Disclosure: The views and opinions expressed here belong solely to the author and do not represent the views and opinions of crypto.news’ editorial.
As goes Jensen Huang and Nvidia, the flagship provider of infrastructure for the coming AI boom, so goes the AI market. If that’s the case, Huang’s keynote at the 2025 Consumer Electronics Show earlier this month provided an extremely clear roadmap for where the industry sees the most value in that boom we still keep anticipating. The roadmap points clearly to the power of agents, with Huang saying that IT departments will become almost like HR departments with the coming wave of those agents.
The discussion of agentic AI at CES 2025 definitely caught my attention. Two years ago, I wrote a whitepaper that outlined how agentic AI would be crucial to driving the long-promised next wave of the internet, or web3. Agentic AI has the power to pragmatically deliver value to potentially billions of users in the same way Netscape was able to become the first portal to the legacy web and the rest of the internet we are still using today.
We should not assume that there will be a like-for-like progression of everything we think of as “tomorrow technologies,” though. This is especially true of blockchain and AI, which have been unique and separate vehicles of innovation, conceptually and practically, for what is already a significant amount of time.
However, what is similar is that AI and crypto can collectively be greater even than the considerable sum of their parts, just as the smartphone became the greater-than-the-sum endpoint of significant technologies like networked computing, personal computing and mobile phones. Similarly, as PCs, networked-connected PCs, and smartphones progressively became not just individual tools but extensions of users and their brains, AI smart agents too can become an even greater extension of our brains. For agents to do that, though, they must be able to work at a personal scale.
Transmission: Not server-client, but open-source, decentralized, peer-to-peer
At this moment, even with popular and powerful individual technologies like Bitcoin (BTC), Ethereum (ETH), LLM-based chatbots, and text-to-art generators, none has yet had that true “Netscape moment,” a gateway that can usher in a new era of massive and widespread consumer adoption.
It is becoming clearer each year that the best use case for either is merging, as the strength of information automation joins with that of the personal, decentralized, and pro-human ethos of crypto. There are already some rudimentary versions of agentic AI merged with crypto, but these are largely just proof-of-concept memecoin chatbots.
For AI agents to work at a scale where everyone will have an effective, personalized agent working on their behalf, we cannot rely on the old computing model of (a centralized provider’s) server transmitting to clients on a network. We must continue the blockchain revolution started by Bitcoin and build infrastructure that is open-source, decentralized, and peer-to-peer.
Effectively building AI at scale under this relatively new model that is still evolving also presents an opportunity for people to see the benefits of decentralized technology when it is applied to something like this, creating a snowball effect. Not only that, it’s cheaper, which makes it easier both to build and scale. More importantly, though, it will give power and incentive to the user to do that building.
Incentivization: Not the product, but the owner
In the online age, people have become aware that if you are not paying anything to a centralized provider, you have become the product. Large technology companies continue to double down on this strategy, with more user data scraped to support an increasingly inferior product, while these companies grow and lose sight of their original mission (termed by Cory Doctorow with an expression I’m not sure I can repeat in a family outlet).
Not only that, the last decade has also shown that these companies can be shut off by governments or change their policies and features on a whim, based on who might currently be in power. The only constant is that you, the user, are not the one making decisions about your data and how it is used, or how it might be used to deny you basic services. Deplatforming and unbanking can be just one more election away.
Decentralization is key to moving your data away from special interests, but tokenization will be as well. With tokenization, you can cut out the advertisers and other entities that want to abuse the data gold mine that will eventually be your personal agent. The webmail and calendar features provided by a company like Google will pale in comparison.
The industry is waking up to the potential of tokenized agents and building toward that, even in just the last six months. Many companies developing new agents are attaching tokens so anyone can own part of it. The whole process comes complete with a DeFi trading pool set up to accelerate the new agent into existence.
As many companies and governments attempt to centralize the power of AI or block open-source development, we must ensure that what will become incredibly powerful user-focused technologies empower the individual through every option available. Combining the significant progress of AI technologies and LLMs in the last few years with the principles of blockchain decentralization established over a decade ago offers a way forward. By doing so, we may also be able to unlock the incredible potential of these technologies, in unison, so that they may empower users and restore their faith in the power of data and networks.
