This morning, I checked my Bitcoin wallet—the inscriptions I minted back then peaked at nearly 50,000 USDT. Bad news: They’re basically worth zero now. Good news: The gas fees I spent can still be withdrawn to an exchange. Bitcoin is different from other chains: even if the inscriptions are worth nothing, the gas fees embedded in them can still be recovered. Overall, I haven’t lost much—consider it an indirect way of making a fixed deposit. If you’ve minted any, go check yours too—this market turn is an unexpected bonus!! https://t.co/XsVlaJ5boF

In this market, everything except social security is falling!

Seeing @AethirCloud’s updated tokenomics felt familiar—after all, they were quite generous with their airdrops back in the day. This update focuses on ATH finally attempting to link real compute business revenue with token value. After all, DePIN projects fear most the disconnect between claiming business growth while relying solely on token emissions to sustain the network. Revenue from genuine customers is the true indicator of whether a project can sustain long-term success. Of course, token burning doesn’t guarantee price appreciation. ATH’s original emission schedule remains unchanged; whether supply and demand dynamics improve ultimately depends on the volume of real orders, the scale of network revenue, and whether the burn rate can offset new supply. Currently, $AGPU is also a customer of Aethir, and AGPU holds a significant amount of ATH—this adds another layer of anticipation for me. If AGPU’s AI compute business continues to scale and generates more real demand, and if Aethir’s revenue and ATH burn mechanism gradually prove effective, then ATH’s value performance could theoretically create a positive feedback loop for AGPU’s ATH asset holdings. In an ideal scenario, both businesses mutually reinforce each other, forming a virtuous cycle—like stepping on one foot to lift the other. But this remains theoretical. Whether it materializes depends entirely on actual compute delivery, real revenue generation, and token supply-demand metrics—not just storytelling. As someone who was once stuck holding both tokens, I certainly hope they can achieve this outcome. Still, beyond short-term price fluctuations, what I truly want to see is how much real revenue flows into the burn mechanism after the November update, how much is actually burned, and how many sustained orders new compute capacity generates. Announcements can set expectations—but on-chain data and real revenue are the ultimate answers. Let’s watch and verify together. I hope this time it’s not just empty promises, but a genuine connection between business growth and token value.

At first glance, it looked like another project had shut down. But upon closer inspection, I realized this is about abandoning L2 in favor of L1—after all, L1s are now faster, cheaper, and carry that prestigious lineage; who wouldn’t love them? Previously, L2s were the hot startup trend; now, they’re being shunned. L2s were originally created because the elite chains were too expensive. But now, L1s are cheaper than some L2s—so what’s the point of L2s anymore? https://t.co/TleUS9nrtR

In 2026, AI infrastructure is undoubtedly one of the most closely watched sectors in the capital markets. Everyone is chasing GPU compute power, data centers, and AI cloud services. And the rising AI compute star, @AxeCompute (NASDAQ: AGPU), is gradually piecing together the most critical components needed to build a new neocloud player. AGPU only transitioned from Predictive Oncology into an AI infrastructure company by the end of 2025, yet it has already secured over $3 billion in multi-year contracts in a short time. This growth is impressive, but the market’s real concern is whether a newcomer has the capability to build, deploy, and sustainably operate such a massive GPU cluster. After all, signing contracts and delivering compute power are two very different things. GPU procurement, data center construction, power supply, network configuration, and cluster orchestration—all require capital and deep expertise. For a new company, building every capability from scratch would pose serious challenges in terms of expansion speed and financial pressure. AGPU’s approach differs from that of some other neocloud players. Companies like CoreWeave and Nebius are investing heavily in building their own AI infrastructure and engineering capabilities. AGPU, by contrast, emphasizes leveraging specialized platforms for cluster deployment and operations, allowing it to focus more on customers, contracts, and capital allocation. This can be understood as a “light organization, heavy asset” strategy: rather than doing everything in-house, it entrusts critical functions to partners with proven experience. The most noteworthy partner here is Penguin Solutions. With over 25 years of relevant experience, Penguin has deployed and managed more than 85,000 GPUs, accumulating over 2 billion GPU runtime hours. It played a key role in building and operating Meta’s AI Research SuperCluster and was selected as NVIDIA’s托管 service partner for Voltage Park. For neoclouds seeking rapid scale-up, this kind of experience isn’t just about building one cluster—it’s about reducing the complexity of replication and ongoing operations. Penguin’s own business performance further validates this model. According to its FY2026 financial report, four of the six new AI infrastructure customers added in the fourth quarter were neocloud clients. This suggests that an increasing number of neocloud companies are turning to specialized service providers to fill gaps in their infrastructure delivery and operational capabilities. Looking at AGPU, its $3+ billion in signed contracts, over $317 million in customer prepayments, and a five-year 55MW colocation agreement provide strong backing across orders, funding, and infrastructure resources. Penguin’s involvement further completes the critical piece: professional deployment and operations. Of course, this doesn’t mean AGPU is without risk. Prepayments are not profits; contract value is not revenue; and securing data center space doesn’t guarantee timely GPU cluster deployment. Ultimately, success hinges on actual delivery, revenue recognition, and cash flow realization. But I believe understanding AGPU requires looking beyond how many billions in contracts it has signed—it’s about whether it can leverage professional partners to turn those contracts into repeatable delivery capabilities. The real competition in the neocloud space isn’t just about building one 2,000-GPU cluster—it’s about whether you can replicate that 2,000-GPU delivery model across 50,000 GPUs or more. A single successful delivery proves project execution capability; sustained replication is what builds scalable competitive advantage. Professional platforms can help newcomers shorten their capability-building timeline—but whether they can replicate success still depends on the alignment of capital, supply chains, power resources, customer demand, and project management. Therefore, what makes AGPU’s partnership with Penguin noteworthy isn’t merely the addition of another vendor—it’s that AGPU may be adopting a fundamentally different growth model: AGPU focuses on acquiring customers, securing contracts, and organizing capital; Penguin delivers professional infrastructure deployment and operations. If this model proves viable, AGPU’s growth won’t need to rely entirely on building every engineering team from scratch—it can accelerate deployment by leveraging mature platforms. The next key milestones will be whether the first clusters launch on schedule, whether revenue recognition begins, and whether delivery scale can be consistently expanded. Contracts have opened the growth potential; professional platforms have provided the execution path; but scalable delivery is the real test AGPU must now pass.

Previously, I paid the minimum required social insurance through my friend’s company, but since last year, I’ve had to make up payments for the full year. This year, I again need to pay the increased portion for January through September. With these annual increases, it’s hard to imagine reaching the 20-year contribution requirement to retire at 65. Isn’t this just like staking in a Web3 project? You put your money with the project team, lock it up, and then unlock it later to claim returns. The only difference is that with social insurance, you get discounted medical care during the lock-up period—unlike Web3, where there’s no such benefit. In essence, social insurance is just a bet: you’re wagering on whether you’ll live long enough to collect.

So fast! It’s already been six years since Filecoin launched—and it peaked right at launch! Good news: The annual release of 68.3 million FIL will stop. Bad news: You’d need a 400x+ price increase just to break even! With the vesting period ending, this doesn’t mean no new FIL supply will be created. Storage providers will still receive block rewards—but the token release structure has changed: 25% is unlocked immediately, and the remaining 75% unlocks gradually over 180 days. In my view, this marks a significant turning point as Filecoin’s economic model enters a new phase. For years, the market had to absorb massive vesting releases; now, that supply pressure is about to disappear—a development worth watching over the long term. But don’t equate reduced new supply directly with price increases. Whether FIL can truly gain momentum depends on real storage demand, ecosystem adoption, investor interest, and whether the market can absorb the existing circulating supply. Improved supply conditions are just the first step. What ultimately determines how far FIL can go is whether demand can keep pace.

Received a package from @OneKeyHQ before National Day—inside was a silver OneKey Pro 2. My first impression: it looks like a mini iPhone. The official images show three colors: Glacier Silver, Sunrise Orange, and Ink Rock Black. While orange stands out, I personally prefer silver—it’s timeless and elegant. The device feels smooth in hand, with rounded edges that don’t dig into your palm. The aluminum glossy frame paired with a matte glass back elevates its build quality far above most hardware wallets. Priced at $299, it comes bundled with a $257 value package: a $179 Prime membership, a $39 SeedCard, and a $49 magnetic leather case. I’ve long criticized the Pro 1 for lacking a tempered glass screen protector—this generation solves that by including the magnetic case outright, finally addressing protection. Compared to the Pro 1, the main improvements are in ergonomics and battery life. The entire body is smaller, about 1.5mm thinner, while maintaining the same screen size—but now with a full-screen design, higher screen-to-body ratio, finer display detail, and increased brightness. The fingerprint sensor remains on the right-side power button and feels more responsive than before; I recommend enrolling multiple fingerprints. With the case on, I mostly use my right index finger. The battery has been upgraded to a 700mAh lithium-polymer cell—the same class used in smartphones—significantly extending lifespan, so you no longer need to worry about long-term storage. Wireless charging is now standard, eliminating the need to always plug into a computer. Honestly, I didn’t even realize the Pro 1 supported wireless charging. The system-level experience feels noticeably better. The UI has been overhauled: for the first time, you can view your assets directly on the device—a feature most other hardware wallets still lack. In terms of fluidity, the Pro 2 feels like a brand-new flagship phone; the Pro 1 feels like an iPhone used for three to five years; other brands still feel like aging Android devices—sluggish when swiping, confirming, or paging. Security features aren’t new to this generation—most are also available on the 1S and Pro 1: SiteScan for dApp scanning, light preview of contract transactions, automatic hiding of suspicious tokens, MEV protection, hidden wallets, connection verification, private sending, and KYT. It covers nearly all common pitfalls during everyday transfers and interactions. For backup, the previous OneKey Lite could be directly attached to the back for import, and the included SeedCard works here too—the process is nearly identical to the Pro 1. Recent security incidents in the space have been frequent. As asset volumes grow, adding an extra layer of security is essential. Today’s interactions are no longer as cumbersome as in the early days—most of the time, you’re just confirming one extra step in your app. Ultimately, my real-world experience: lighter, thinner, smoother—and the included accessories bring down both protection and usage costs.

Recently, both Blast and Abstract announced the termination of their operations. While consolidating assets, we found that funds on the Debank chain currently have no bridge available to return them. This project launched XP, sold domains, and offered VIP memberships two years ago—all aimed at a promising future. However, its Twitter account hasn’t been updated in over a month, and there’s been no further news since. Is there still hope for the DBK chain? https://t.co/ulMsNqtw91

The combined 24-hour revenue of the former Four Kings is less than half of the new L2 star RH. It’s upside down and highly unusual. Meanwhile, our Fule, despite having already TGE’d, still generates nearly $6,000+ in 24-hour revenue. Abstract, which hasn’t TGE’d yet, is still bringing in over $3,000—likely all from rug-pull contributors. Currently, nearly 80% of all chains generate virtually no revenue, while fewer than 5% of projects remain profitable. Who’s next? 24-Hour Revenue of the L2 Four Kings: Arbitrum - $26,116 Starknet - $9,135 OP - $3,462 ZKsync - $465