Long before blockchain became a household term, Yat Siu was already shaping the architecture of the internet.
From founding Outblaze to co-founding Animoca Brands, Yat shares his vision on digital property rights, the agentic web, and why Web3 is the non-optional foundation of the future internet.
Long before blockchain became a household term, Yat Siu was already shaping the architecture of the internet. Beginning his entrepreneurial journey during the pioneering Web1 wave of the mid-1990s, Yat established himself as a forward-thinking technologist by founding Outblaze, an enterprise connectivity and messaging giant that later saw major asset acquisitions by IBM.
Instead of remaining inside the comfortable confines of legacy enterprise tech, Yat recognized that the next great paradigm shift would occur at the intersection of culture, community, and coordination. He co-founded Animoca Brands, transforming it from a mobile gaming studio into one of the world’s most influential Web3 investment powerhouses and ecosystem builders. Today, with major milestones ranging from early backing of era-defining projects like CryptoKitties to launching the decentralized identity ecosystem Moca Network, Yat continues to champion digital property rights as the absolute foundational infrastructure for the next generation of the internet.
Background & Bio
- You were building internet businesses in the mid-1990s, long before social media, smartphones, or blockchain existed. Looking back across Web1, Web2, and Web3, which assumptions about the internet proved completely wrong, and which have remained surprisingly constant?
As with many other complex systems, the internet evolved because its many contributors addressed what they perceived to be missing in each iteration. This is easily seen in the evolution of the internet since the advent of the world wide web in the 1990s:
The early read-only web (Web1) made discovery possible at a global scale, and suddenly people, ideas, and projects gained cross-border visibility in a manner that had been impossible until then.
Then, around 2004, the read-write web (Web2) emerged via the addition of social participation and instant interaction, which dramatically increased collaboration, adoption, and engagement, and accelerated innovation.
Over the last decade, the decentralized “read-write-own” web (Web3) introduced digital property rights to the online realm, which I think is the single most important structural change yet because it directly challenges the digital manorialism that became the norm in Web2. Ownership finally enables netizens to own their assets, data, and identity in digital environments, instead of merely “renting” them from powerful dominant platforms.
And now, with the popularization of AI — specifically mass access to LLMs and agentic AI — the web is inexorably evolving to incorporate these powerful, world-changing technologies at the foundational layer. The intelligent or symbiotic web (Web4), centres around the growing interactions between organic and artificial intelligences and the integration of AI in countless aspects of business and personal life.
The change has been incredible, but two things remain constant: each phase of change dramatically expands human capability, and each phase is initially dismissed as hype before its real utility becomes obvious. The value of Web4 to me is about agentic AI operating on top of ownership rails, which makes Web3 systems significantly easier to use and more accessible. This means that more people can actually participate in the digital ownership paradigm instead of being excluded by the relative complexity of blockchain.
- After building Outblaze into a successful internet infrastructure company and completing a deal with IBM, you could have stayed in enterprise technology. What convinced you that gaming — not enterprise software — would become the most important gateway to digital ownership?
I never saw it as a choice between enterprise software and gaming in the abstract. In fact, even while we were operating the Outblaze enterprise messaging business we were also setting up gaming companies and projects. We saw that gaming was an excellent indicator of new digital behavior, allowing us to see emerging patterns early and at scale. I’ve always been a gamer myself, and what we saw very early on was that games were not just entertainment, not just drivers of technological development, but represented the first large-scale digital societies. That was true of early MMORPGs starting at least as far back as the 1990s, then with the mobile games revolution, and later when blockchain products became readily accessible.
When the iPhone first appeared, a lot of serious gamers and game developers dismissed mobile gaming entirely because the experience was so simple and limited compared with much more sophisticated games for PC or console. But in the world of technology, which tirelessly improves on and refines itself, that kind of thinking is a massive shortsightedness trap.
We believed mobile gaming would become huge for the simple reason that eventually everyone would own a smartphone, which would make mobile gaming accessible to new audiences, including many people who had never considered themselves gamers before (the rise of casual gaming). We were correct about that trajectory.
The digital ownership thesis came quite a bit later, when our work in blockchain, including exposure to CryptoKitties, made it clear that games were also the most natural entry point for digital property rights. People understand ownership intuitively in games, because they already invest time, money, identity, and status in their highly personal game experiences. So for us, gaming became the de facto gateway to decentralization because it was the broadest and most “human” path into the relatively complex world of digital ownership.
And now, with agentic AI, I think gaming will again be one of the early adoption use cases, especially through agentic user-generated content and agentic NPCs. Games are where so many people learn to interact with new forms of technology (e.g. from 3D graphics to AI).
- You’ve said that NFTs, rather than Bitcoin, were your entry point into crypto. What did you see in the early CryptoKitties phenomenon that many experienced technology executives and investors completely missed at the time?
What many people missed around a decade ago was that digital property rights could finally become real on the internet in a way ordinary people could understand. This was the fundamental “aha” moment resulting from the work one of our subsidiaries did on the launch of CryptoKitties.
In Web2, people often feel that they own their digital items, especially in the many cases where those items represent significant effort and/or cost. But in reality, there’s zero actual ownership. You are typically merely “renting” access to what you think is your property under the platform’s or publisher’s unilateral licensing rules. You cannot freely transfer your assets, build on top of them, or take them elsewhere. You simply generate value for the platform/publisher’s property: digital manorialism.
CryptoKitties changed that. What became obvious to us very early was that once an asset is truly owned by its user, third parties can build around it without asking the original issuer for permission. That was the real breakthrough, because it creates the conditions necessary to generate new businesses, services, and network effects based primarily on ownership. In the physical world, ownership is what makes an economy around an asset possible (e.g., you can genuinely own and sell your home and your car, and this leads to property agents, car resale businesses, aftermarket industries, etc.). CryptoKitties conclusively demonstrated that this was possible in the digital world.
Vision & Thesis
- For years, you’ve argued that digital property rights are the missing layer of the internet. As AI agents begin creating, transacting, and making decisions autonomously, how does your thesis evolve when ownership is no longer exercised only by humans?
If anything, AI agents make digital property rights more important, not less. Humans have limited time, limited attention, and limited capacity. Agents can operate continuously, at machine scale, across vast amounts of content, data, assets, and transactions. So the question of who owns what, who can use what, and who should be compensated becomes much more urgent. If agents are going to create, decide, trade, and interact on our behalf, then ownership cannot remain vague or dependent on the discretion of platforms. It has to be clear, enforceable, and machine-readable.
That is why tokenization matters. When assets, rights, or economic interests are represented on-chain, they are easily legible to AI agents and can be used, exchanged, and governed in a way that preserves attribution and compensation. In the agentic world, digital property rights are not just a philosophical principle, they are an essential part of the operating system for the agentic economy.
- Many people describe Web4 as the convergence of AI, identity, and decentralized networks. In that future, do you expect AI agents to become economic actors in their own right — with wallets, reputations, and assets — or will they remain tools acting on behalf of human owners?
Many agents simply act on behalf of humans or institutions, but some can and do become genuine economic actors with wallets, reputations, transaction histories, and delegated authority to make decisions within defined boundaries. Contrary to some public perceptions, that makes humans much more relevant, because it is humans who set the intent, values, constraints, and accountability of agents, while the agents handle discovery, coordination, execution, and settlement on behalf of their human stewards.
To me, the real shift in the agentic age is not just automation, but the reallocation of human attention. If agents take over repetitive and operational work, if agents can more efficiently locate information and cut through the hassles of intrusive advertising, then people are able to spend more time on judgment, creativity, strategy, and original thought.
In addition, digital identity becomes foundational, especially for agents, because it is necessary to know which agent is acting for whom, with what permissions, and with what reputation. That is what our project Moca Network is building: a digital identity and reputation layer for the open internet, where users and AI agents can carry verifiable, portable credentials across apps, platforms, and chains without relying on a central gatekeeper.
Market Structure & Personal Perspective
- As AI ecosystems become dominated by a handful of frontier models and infrastructure providers, are we at risk of recreating Web2’s power structures on an even larger scale?
If AI develops mainly through closed models, closed distribution, and closed identity systems, then yes, we risk recreating Web2’s concentration of power in an even more extreme form. But I don’t think this is where we are heading. In fact, I think the agentic web is developing differently because it is being built on more open infrastructure, especially Web3 rails such as wallets, tokens, and decentralized identity. Agents cannot participate economically in a truly native way through traditional systems, but they are perfectly suited to the machine-to-machine interactions of blockchain-based systems. Web3 gives agents a way to transact, coordinate, and use assets in an open environment rather than being confined to a single closed platform. If we want an AI future that does not simply repeat Web2, then digital property rights and decentralized infrastructure must be part of the foundations.
- You’ve spent much of your career investing in ideas before they became mainstream. Looking toward 2035, what is one belief about AI, Web3, or the future internet that you hold today but that most of the technology industry still disagrees with?
2035 is a long time from now in the context of technology innovation cycles, which are becoming shorter and shorter. But one belief I hold very strongly is that digital property rights will prove to be one of the most foundational layers of the future internet. AI will be everywhere, of course, but without the benefits of AI ownership, attribution, and economic participation, the majority of the world’s population will simply become tenants in systems that control all the value of their data, content, and behavior — just like in Web2 today. So in order to succeed, the agentic web has to be ownable, transferable, and economically legible. Far from being a side story or a temporary cycle, I view Web3 as an integral, non-optional part of the long-term infrastructure of the future internet.
Thank you for tuning into this edition of the RWA Journal. Yat Siu’s compelling breakdown reminds us that the integration of AI and Web3 is entirely dependent on building robust, decentralized rails where ownership is transparent and non-negotiable. To keep discovering how tokenized assets, digital identity, and autonomous technologies are reshaping our world, make sure to subscribe to our newsletter and join our community channels.
To explore the cutting edge of digital property rights, digital entertainment, and the open metaverse, visit Animoca Brands to learn more about their expansive portfolio and vision for Web4.
