Key takeaways

  • Binance Online 2026 brought together the heavyweights of crypto and finance, including CZ, Chamath Palihapitiya, and Anthony Pompliano for a roundtable discussion on where the smart money is flowing.

  • Chamath laid out what he calls a 'dirt to token' four-layer framework, covering the physical inputs behind AI, the software control plane, and distributed computing.

  • The convo also dug deeper into how CZ and Palihapitiya are zeroing in on AI infrastructure, building agentic use cases, and the trend of merging AI with finance.

At Binance Online 2026, discussions about capital flows are no longer stuck on the old question of 'crypto vs. stocks.' The new question is how AI infrastructure, cryptocurrencies, and traditional finance are intersecting.

This is the backdrop for the roundtable discussion 'Where Smart Money Is Moving Now.' The discussion was hosted by Anthony Pompliano, founder and CEO of Professional Capital Management, with guests including Binance and Giggle Academy founder Changpeng Zhao (CZ), and venture capitalist, entrepreneur Chamath Palihapitiya.

Chamath's investment thesis for 2026: starting from 'Dirt to Token.'

Palihapitiya proposed a four-part framework to explain how he allocates capital within the AI tech stack.

"One theme is the concept of 'dirt to token.' It means: the land that houses the data center, which houses the racks, which ultimately houses the chips, resulting in tokens."

The first layer is the data center tech stack itself: the land, power, and racks that house the chips. The second layer is what he calls AI 'physical pivot assets,' including prismatic LFP batteries and rare earths. The third layer is the AI control plane, which is the software layer where agents interact with users and models. The fourth layer is distributed computing, i.e., inference and training closer to the end user.

On the control plane aspect, Palihapitiya is trying to build this layer through his proprietary AI-native software development platform 8090.

"This is a concept: there will be a very powerful 'connector' between humans and agents to cover the entire interaction spectrum of software development. It must be model-agnostic. Workflows can be broken down into tasks: some tasks might be better suited for Anthropic, while others for OpenAI. You then reassemble these components. There will be an agent responsible for cost and optimization management. Humans will also collaborate with these components, many of whom are non-technical."

The logic is that no single AI model can excel at all tasks. The control plane sits above the model layer, routing tasks to the provider best suited to handle that job. The control plane can also optimize costs and allow non-technical users to orchestrate complex workflows without needing to write code.

CZ focuses on 'selling shovels.'

CZ mentioned that his personal allocations outside of cryptocurrencies are primarily concentrated on infrastructure powering AI rather than the models themselves.

"I'm actually very interested in AI infrastructure. These AI models are indeed fascinating; I believe they have tremendous potential. But I'd rather invest in the 'shovels and picks,' which are various infrastructures that power AI."

CZ also views AI as a catalyst for the next leap in robotics and biotech. Even so, most of YZi Labs' funding (about 70% to 80%) is still allocated to Web3, with AI and related fields making up the rest.

Two investment approaches.

The roundtable discussion also revealed significant differences in investment approaches between CZ and Palihapitiya. Palihapitiya described a research-oriented process, centered around what he calls a 'prepared mind,' and builds through his Learn With Me series: he will publicly outline a systematic review of an industry or technology before investing.

Palihapitiya stated, "Everything is rooted in learning first, then acting."

CZ, on the other hand, takes a different approach: he starts with a small position and then uses that as an opportunity to dive deeper into the asset. Having 'skin in the game' provides the motivation to learn.

The intersection of AI and finance.

After shifting away from investment topics, the discussion touched on the overlapping areas of AI and financial services. CZ believes that crypto infrastructure is actively facilitating compatibility with autonomous agents.

"We want to make all crypto infrastructure agentic-ready. This way, when agents want to use cryptocurrencies, they can call a skill or API, and the agents can use it directly."

CZ described an AI agent capable of holding wallets, transferring stablecoins, and executing trades without human intervention at every step. While cautioning about risks, CZ also believes that allowing AI to manage capital at scale is still in its early stages.

"Letting agents handle large amounts of money, or even any funds, today is a bit risky because AI is still new, and they lack sufficient guardrails."

Although he hasn't tested the feature himself, CZ mentioned Binance's own AI trading bot as a viable alternative: the agent operates in an isolated wallet and is set up by users. Users can deploy strategies through natural language prompts, significantly lowering the barrier to algorithmic trading. CZ also believes that AI will eventually surpass ordinary human financial advisors since most advisors provide generic advice rather than truly personalized analysis.

In this space, Palihapitiya is incubating two companies: a next-gen quant trading firm run by a former hedge fund portfolio manager, and an AI-powered wealth management firm he's co-supporting with friends.

Palihapitiya stated, "All these things should just be a prompt, plus a human's 'yes/no' confirmation, and then achieve flawless execution."

Financial democratization, and the fusion of crypto with traditional finance.

In CZ's view, blockchain tech and stablecoins play a crucial role in regions where traditional finance (TradFi) systems can't reach.

"We've seen that in countries or regions with low bank penetration, like Africa and Southeast Asia, people are using blockchain to access stablecoins that were previously unavailable to them. You might think it's just a simple stablecoin, but for many around the world, it's actually financial accessibility."

In economies facing hyperinflation, holding USD through traditional banking channels is often fraught with challenges or even impossible. CZ believes that in many parts of the world, stablecoins are the most practical on-ramp for USD deposits.

Pompliano emphasized at the end of the discussion that 'crypto' and 'finance' are becoming the same thing, comparing the development path of crypto to that of the internet.

"I think we are heading in the same direction as the internet. In the past, there were internet companies, internet assets, internet entrepreneurs; now we just call them companies, assets, or entrepreneurs. The same is happening in the crypto space. I always use BlackRock as an example: are they now considered a crypto and Bitcoin company? Because their biggest, most profitable product is a Bitcoin product? No, it's just a company; it belongs to finance."

Summary

The Binance Online 2026 roundtable discussion 'Where Smart Money Is Moving Now' gives us a glimpse into CZ and Palihapitiya's big bets for 2026: from AI infrastructure (like data centers and chips) to the software company Palihapitiya is building, aimed at acting as an AI control plane.

If you missed this roundtable discussion or want to watch other segments of Binance Online, head over to Binance Square for the full replay.

Further Reading

Please note: The original English content may differ from the translation (the translated version may be AI-generated). For any discrepancies, please refer to the original English version for the latest or most accurate information.