Recently, this stage is quite interesting, BTC is staying still, and various alts are starting to take turns performing.

The tactic is actually quite old: first, use a big drop in BTC to shatter the mentality of retail investors, wait for them to cut their losses and leave the market, then pick a few alts to pump, making those outside the market envious. I've been in this market for so long, I've seen this kind of wash trading technique too many times. When it drops, they panic and sell; when it rises, they chase the highs to buy in. Human nature is repeatedly harvested like this.

To survive in this market for a long time, you must learn to control your emotions. Hold onto the tickets you should keep, and patiently wait for those you should observe. Don't let the ups and downs of this volatility throw you off balance.

Speaking of projects, I recently noticed GAIB, which has an interesting approach - packaging AI computing power infrastructure as an on-chain yield product. In simple terms, it turns the business of GPU leasing into a model similar to RWA. Cloud service providers lack the funds to buy graphics cards, and GAIB helps with financing, then shares the profits generated from the computing power with investors.

This logic reminds me of the hash power mining tokens I traded last year, except back then we were mining Bitcoin, and now we are renting it to AI companies to train models. Essentially, both are about assetizing hardware computing power.

Staking AID tokens for returns is indeed a familiar play. However, the core issue lies in whether the actual yield from AI computing power leasing can outpace token inflation and operational expenses. A single NVIDIA H100 card costs tens of thousands of dollars, and the payback period may be longer than expected. Moreover, the demand for AI training is not consistently explosive; market fluctuations depend on the real needs of downstream customers.

Whether the path of tokenizing computing power can be successful depends on the actual operational data.