CZ: AI-Powered Trading and Payments Will Arrive in Months, Leveraging Blockchain

On June 20, Binance founder Changpeng Zhao (commonly known as CZ) sat down for an interview on the Galaxy Brains podcast with Alex Thorn, director of Galaxy Research. CZ revealed he’s installed and subscribed to nearly every major mainstream AI large language model (LLM), adapting his use of each product to match specific needs. OpenAI was the first LLM he set up, Claude works best for coding tasks, and when operating in regions like Hong Kong, he turns to Chinese models such as DeepSeek and Kimi. He stressed he doesn’t choose tools based on political alignment—his rule is simply “use whatever works.” Overall, he observed that American LLMs tend to be more accurate, while Chinese open-source models are strong though sometimes less comprehensive. CZ forecasts AI-driven trading and payments will roll out in months, not years. He explained AI can help users book tickets or execute trades, but it still requires integrated payment and transfer capabilities. Traditional payment systems aren’t well-suited for direct AI use: for example, an AI can’t manually swipe a Visa card, and it struggles with processes like two-factor authentication (2FA) and KYC. In contrast, blockchain is API-driven, making it far easier to integrate with AI systems. Since AI needs to spend funds and complete transactions, CZ believes it will rely on blockchain technology. When talking about AI competition, CZ stated there will absolutely be rivalry between the U.S. and China in the AI space, spanning areas like chips, data centers, and manufacturing capacity. He added that competition can fuel growth, but the pace of AI development needs careful handling—AI is extremely powerful and carries major risks. CZ noted AI regulation is more complex than crypto regulation. Crypto rules mainly address issues like fund security, KYC, and anti-money laundering (AML), but AI poses diverse threats: it can support hacking attacks, assist in creating weapons, or control robots and drones. Right now, there aren’t enough experts devoting deep thought to how to fully regulate AI.