追风Lab

Seeing Aethir update its tokenomics feels oddly familiar, especially since the airdrop back then was pretty generous.

The main thing this time is that ATH is finally trying to link revenue from its real computing-power business to the token’s value. After all, the biggest risk for DePIN projects is talking about business growth while relying on token emissions to keep the network going. Revenue from real customers is an important indicator of whether a project can go the distance.

Of course, burning tokens doesn’t guarantee a price increase. ATH’s existing emission schedule hasn’t changed, so whether this can improve the supply-demand balance still depends on the volume of real orders, the network’s revenue, and whether the amount burned can offset new supply.

At the moment, $AGPU is also an Aethir customer and holds a large amount of ATH. That gives me another reason to be hopeful.

If AGPU’s AI computing business continues to gain traction and generate more real demand, while Aethir’s revenue and ATH burn mechanism gradually get up and running, then ATH’s performance could theoretically have a positive impact on the value of the ATH assets held by AGPU.

In an ideal scenario, the two businesses would boost each other and create a virtuous cycle—kind of like stepping on one foot with the other to climb upward.

But this is only an idealized scenario. Whether it can happen depends on computing-power delivery, actual revenue, and token supply-and-demand data—not just wishful thinking.

As someone who’s been left holding the bag on both of these before, I’m obviously hoping they can make this kind of move. But more than short-term price swings, I want to see how much real revenue actually enters the burn mechanism after it launches in November, how many tokens are actually burned, and how many sustained orders the additional computing power can bring in.

Announcements can talk about expectations, but on-chain data and real revenue are the final answers.

For now, I’ll keep watching and verifying. I hope this time it’s not just a pipe dream, and that they can truly connect business growth with token value.

As of now, we can confirm that quite a number of L2 / Rollup projects have ceased operations, but the reasons vary greatly, including:
· Proactive shutdown / strategic pivot
· The economic model cannot be sustained
· Shutting down after being attacked
· Bankruptcy
· Team disappears / suspected rug-pull / running off

And our Blast, from 2B➡️32M, is mainly shut down for the same reason:

L2 operating costs > the revenue generated by L2

This actually points to a very real issue:
TVL ≠ revenue ≠ business model

If we reclassify by “cause of death,” it becomes even more interesting:
① The economic model couldn’t hold up
Blast, Botanix, Redstone

Shared characteristics: technical capability, funding, and users/TVL—but the real fee revenue cannot cover the ongoing operating costs.

Blast is especially典型：once it had $2B+ TVL → around $32M, and ultimately the official team directly admitted that the revenue could not cover the costs.

② Hacker / security incidents

Kinto isn’t really about “not having users,” but rather:
Exploit → asset losses → token collapse → debt / funding pressure → shutdown.

③ Strategic pivot
Aztec Connect, Fluence, Sophon, ZERO

These shouldn’t be called “running off.” A more accurate description is: giving up maintaining their own L2 and shifting resources to businesses with higher commercial value.

For example, Fluence directly shut down its L2 and migrated assets back to Ethereum; Sophon shifted its chain-migration strategy toward Base applications.

④ Bankruptcy

The most worth listing separately right now is: Movement Labs

This is strictly a company-level Chapter 11 bankruptcy, but it’s also not the traditional “exit scam/rug-pull with cash.”

⑤ Suspected rug-pull / ran off

Gemholic / GemSwap: the project previously locked about 921 ETH due to smart contract issues; after funds were restored, the team transferred approximately $3.3M in assets and deleted/disappeared their social accounts, and was then hit with rug-pull allegations.

In this market, it’s really not easy to survive—every month and every week, projects across various tracks stop operating. As the saying goes: watch your pocket, and safety should always come first!

Mid-Autumn Moon Viewing, and Binance also rolled out a “Moon Landing Plan.”

In plain terms, it’s not complicated: complete the event tasks to share in the 10,000U, plus there will also be 170 random prizes—30 custom Mid-Autumn shoes, iPhone Duo, Apple Watch, and more, along with cash red packets and merchandise...

I just received the shoes this morning—Mid-Autumn limited. Wearing them out really feels ceremonial. The rules are also simple: repost this main post, and include the event link and hashtags—#BinanceMidAutumnMoonLandingPlan

Mooncakes can be eaten a bit later; I got the tasks done first. For events like this, not joining would be a waste. First, help yourself step on the Moon’s surface—whether you win or not depends on your luck.

I took stock of the exchange hack incidents that have occurred over the years. What has been publicly disclosed is only a portion of them—some incidents have not been fully revealed to the public even to this day.

When you look at the past few years, most CEXs have, to one degree or another, experienced security incidents. The difference may lie in whether they have the ability to withstand the fallout, and whether they can stabilize users’ losses and confidence.

When it comes to CEX security, nobody can guarantee that problems will never happen. What truly matters might be whether, after something goes wrong, the platform has sufficient reserves, the right mechanisms, and the sense of responsibility to handle the situation properly.

After all, for users, security is never just a slogan—it’s whether the platform can hold up when it matters most.

I’ve recently been paying more attention to a product by DeAgentAI: the AI Token Smart Router. Many people are still asking, what exactly is the AI Token Smart Router?

In simple terms, it’s not building yet another new large language model. Instead, it consolidates the different AI models that enterprises previously had to integrate and call separately into a single unified entry point.

Previously, companies had to connect to multiple models, which often meant working with different vendors. Now, the Router is more like an “AI model master router”—it unifies multi-model calls so that enterprises can choose, measure, and reconcile more easily. Even more importantly, this product has moved beyond just a slide-deck stage.

According to data recently disclosed by the official team, during the trial operation phase, the AI Token Smart Router already had B2B customers using it. The latest official disclosures also show that since the start of its internal testing, it has served 8 B2B customers, processed more than 54 million external model calls in total, and monthly revenue has already exceeded $1 million. The product has now also been officially opened to B2B customers.

So I think what’s truly worth watching now is:
How many do these businesses actually use?
Can the Router keep generating revenue?
And how does that revenue ultimately connect back to $AIA?

That’s also what I find particularly interesting about the relationship between AI Token Smart Router and $AIA.
It’s currently following a fairly direct value chain: enterprises call AI → the Router generates commercial revenue → protocol/business revenue is generated → part of the funds is used for $AIA buybacks and burns.

DeAgentAI has also previously announced a mechanism for $AIA buyback and burn based on real protocol revenues.
So if, going forward, the number of the Router’s B2B customers, the volume of model calls, and revenue can continue to grow, then it’s effectively turning “AI usage volume” into “real business revenue,” which in turn creates value inflow back into $AIA.

Personally, I’m keeping a close eye on three data points:
① Whether the number of B2B customers continues to increase
② Whether the volume of model calls can keep rising beyond the tens-of-millions range
③ Whether the revenue → buyback/burn of $AIA link can continue to be realized

This is easier to verify than just telling the narrative of an AI project.

In plain words, what the AI Token Smart Router really needs to prove is one sentence: whether there are real enterprises using it, and whether these real calls can ultimately translate into real revenue.

After recently looking at Binance’s TradFi Perps, I’m increasingly convinced that competition has shifted from “who can launch first” to “who can make the market deeper.”

Take SNDKUSDT as an example. Binance’s one-way OI is about $626 million, accounting for roughly 59.5% of the selected major venues—about 3 times that of the second place. Over the past 7 days, trading volume was about $16.7 billion. With both OI and volume leading at the same time, it indicates clear advantages in both capital staying power and trading activity.

Behind this is a classic scale effect: more assets → more users → deeper liquidity → it’s easier to make new listings deep as well.

Today, Binance already covers 135+ TradFi perpetual trading pairs, spanning everything from U.S. stocks, Hong Kong stocks, Korean stocks, and Japanese stocks to commodities, and it’s further expanding into Chinese tech assets such as CXMT.

So the real competition in TradFi Perps isn’t about “who can launch,” but rather who can truly concentrate capital, trading activity, and liquidity.

Three ancient projects share a common trait: they launch an IPO only to fail, then return to announce their own token, and afterward restart another PUA cycle. Instead of properly developing their own lane, they insist on doing things they’re not good at. In the end, they end up with a “四不像”....

OpenSea was founded in 2018, with its main track being NFTs. In October 2025, it announced SEA, and then listed on a DEX to kick off “Odyssey.” The number of team members who resigned exceeds 20!!

Polymarket was founded in 2020, with its main track being prediction. In May 2026, it hinted at POLY. Soon; In April 2026, it launched Perp. The number of team members who resigned exceeds 15.

Metamask was founded in 2016, with its main track being wallets. In October 2025, it confirmed Metamask, and then launched an points rewards program—its wallet includes a cross-chain bridge, DEX, Perp... The number of team members who resigned exceeds 9.

Harmony is ruined. I originally thought it would just be made into 4 billion, but now it’s 3 trillion. The last time something reached the level of billions and beyond—when I thought of $LUNA

Another long-established Layer 1 blockchain, Harmony, has a bug again—$ONE plummets by dozens of points!

The attacker exploited an unoccupied block to mint 4 billion $ONE without authorization. 97% of the freshly minted tokens have already been transferred to a CEX to dump the price, causing a massive crash. The project team has urged CEXs to freeze several transfer addresses in an attempt to regain control of the situation. However, the community is the one most harmed.

$ONE hit an all-time high price of 0.38U, but now it’s only 0.0008U—a drop of over 99%. What does that mean? If someone bought $10,000 at the peak, there’s now only about 20 left!

In the past three years, there have been 322 documented cases involving Web3, spanning 57 countries. Of these, 101 involved kidnapping, 69 involved robbery, and 12 involved murder, with total losses exceeding $625 million.

In some cases, the most dangerous thing isn’t your private key being exposed. It’s your wealth, identity, and address all being exposed at the same time.

Wallet security does not equal personal safety.

A truly mature asset security system for the future should include at least four layers:
First layer: Asset security
· Cold wallets, multisig, MPC, permission isolation.

Second layer: Account security
· 2FA, hardware keys, independent devices, email isolation.

Third layer: Privacy security
· Reduce public exposure of wealth, addresses, home addresses, and family information.

Fourth layer: Personal safety
· Don’t allow on-chain assets to form an overly clear mapping to your real-world identity, residence, and family members.

It’s undeniable that TEDA really has money. Just this year’s Q2 alone, it bought 14 tons of gold.

It has now become the world’s largest private holder of gold, with a total amount of 146 tons. Its value is $18.8 billion. As of now, the amount of circulating USDT is about 18.355 billion USDT, and the total supply (infinite) is about 18.909 billion USDT. Essentially, it’s already pegged to the value of gold.

With so much gold, even if you use trucks with a 10-ton capacity, it would take 15 vehicles to transport it all.

SpaceX’s 2Q financial report in 26 years: Revenue hits $7.814 billion, up 92% year over year; net loss narrows to $541 million, improving a lot versus the same period last year; adjusted EBITDA $3.538 billion, up 191% year over year.

All three metrics indicate that the core business is already very healthy. Starlink, AI, and rocket launches all grew.

At present, SpaceX is operating on a three-wheel drive structure: “Starlink earns money to support the family + AI is expanding aggressively + rockets continue to burn cash to work on Mars.”

This evening, the first batch of 910 million shares from SpaceX will be unlocked. The float will expand from 640 million shares to 1.55 billion shares! Can they drop the price to give retail investors a chance to get on board?

Saw two resignation announcements early. Coincidentally, both are stepping down from their current roles and continuing on as consultants.

1: X platform product lead Nikita Bier resigns as X product lead and shifts into a consultant role

2: Mysten Labs co-founder and CTO Sam Blackshear will leave Mysten Labs and join Anthropic to work on defensive security research. He will continue to serve as a consultant for the Mysten and Sui ecosystems, and participate in the Move Foundation that’s currently being prepared.

Everyone should thank Sam Blackshear for his vision—so that $SUI can become the stuff of legend. No need to do anything else; just join DC early to qualify.

Such opportunities are rare—hope to see you again on the next project!

These past two days, the technology stocks and the overall market are really going crazy! Buoyed by the strong positive news from both Q2 revenue and adjusted earnings per share all beating expectations across the board, the stock price of Big Bull Palantir surged by nearly 30% in after-hours trading and the following day, directly igniting market sentiment—once again proving that AI remains one of the global capital’s most watched main themes.

Worth noting is that funds have not only poured into the US stock spot market, but have also rapidly flowed into the crypto market’s $PLTR contract trading. CoinGlass data shows that PLTR’s total open interest across the network is approaching $38.2 million, with tradeXYZ, Binance, and Bitget ranking in the top three by open interest, together accounting for roughly 64% of the market share, forming the core tier. Looks like many people have already positioned themselves early with contracts or chased the rally.

The boundary between traditional finance and the crypto market is becoming even more intertwined. When a popular US stock issues good news, capital can complete price discovery and risk hedging on-chain 7×24 hours a day, without being constrained by traditional trading hours. Perhaps this is the direction global asset trading will move toward in the next phase.

Lately, it seems like everyone’s jumping into US stocks and Hong Kong stocks. Binance Wallet just dropped a big move, launching a 'Double Fee Waiver' event specifically for ONDO’s tokenized securities: service fee 0 + on-chain gas fee 0!

Binance really knows its users; wherever the hype is, that’s where the service goes. ONDO itself is a representative of bringing traditional assets like real estate and bonds onto the blockchain, marking it as a new sector. Now, you can enjoy stable yields from RWA without worrying about fees.

Of course, after the 2-month promotional period, they will start charging gas fees + service fees again. While the event is still live, I’d recommend checking it out. Saving a bit is still saving, and getting in early might net you some extra rewards.

Binance's new client app is super convenient after giving it a spin.
As an old 'hodler', my expectations for the desktop version are pretty straightforward: stable, fast, and comprehensive. Honestly, the previous version had me feeling a bit 'mixed feelings' with disconnects and delays really messing with my trading psyche. Since the update, I haven't experienced any of those issues!

The most noticeable change this time is the smoothness. The dreaded lag, delay, or unexpected crashes seem to have vanished after this update. The optimization of the underlying architecture has made the whole operation logic super responsive; that millisecond-level responsiveness is exactly what a professional tool should deliver.

For those with asset allocation needs or writing APIs for strategies, the restoration of sub-account functionality is like a 'much-needed rain after a drought'. Now you can manage account permissions, transfer assets, and set up APIs all in one place on the desktop, eliminating the need to jump back and forth between the web and the app, which feels really empowering.

The Pay feature now has a quick access point in the navigation bar, and the payment process has been streamlined to the max. This actually reflects Binance's broader direction of evolving into a super financial app.

The only downside is that the client seems unable to take screenshots of candlestick charts, which could use some optimization.

Binance has always focused on enhancing the user's fundamental experience, and this comprehensive performance boost for the desktop version, along with the addition and refinement of features, showcases Binance's grand vision of creating a 'super financial app' by honing the user product experience across all platforms. This enables users to enjoy convenient access while also embodying the principle of financial equality, lowering the barriers to product usage.

If you're interested, you can go download and give it a try
https://binance.com/zh-CN/download?utm_source=copylink&utm_campaign=app_square_share_link

What's up with Panda King? Just drew a天地针 (doji candlestick)!
It's dropped over 99% from its peak, where did that once-celebrated #ZK go?

AI Agent = LLM + Tool Use + Agent Loop + Rules