Binance founder and Giggle Academy founder Changpeng Zhao said he would not advise anyone to sell an apartment and go all-in on Bitcoin if that apartment makes up most of their wealth, even though he famously did exactly that in 2014. Speaking during a community Q&A at Binance Clubhouse Bali 2026, Zhao said the more practical approach for most people is dollar-cost averaging, or DCA, by putting 1%, 5%, or 10% of monthly income into major crypto assets they trust.
The podcast aired on Aug. 23, 2026, with Jessica Walker as host. According to the compiled transcript, Zhao discussed risk management, shifts in his own market views, the four-year cycle, AI and crypto infrastructure, Giggle Academy’s education experiments, and what Indonesia still lacks as it tries to build a larger Web3 ecosystem.
The input also includes a disclosure stating that Zhao is the founder of Binance, one of the world’s largest crypto exchanges, and that both his personal wealth and Binance’s business are highly tied to crypto asset prices. The program touched on BTC, BNB, as well as sectors including real-world assets, or RWA, and decentralized finance, or DeFi. The original note said the piece presents the guest’s views faithfully and does not constitute independent third-party judgment or investment advice.
Zhao says his 2014 all-in move is not a template for most people
Walker asked Zhao what the 2026 equivalent of his 2014 apartment sale would be, and whether he would recommend that kind of move to a 25-year-old today. Zhao’s answer was blunt: “I wouldn’t recommend anyone sell their apartment and go all in if that is most of your wealth.”
He said that decision has to be personal because crypto prices are highly volatile and people need to understand the downside and the risk they are taking. Zhao drew a line between his own position at the time and what most people face today. He said he had very strong conviction that Bitcoin would rise over the long term, and his daily life did not depend on Bitcoin’s price. Even if Bitcoin had gone to zero, he said, he would still have been fine. Most people are not in that position.
Zhao did not retreat from the long-term case for crypto. He said he still believes crypto will continue to grow over time, and that Bitcoin, BNB, and other strong projects will keep growing. But he separated that broad conviction from the question of portfolio construction for ordinary investors.
For most people, he said, DCA is the better answer. “I would recommend most people DCA, put 1%, 5%, or 10% of your monthly income into major crypto assets,” he said. In the compiled piece, the examples for those major assets were BTC, ETH, and SOL, with the broader point being to avoid leverage and stay away from random smaller tokens.
The article’s summary framed that range as one of the most actionable figures from the session. Using the same percentages, a person earning $10,000 a month would be allocating roughly $100 to $1,000 into regular purchases of large-cap crypto assets. That example was presented as an application of Zhao’s rule rather than a separate rule of its own.
He says he changed his mind on NFTs, meme coins, and RWA
Asked which crypto belief had changed the most over the past decade, Zhao said there was more than one answer.
First, NFTs. He said he did not expect NFTs to become popular, but they did. Second, meme coins. He said he also did not expect meme tokens to become so big, and that he still does not fully understand the logic behind them, even now.
Third, RWA. Zhao said he did not think much of the sector a year and a half ago, but his view has changed. He said he now sees real attraction in the model, pointing to 24/7 trading, transparency, low fees, and global accessibility as concrete advantages.
He tied those changes together with a broader point about how he updates his views. “There are many things I didn’t see at first, but once they became popular, I went back and tried to understand them,” he said, adding that he does not want to stay stuck on an earlier judgment just because it came first.
The four-year cycle still looks strong, but he says the next trigger is hard to call
Zhao did not introduce a new market framework when asked about the four-year cycle. Instead, he said the pattern still appears to hold.
“The four-year cycle still looks very strong,” he said. He added that four years ago the market was also roughly in a bear phase, which he described as similar to current conditions, and said the pattern has not been broken so far.
That confidence did not carry over to predictions about the next major catalyst. Zhao said the next trigger is very hard to predict. He pointed to the 2020 DeFi Summer as an example, saying he could not have called it even six months in advance.
The compiled article listed RWA, perpetual decentralized exchanges, and AI agents as areas now attracting momentum, but Zhao did not say which one, if any, will push the broader crypto market to new highs. “I don’t have a crystal ball,” he said. In his view, the sensible approach is to watch developing trends rather than pretend certainty where none exists.
On AI, Zhao says crypto-native payment rails fit machine agents better than cards
Zhao described AI as eventually taking on tasks such as making payments, placing trades, and booking hotels. “That day is not here yet, but it will come,” he said.
Walker framed one question around the fear that AI could front-run retail traders, discover vulnerabilities faster than humans, and turn into a nightmare for smaller market participants. She also asked whether Binance should build guardrails.
Zhao challenged the premise that AI will remain concentrated in the hands of a few users. He said he expects nearly everyone to use AI.
He acknowledged that AI can identify vulnerabilities faster, and that some people will use AI for trading earlier and more effectively than others. But he said that same capability can also help developers patch vulnerabilities faster and build safer systems. Early adopters may gain an edge, but others will learn from them, much as they did with the internet and with technical analysis tools.
“I’m not planning for Binance to protect humans from being harvested by AI bots,” Zhao said. He argued that AI tools should be opened to everyone. At the trading API level, he noted, what an exchange sees is an incoming order, not whether the decision behind it came from a human or an AI system. His position was that broadening access makes more sense than trying to wall tools off.
When the discussion moved to AI agents holding wallets, making payments, and interacting with smart contracts, Zhao said blockchain infrastructure will be very important for AI. He argued that crypto-native payment systems are the smoothest option for AI because they are programmable, neutral, borderless, fast, and low cost.
He contrasted that with credit cards, which he described as a payment system designed for humans. If AI has to use cards, he said, a person keeps getting pulled back in through face verification, SMS checks, and random transaction declines. In his view, that friction makes card rails a poor fit for autonomous agents.
How he uses AI today, and where he stopped using it
Zhao said his most common use case for AI is information retrieval. Before, he would search the web, sort through ads, and read multiple articles to understand something. Now, he said, he can ask AI directly and skip much of that process.
He did not present AI as a universal tool that works equally well for everything. Zhao said he tried using AI to write code that would filter Gmail spam, spent a fair amount of time on it, and eventually gave up because the result was not good enough.
He added that he does not write much code anymore, so AI coding has limited personal value for him. For developers, though, he said AI may improve productivity by 5x to 10x. Even then, he stressed that engineers still need to understand every line of code AI generates rather than treat it as a black box.
On news consumption, Zhao said he does not use AI either. He said he does not follow traditional news because he sees too much subjective opinion in it, and instead goes straight to the people he follows on X.
Giggle Academy says pure AI teaching did not work for young children
Zhao also spoke about Giggle Academy’s direct experience using AI in education. He said the platform now has more than 1 million children learning on it and focuses mainly on content for ages 2 to 6.
According to Zhao, the team tried fully AI-based teaching and found that it did not work. One problem was that chat-style AI does not proactively ask questions, and young children often stop after asking three or four questions. Another problem was engagement. He said purely AI-generated content was less appealing to children, and that even very young kids could sense that it was not made by a person.
That led the team to keep humans in the driver’s seat while using AI to generate content components. Zhao used storybooks as an example: fully automated one-click AI storybooks were not popular, while stories written by humans with some AI-generated components performed better. He said a team of 60 people is working on that process.
He also pushed back on the importance often assigned to decentralized identity. Zhao said he does not think decentralized identity matters as much as many people assume because blockchains themselves are too transparent. Combined with AI and KYC information from a few centralized exchanges, he said, tracing on-chain activity becomes very easy. In that comparison, he said banks and cash are actually more private.
On Indonesia, he says the country has users but lacks local assets
Asked what Indonesia still needs if it wants to become a leading Web3 ecosystem in Asia, Zhao pointed less to user growth and more to asset formation.
He said the Indonesian community is already very active, but the country is still missing several key pieces. It does not have its own stablecoin, and it does not have its own RWA offering at scale. He listed tokenized Indonesian stocks, real estate, gold, oil, rare earth minerals, and government bonds as examples of assets that are still missing.
Zhao also used USDT as a comparison point. He said USDT is backed mainly by U.S. Treasuries, which he described as an indirect version of tokenized government bonds. From that perspective, he asked, what country would not want hundreds of millions of people around the world to be able to buy its government debt directly?
He added that having a stablecoin tied to the local currency is also important because it provides a price discovery mechanism. More broadly, he said these gaps point to the same underlying issue: Indonesia needs clearer regulation.
His advice to a 25-year-old: ignore most specific tips and spend 30 to 60 minutes learning every day
When Walker asked what advice he would ignore if he were 25 years old today, with no money and no network, Zhao again avoided giving a token list or a project list.
He said people naturally look at others through their own worldview, but circumstances differ from person to person. For that reason, he would ignore most specific advice, especially instructions such as buy this, sell that, do this project, or avoid that one. In his view, that level of specificity often does not fit the person hearing it.
His replacement was deliberate learning. Zhao said young people should spend 30 to 60 minutes a day learning, and that the younger you are, the more time you should put into it. He said that can include YouTube, X, and asking AI questions, but the process needs to be intentional rather than passive scrolling.
He also argued that each question should be pushed three to five layers deeper. In the source summary, he gave market questions as an example, saying broad prompts such as asking why the market went up or down are not enough by themselves.
Zhao strongly recommended reading, especially nonfiction and self-improvement books. He said a book on Amazon costs $9.99 and called that one of the cheapest ways to acquire knowledge.
His final point was about compounding on the cognitive side. If someone studies for an hour every day, he said, their way of thinking changes over time, their analysis becomes deeper, and their decisions improve. It does not happen overnight, but it does happen.