BlockBeats News, August 27, Binance founder CZ stated at the Bitcoin Asia 2026 conference:

“I think that (AI and crypto integration) may start with stablecoins. Once you can accept stablecoins, adding Bitcoin becomes very easy. Then, hopefully, adding BNB, Ethereum, Solana, and other blockchains will be just as easy. Also, some large AI companies may want to issue their own tokens, so they are likely to integrate these as well.

I have discussed with several top AI companies on how to issue tokens, and there are many different ideas. You are all very early now. For example, building a data center requires a large amount of capital, and 1 gigawatt of computing power costs around 300 to 500 billion dollars. Many AI companies, although they are highly valued, do not have enough funds; many companies do not have enough money to build computing power at that scale. Some of the AI companies I have talked to want to build hundreds of gigawatts of computing power in the next few years, which would cost trillions of dollars, right? So we will discuss how it might be possible to issue a data center token that would allow holders to be rewarded for using that data center in the future.

If you are going to issue a token, you want to maximize the token's utility as much as possible, so naturally, you would want to allow the token to be used for subscription services, and then you also want to link it to different language models and services. So there are many different patterns that can be combined. Hence, it is not unimaginable for AI companies to issue their own utility tokens and drive them. I think we have already seen some companies closely related to OpenAI issue tokens, such as Worldcoin. Although it may not have a great deal of utility today, those people are very close to AI companies.

So I think crypto will come into play. When AI is processing payments for us, they will use cryptocurrency. But today, I think AI payments might come after AI transactions. For example, using AI agents to book hotels, pay bills, and so on. Based on my discussions with some top AI companies, today, payments are not their priority. They are trying to make agents smarter to help you find the best deals, but they do not see payment as a big problem that AI needs to solve today. After AI helps people find the best hotel, most people can book by themselves using a credit card.

But in the field of transactions, AI may be more suitable because you have to quickly gather a large amount of information, must react quickly to news, and analyze charts. Whatever trading you do, AI can probably increase efficiency tenfold to facilitate this process. And to do this, you wouldn't want AI to just help you analyze and then you press the trade button yourself. For efficiency or timing reasons, you would want AI to directly place orders for you. So AI might first help you trade and then help you make payments.”