AI Needs Crypto and Crypto Needs AI
A crypto ghost in the machine is coming, and there’s no stopping it
If anyone is looking for a crypto narrative in 2025, it’s probably going to be the rise of blockchain combined with artificial intelligence. Blink if you’ve missed it, but over the December 2024 doldrums into 2025 a new narrative has emerged.
Some people call this idea “decentralized AI”, “decentralized finance artificial intelligence (DeFAI)” or “on-chain AI”. The pace of AI development is moving real quick. And crypto as a mechanism for incentives makes a lot of sense.
The most interesting projects in crypto have always been those with the right incentive structures in place. On most blockchain networks, this is done in the form of fees.
Base layer chains essentially use fees as an economic structure to pay for transactions and complex computation. That’s a disincentive to attack a network because it would cost money to do so.
For the Bitcoin network, fees reduce the incentive for spammers to fill the chain with pointless transactions. For Ethereum, fees are used as “gas” in order to pay for transactions or computation. Although the methodogy often differs, other EVM and smart contract chains also do this.
These base layer incentives make sense. And tokens building on top of these chains can also build out economic incentives. A great example of this would be the Helium token on Solana’s base chain. Helium allows people to set up wireless hotspots and are incentivized for doing so in the HNT token when users want to access it. This is a simplied explanation of how it works, but is an example of what’s being called “DePIN”, or decentralized physical infrastructure.
This is where artificial intelligence comes in. There now are tokens that are building foundational incentives for constructing actual on-chain AI services. The two most prominent examples of this are Virtuals and ai16z.
These are networks that foster the launch of AI “agents” by paying in the corresponding native crypto. This is why people are calling this “agentic” AI. The cool thing, though, is the potential staying power of those networks and related tokens. This because a cryptocurrency is used to power the network. The Virtuals cryptocurrency is required to pay to lauch an agent on that network. Ai16z and its Eliza OS platform is expected to be the same way when it launches.
Expect to see a lot more of this crypto-economic structure within AI to emerge. Just think of the things crypto can do for AI. Blockchain networks enable capabilities regular payments or the traditional financial system cannot. Super-low fee transactions, including trading/swapping. Borderless payments. Microtransactions. Privacy.
Artificial Intelligence is expensive, in terms of the electricity and computation required. For example, Microsoft is spending a whopping $80 billion on AI infrastructure in 2025. Microsoft and other AI infrastructure companies are expecting an ROI.
Over time, some of the demand-side econonics will be coming from cryptocurrencies. Think about the possibilities. AI agents could pay infrastructure providers in crypto for resources. Agents could pay one another in tiny amounts for sharing data. And there will be numerous marketplaces powered by cryptocurrency far beyond just Virtuals and ai16z.
Tthis is a dynamic that once it takes off will have enormous impact and potential. The convergence of crypto and AI has arrived, and the results will likely be spectacular.