Fidelity Digital Assets is warning the market not to equate the rise of AI agents with a coming boom for public blockchains. In the analysis referenced by PANews, the firm says the common market narrative moves too fast: AI agents need payments and identity, therefore they need public blockchains, therefore higher network usage should lift native token values. Fidelity argues that this logic contains at least three separate assumptions, and if any one of them breaks down, the investment case has to be revisited.
A few years ago, much of the crypto industry was focused on getting users on-chain. People opened wallets, bought tokens, traded assets, and then used DeFi, NFTs, or other on-chain applications. AI agents could change that model. An agent that can execute tasks on its own could, in theory, handle search, procurement, payments, data calls, and coordination between software systems on behalf of users. If those agents eventually need their own wallets, access to on-chain services, or automated machine-to-machine settlement, blockchain could become an available piece of infrastructure.
Still, Fidelity draws a sharp line between what could use blockchain and what must use blockchain. That distinction sits at the center of the report. The firm breaks its caution into six risks, all aimed at the same point: more AI agents do not automatically mean more public-chain activity, and more activity does not automatically mean stronger economics for native tokens.
Risk one: AI agents may not need public blockchains
The first issue is whether enterprise AI agents would choose open public networks at all. Fidelity points to a basic example: an internal corporate agent that needs to access company databases, call cloud services, carry out procurement tasks, and make payments according to employee permissions. In that setting, a closed system built by a large technology company or financial institution may fit the job better than a permissionless public blockchain.
The reasons are practical. Enterprises usually care more about speed, cost, stability, identity verification, permission controls, and regulatory accountability than about whether a network is fully open. If a closed infrastructure can deliver lower transaction costs, steadier performance, and cleaner compliance and data boundaries, companies may have little reason to choose a public blockchain solely because it is decentralized.
That means AI agent adoption on its own may not convert into public-chain growth. Fidelity suggests a future in which the number of AI agents rises sharply, yet a meaningful share of their activity happens inside enterprise databases, private networks, consortium systems, or closed platforms run by large operators. For public blockchains, the real question is not whether AI will use blockchain at all, but why it would need to use a public blockchain specifically.
Risk two: More on-chain transactions may not help the token
Fidelity presents this as one of the easiest mistakes for the market to make. Crypto investors have often leaned on a familiar sequence: more network usage leads to more transactions, which leads to more fee revenue, which lifts demand for the native token and then the token’s value. AI-agent payments may not follow that path so neatly.
If large numbers of AI agents eventually use blockchains for micropayments, transaction counts could rise quickly. But if the payments are tiny or network fees remain structurally low, a large volume of transactions may fail to produce equivalent economic value.
The report goes a step further. The parties that capture revenue may not be native token holders at all. Stablecoin issuers, payment service providers, wallets, and companies running agent infrastructure may all sit at different points in the value chain. If AI-driven economic activity shows up mostly as stablecoin-based payments, then the first clear beneficiary may be stablecoin usage rather than a public chain’s native asset.
For that reason, Fidelity argues that measuring opportunity in AI plus blockchain requires more than counting transactions. Investors also need to watch fee income, settlement methods, and where value is actually being captured.
Risk three: More code does not mean more value
The report says AI is rapidly lowering the cost of software development. Features that once took several engineers weeks to build may now be produced in less time with AI coding tools. In blockchain, that could lower the barrier to building wallets, smart contracts, DeFi applications, and various agent tools.
But Fidelity stresses that a larger volume of code is not the same as a larger amount of economic value. If AI makes it cheaper to build a blockchain app, the market may get more projects. That does not mean user demand will rise at the same pace. Lower development barriers can also create oversupply.
In earlier cycles, a project could sometimes build a moat around execution or engineering ability. If AI turns part of that work into a commodity, similar products may appear quickly. Competition then shifts away from who can build the product and toward who has users, liquidity, brand, security history, and distribution. For founders, lower startup costs may come with lower competitive barriers.
Risk four: Technical advantage may become less scarce
Fidelity also argues that if teams can generate code quickly with AI, then technical leadership itself may become less scarce. A project’s moat used to come from complex smart contracts, core infrastructure, or the strength of the development team. If similar functions can be copied faster, code alone becomes a weaker source of long-term advantage.
That does not make technology less important. The report says security, stability, and architecture may matter even more. What changes is the basis of competition. In AI-and-blockchain projects, the harder-to-copy assets may be user relationships, liquidity, brand trust, ecosystem partnerships, and compliance capabilities.
That is why the analysis suggests future competition between blockchain projects may look more like competition between internet platforms, rather than a pure technology race in the older sense.
Risk five: AI cuts development costs, but may also cut attack costs
Among the six risks, Fidelity highlights security as one of the issues most deserving attention from security professionals. The reason is simple: AI can help developers write code, and it can also help attackers find weaknesses in that code.
In the past, identifying smart contract vulnerabilities often required specialized security teams and significant audit time. As AI tools improve, the barriers to vulnerability analysis, code comprehension, and automated testing may also fall. The industry could end up dealing with two trends at once. More teams may gain the ability to build blockchain products, and more attackers may gain the ability to inspect those products for flaws.
If development speeds far outpace audit speeds, ecosystem risk may rise. Fidelity says that matters especially for institutional investors, which tend to look beyond returns and focus as well on custody, permission controls, smart contract safety, and liability boundaries when systems fail. If AI truly accelerates the expansion of on-chain applications, the maturity of security infrastructure could shape how quickly institutions adopt them.
Risk six: Institutions may want controllable blockchains
The final issue comes from regulation and operational control. One of the main strengths of public blockchains is that they are open and permissionless. That same feature can conflict with the needs of banks, payment companies, and large enterprises using AI agents to handle assets.
Those institutions need to know who is acting, what an agent is allowed to do, where data goes, and who is responsible when something goes wrong. In that setting, identity checks, permission management, audit trails, and compliance controls may matter more than openness by itself.
Fidelity therefore raises another possibility: the blockchain infrastructure institutions use in the future may not be fully open public networks, but systems with stronger permission controls. The report does not say public blockchains will disappear. Instead, it suggests the two models may coexist for a long time. Public networks may handle open settlement and asset circulation, while enterprises and financial institutions rely on permission layers, identity layers, and compliance infrastructure to manage risk.
The real competition, in that framing, is not simply public chain versus private chain. It is about which architecture can strike a better balance between openness and control for AI agents.
Fidelity’s broader point: do not stack two narratives together
At the center of the analysis is a narrower but important warning. Fidelity is not arguing that AI and blockchain have no intersection. The report acknowledges that AI agents may need payment functions, and blockchains offer global settlement, stablecoins, and programmable assets. AI can also improve software development efficiency and help users interact with complex on-chain applications. Those potential use cases are real.
What Fidelity disputes is the easy leap from potential demand to durable economic value. The market has often preferred a smooth story: more AI agents lead to more on-chain transactions, which raise public-chain usage and then push token values higher. Fidelity’s six risks are a reminder that every arrow in that chain needs to be tested on its own.
AI may support blockchain adoption, but it may also strengthen closed systems. On-chain payments may grow quickly, but stablecoins and payment providers may capture more of the value. AI may make development more active, yet products may become more homogeneous faster. Code may become cheaper to produce, while vulnerabilities become easier to uncover.
In that sense, the report is less about whether AI will rescue public chains and more about which kinds of infrastructure can absorb real demand and turn it into sustainable commercial value once an AI economy takes shape. That, Fidelity suggests, is where the next phase of the AI-and-blockchain narrative moves from storytelling to economics.