In July, the spot and derivatives trading volume on CEX dropped to $3.76T, a month-on-month decrease of 23.9%; meanwhile, the spot share of DEX rose to a historic high of 19.5%. Previously, we assumed liquidity was on CEX, but now more spot trading, long-tail assets, meme tokens, and on-chain routing are staying on DEX. CEX certainly won't die, but its monopoly feel is indeed declining. However, the rise in DEX's share does not mean all DEX tokens will increase in value. The real beneficiaries are routing, MEV, wallets, aggregators, and market-making infrastructure, because these form the foundation of wallet entry points. The key shift from CEX to DEX lies in the change of entry.

MoneyGram has launched Ramps on Solana, providing cash access in 170+ countries/markets through a single API, and supporting deposits in 25+ markets. This kind of payment and on/off-ramp infrastructure may not cause $SOL to skyrocket in the short term, but it will definitely expand Solana's usage scope. Many public blockchains promote TPS, but the real issue for ordinary users is how to convert cash into on-chain balances and how to withdraw local currency from the chain—how salaries, remittances, and living expenses flow in and out. Cash access in 170+ markets is undoubtedly closer to everyday life than millions of TPS. As I always say, whoever controls the entry point is closer to the real users

Just now: Harmony is suspected of minting 4B $ONE, accounting for 26% of the supply. Among them, 2.8B has been transferred to exchanges. The team is cooperating with relevant exchanges to freeze the funds and is advancing patch and rollback plans. This is not an ordinary contract exploit; this is an incident at the supply level. When a chain can abnormally mint 26% of the supply, what users should worry about is not price volatility but the credibility of the ledger itself. Additionally, rollback can sometimes be life-saving but can also harm the narrative of immutability. At such times, whether a rollback happens or not, someone will be dissatisfied. If you hold $ONE or related ecosystem assets, don't rush to bet on a rebound. First, watch the progress of exchange freezes, official patches, whether a rollback will occur, and whether the minted tokens have entered the secondary market.

What you need to know before the US July CPI is released tonight at 8:30 PM: 1. Bloomberg and multiple institutions have a consensus forecast that July CPI will rise 0.1%-0.2% month-over-month, with the annual rate dropping from 3.5% in June to 3.4%; core CPI is expected to rise 0.2% month-over-month, with the core annual rate falling from 2.6% to 2.5%. This will be the first full inflation data this year to include the oil price transmission effect caused by the geopolitical conflict in the Strait of Hormuz, and market reactions are usually more intense than similar data in the following months. 2. Brent crude oil is still above $80, and the market will focus on whether "oil price-driven overall inflation" and "sticky core inflation" can be separated. Ajian believes that if core CPI really drops to 2.5%, it indicates that the stickiness of service sector inflation is easing, which is a substantial positive for a rate cut in September; but if the core data rebounds unexpectedly, combined with rising oil prices, the option of raising interest rates will be back on the table. These two scenarios have completely opposite impacts on crypto assets

Firmus, a company that started Bitcoin mining in Tasmania five years ago, has recently completed a $2 billion financing round, bringing the company's overall valuation to $10.5 billion. Its business focus has shifted to AI infrastructure. This is currently one of the larger publicly disclosed cases in the narrative of mining companies transitioning to AI computing power. The power and cooling infrastructure for Bitcoin mining is being recognized by the capital market as real assets that can be directly repurposed for AI computing centers. Therefore, the valuation of such transitioning companies should not follow the traditional mining company valuation logic, nor simply apply AI company valuation logic. Instead, it should consider the actual utilization rate of their power capacity and data center assets, and whether they have secured real computing power leasing contracts. This information is more critical than the financing amount itself. In the past, when discussing the future of mining companies, the default answer was usually that profits would be squeezed after halving, making mining companies increasingly difficult to operate. However, the Firmus case shows that the most valuable asset in the hands of mining companies may not be mining itself, but the power and cooling infrastructure they have already built. This asset has become a scarce resource amid the current shortage of AI computing power. So, if you are investing in stocks, rather than focusing on first-hand data like AI chips or computing power leasing (which is hard for ordinary people to obtain), it is better to look at whether traditional energy-intensive industries (such as mining companies) are willing to transition to AI infrastructure as a secondary signal. If even mining companies are moving in this direction, it indicates that the shortage of AI computing power is severe enough to attract cross-industry capital to reallocate assets. This signal is much easier to obtain than first-hand data but still carries significant information. If you are already following mining-related targets, the key indicators to watch next should shift from hash rate share to power capacity and AI computing power contract status. Using old indicators to evaluate such companies will lead to incorrect conclusions.

Polymarket adopted the TWAP mechanism on August 7 to replace single-point snapshot settlement. I wonder if any friends have been paying attention to this. One of the reasons behind this is that some accounts profited about $8.2 million through price manipulation, such as placing large orders on Binance in the last few seconds before the settlement window closed to push Bitcoin's price past the exercise threshold. Of these losses, 93% fell on retail traders. For the sake of retail traders' interests (not really), prediction markets are increasingly aligning with serious finance. This process will inevitably eliminate a batch of front-end users, arbitrageurs, and rough settlement designs. This will be a devastating blow to those ecosystem projects that rely on early dividends to survive.

Unexpectedly, the BIP-110 issue is still fermenting: BIP-110 attempts to limit large on-chain data through a soft fork, but miner support is only 2.53%, and some minority chains stalled after producing just two blocks; moreover, the removal of Luke Dashjr's editing rights from the Bitcoin BIP editorial team is reportedly also related to the BIP-110 controversy 😂 This is the strongest aspect of $BTC — even with such disputes, the main chain continues as usual. Keep in mind that proposals with low support can still create real risks. Although the main chain's price seems almost unaffected; in the protocol world, some events change not the price but trust, such as the removal of Luke Dashjr's editing rights. So if you write about Bitcoin, you can't just say it's immutable; you have to explain who can propose changes, who can edit the process, who judges neutrality, and how the community rejects minority factions pushing through changes. Above is my take

SushiDAO has released an RFC proposal planning to restructure the $SUSHI economic model by cutting xSUSHI buybacks, introducing a SUSHI Reserve, allocating funds for Sushi Ops, and planning to deploy $10M-$20M protocol liquidity to Robinhood Chain. This move is very much in the spirit of veteran DeFi players. It's not that the old protocol has nothing to do, but rather it's rethinking who the fees should ultimately benefit. Should buybacks go to holders, or be kept as reserves? Should liquidity stay on the original chain, or move to new distribution channels like Robinhood Chain? In my view, DeFi protocols are now entering a corporatization phase: previously, when protocols made money, the most popular news was buybacks and burns. But now protocols need to expand chains, build markets, support teams, and open new channels. If all the money is distributed, the team runs out of ammunition; if all the money stays in the treasury, holders feel diluted. This can be seen as a harsh test of shareholder mentality: do you want a dividend stock or a growth stock? Different allocation methods have completely different impacts on token value, so this liquidity migration doesn't necessarily mean a price increase. On the contrary, it may trigger dissatisfaction among original xSUSHI holders regarding profit distribution, which is a risk that cannot be ignored.

$WLD has recently seen an increase in both trading volume and open interest; $XMR's price also once touched $400, and a whale opened a $36M leveraged long position. These two coins are rising simultaneously—one focuses on identity, the other on privacy. Although these directions seem opposite, they both stem from the same anxiety: in the AI era, how do people prove themselves and protect themselves? Perhaps in the future, when paying attention to AI/identity/privacy themes, we can pay more attention to people's genuine emotions and usage scenarios