8004 standard is just beginning to be adopted, and the infrastructure for the Agent economy is being defined.
BNB Chain’s strategy is not to educate crypto users to use AI, but rather the opposite—to let AI developers unknowingly use on-chain capabilities. At least logically, this direction makes sense.
Actual implementation
From hackathon projects to deployed Agents, the use cases are indeed expanding.
But frankly, it’s still too early to judge the quality and actual usage of these Agents. Growth in numbers is one thing; how many truly valuable applications emerge is another.
However, one thing is certain: in scenarios where Agents need on-chain identity, autonomous payment, and verifiable work capabilities, Binance, as an industry leader, now provides relatively complete infrastructure.
As for whether it can become the "default choice" for AI Agents, we’ll have to see how the ecosystem develops over the next 6 months to 1 year.
Currently, the data is there, and the tools are gradually coming online. As practitioners, it’s worth continuously monitoring progress in this direction, after all, the big direction is AI! The big trend is also AI! #AIonBNB
@0xsandy1913 @BNBCHAINZH
In the current market, if a Perps project has been online for just a few months and its TVL has hard-capped at over 100 million USD, it likely did two things:
Either it’s madly distributing points in preparation for a run, or it has truly found the Achilles' heel of the market makers and whales.
I went through @StandX_Official's operations and found that they are taking the extremely rare second path.
Looking closely at this 450 million in daily trading volume, it’s not something that retail investors with a few hundred or thousand USD slowly grind out.
The bulk of this actually comes from their SIP-1 large transaction module.
---
The on-chain environment is too dirty right now.
Even a few tens of thousands of USD orders can get squeezed by MEV.
StandX has created this execution mechanism with zero slippage at any scale.
In simple terms, they have carved out a VIP highway on-chain.
Those massive funds that dare not move elsewhere have finally found a safe outlet here.
But it’s easy to lure in big players; how do you lock their money on the platform?
They’ve taken a knife to the margin.
When we usually trade contracts, the margin we put in is like a lamb waiting to be slaughtered.
But here, even if you have an open position, the DUSD used as margin is still being drained in the background.
>It stitches together the market's Base Yield,
>position returns,
>and platform fee dividends,
to create a hard-hitting combined APY of around 10%.
This doesn’t require you to click any Staking buttons.
The capital efficiency has been squeezed to an outrageous level.
---
➤ Yesterday, even Fiona from the former IOSG personally stepped in to manage their growth.
This veteran, who has crawled through the capital circle,
gave up comfort to take on this hardcore business.
What she sees is definitely not a simple Fork imitation,
but a clear path for capital reconstruction.
Stop guessing about the market's ups and downs.
Look at those infrastructures that can make money while you sleep,
where big orders can drop without making a splash.
That’s the gentle place where big funds truly dare not leave.
The May Day holiday just ended, and I’m finally coming to my senses after the traffic jam on the highway.
As soon as I opened my computer, I was slapped in the face by the Hong Kong stock market's opening surge.
On Twitter, people are still arguing about the AI singularity every day.
The grand narratives of giant valuations are hanging in the air.
But today, the money in the market is extremely real.
It directly ran to vote for the most struggling manufacturing sector.
➤ Xiaomi surged nearly 7 points, and semiconductor stocks like SMIC and Hua Hong also took off.
The logic is very straightforward: after being battered by geopolitical risks and regulatory washouts several times, smart money has long stopped looking at PPTs.
If you can deliver 30,000 cars in a month, if you have real chip orders, the funds will flow to you.
But to be honest, looking at these bullish candles, I didn’t feel excited; instead, I felt an extreme sense of defeat.
Because I realized that as a person, I am becoming more and more like a useless player in this game.
——————
I just went to check the backend logs of my self-developed Web3 monitoring tool.
In the seconds before the spot market moved, my AI Agent had already captured abnormal capital flows from the Perp order books of Hyperliquid and Aster’s tokenized stocks.
It directly issued a warning in natural language.
While I was hesitating in front of the screen about whether to chase the high, still weighing the situation in the Middle East in my mind,
the machine had already calculated the slippage without emotion and made the risk control decision.
This is the most despairing gap.
You think you are competing with the big players,
but in reality, you are fighting against code that responds tirelessly in milliseconds.
In a volatile market, those few seconds of early reaction from the machine save you the principal you would have lost.
I’ve completely given up on the idea of "manual trading".
These days, I don’t want to do anything else; I just plan to focus on iterating my AI Agent’s cross-market monitoring and risk control modules.
➢➢➢
The trend has already changed.
Tools serve people, and people are in turn being eliminated by tools.
For those still relying on their hands to draw support lines and chase hot stocks based on market feel, good luck to you.
The money you lost on slippage will ultimately turn into electricity fees for the algorithms.
Powerpei
Recently, aside from writing some scripts, I've basically been glued to the Hong Kong and US stock markets.
But the recent trends in the AI sector have left me a bit confused.
➢ First, let's talk about the Hong Kong stocks, like SenseTime (0020), Zhizhu (02513), and MiniMax, this batch of AI newcomers.
The recent script seems to be cut from the same mold.
As soon as there's news of new financing or a new model release,
the opening is definitely an enticing big bullish candle, only to be smashed down during the day.
This frequent occurrence of high opening and low closing
is clearly old money taking advantage of retail investors' faith to frantically distribute shares.
>>>>>
➢ Turning to the US stock market, the situation is even harsher.
Microsoft and Meta just announced a combined capital expenditure (CapEx) plan of $650 billion for 2026,
all poured into AI infrastructure. And the result?
Aside from the rock-solid Nvidia (NVDA) being able to actually take these orders,
the other giants faced a direct drop during the day.
Those Wall Street folks are quite clever.
In the past, when you talked about big model stories, they were willing to buy in.
Now they just pull out their calculators and ask one question: you spent so much money on cards, when will this ROI (return on investment) actually show cash?
If you can't answer, no matter how big your brand is, they'll still smash your stock.
This is actually the most brutal market bottom line this year:
The honeymoon period of speculating concepts based on a parameter arms race is completely over; now it's the bloody verification stage.
---
➤ Since the market wants to see real cash, my operational discipline must also be ruthless.
I've thrown away all my beliefs from last year and set extremely cold-blooded rules for myself.
I've directly blacklisted those who still post daily press releases boasting about big model parameters.
That stuff is already everywhere.
This year, the only ones that can really rise are the agents that have completed the business closed loop and embodied intelligence.
You see, Zhizhu and MiniMax recently showcased an agent landing demonstration,
and the sector immediately gave a nod, rising over 20%.
---
➢ But my operations remain heartless:
As long as you can't produce real enterprise-level purchase orders,
even if the demo looks like a Hollywood sci-fi blockbuster,
I still treat it as just a wave of emotional premium.
If there's profit, I'll sell; no holding back.
As for those big companies in the US stock market, I don't even look at how much computing power they claim they want to buy anymore.
During the recent earnings season, I personally witnessed Amazon's stock drop 6% to 10% because of high CapEx without seeing a return.
In contrast, Alphabet, because its cloud business's gross profit really improved, managed to hit a new high against the trend.
This tells me a hard rule: if capital expenditure doesn't turn into profit within two or three quarters, the market will ruthlessly restructure the valuation.
Now when buying tech stocks, I only look at AI revenue growth and gross margin.
The last and most painful lesson is to control my hands.
The flash crashes and rotations in the AI sector over the past few months have been terrifyingly fast.
I've personally experienced a situation where good news came out, and the next day it opened high and then dropped 10%.
So now, my position in a single stock is firmly capped between 5% and 8%, no heavy bets.
I've even forced myself to develop a habit:
> Whenever there's an earnings report or a major catalyst approaching, I don't care about anything else and first reduce my position to lock in profits. I always keep over 30% cash on hand. The geopolitical games and energy bottlenecks outside can explode at any time.
Having cash means you're the hunter during a big drop.
If you're fully invested and holding on, you're just someone else's prey.
Mr. Market never tells you about a ten-year bull run; he only values those companies that truly generate cash flow.
Enough said, I need to go re-evaluate those Hong Kong stocks I have; any without order flow will be cleared at the opening.
Just now, I calmed down and went through the terms of the WLFI token reconstruction proposal.
To be honest, it's understandable that long-time players who have waited a year and a half are faced with the choice of re-locking their assets.
It's normal for people in the community to vent their frustrations.
But I actually think this situation is a sign that the bad news has been fully priced in.
The team directly burned 10% of their share and proactively extended the release period.
This attitude is quite clear.
They just want to clean up the historical burden hanging over their heads all at once.
There's no choice; the current market cap of USD1 is too solid.
A stable market cap of 4.2 billion.
Zebec is using it to pay salaries.
AI agents are also on board.
The roadmap still aims to push towards Solana and Base.
With the front-end business running so strong,
if the back-end token model remains a ticking time bomb,
institutional funds won't dare to make heavy bets.
Although the handling this time seems a bit rough, it's better to take the short pain than to endure long-term suffering.
Washing the chips into the hands of those willing to hold long-term is definitely more reassuring than having to guard against selling pressure every day.
Only when the foundation is repaired can the car be driven fast. I'll keep an eye on its real cash flow after landing on several new public chains.
