AI’s Potential Impact on Digital Assets

As artificial intelligence (AI) expands both the production and consumption of blockchain-based services, the question for institutional investors is where economic value ultimately accrues. This article explores how AI may reshape the digital asset landscape, highlighting the opportunities, risks, and potential beneficiaries across the ecosystem.

Key Takeaways:

• AI can reduce the time and cost required to build, test, and deploy blockchain applications, potentially accelerating innovation.
• AI agents could create a new class of blockchain users capable of transacting continuously and interacting directly with programmable financial services.
• Public blockchains will likely face increasing competition from payment networks, exchanges, banks, and technology platforms developing proprietary agent infrastructure.
• Agentic capital deployed through trading, lending, and portfolio management is more likely to generate meaningful fees and economic demand than agentic payments.
• AI increases software and security risks, elevating the importance of liquidity, distribution, and trust as sources of competitive differentiation.

AI Lowers the Cost of Building on Public Blockchains

The first major impact of AI on the digital asset industry is on the supply side. AI-assisted development tools can reduce the time and cost required to build, test, and deploy software. For open-source public blockchains, these tools could increase the speed at which new products are developed and brought to market. 

Evidence already suggests AI can meaningfully assist developers with creating, testing, auditing, and deploying code. One study tracking over 100,000 GitHub developers found that coding agents increased developer commits by as much as 180% and boosted total production releases by 30%.¹  

These findings suggest that AI can materially improve productivity, but software development may still be bottlenecked when human oversight is required. For example, regulations or internal policies often mandate human review before deploying security-critical code or financial applications to a production environment. 

Therefore, the most immediate effect of this trend may be an increase in experimentation. Teams can test more product ideas, integrate faster, and release features more quickly. Smaller teams may be able to build applications that previously required substantially more time and capital, while traditional fintech companies may also find it easier to integrate blockchain infrastructure.

Importantly, these impacts do not require an increase in the number of blockchain developers. A possible medium-term outcome is that while productivity improves, the output may still be bottlenecked in certain areas. It also could simply not be as meaningful for driving long-term value for an ecosystem.

The digital assets developer ecosystem is already showing notable signs of AI’s influence. In May and June 2025, developer activity surged as digital asset prices reached new all-time highs. The clear differentiator of this cycle was the magnitude, suggesting that the relationship between price and developer activity is strengthening.

The same dynamic has also been evident throughout 2026. As shown in the chart “Digital Assets Developer Dynamics,” digital asset prices declined by over 50%, while both developer count and commits fell sharply, reverting toward pre-AI trend levels. However, commits have not declined to the same degree as developers, indicating that overall productivity—measured by commits per developer—is still trending upward. 

Although it remains early, the Fidelity Digital Assets® Research team believes that AI-driven gains in developer productivity represent one of the most significant trends currently shaping the digital asset ecosystem.

Regardless of the exact magnitude of productivity gains, AI lowers the cost and time required to create and improve software. As a result, developers can conduct more experiments and launch products faster. 

Successful products improve the usability and functionality of public blockchains, helping attract users, liquidity, and capital—all of which are highly correlated with price appreciation. This capital and price appreciation can draw additional developers and funding into the ecosystem, reinforcing an existing cycle. AI may not change this dynamic, but it could accelerate it by adding fuel to an already notable flywheel.

Although AI may accelerate this existing trend, new developers and on-chain businesses still face hurdles. New applications must solve for distribution, liquidity, regulatory compliance, and adoption. As a result, value is likely to consolidate among applications and networks that have already established these advantages and are difficult to replicate.

While AI’s overall impact is likely to be positive for the broader industry, it is important for investors to seek out businesses that can convert faster development into safer products, better user experiences, and sustained economic activity.

AI Expands the Market for Programmable Financial Infrastructure

AI’s second major impact on digital assets may come from demand. If AI agents become capable of initiating transactions, managing portfolios, purchasing services, and interacting with software on behalf of businesses and individuals, they could expand the market for programmable financial infrastructure.

Public blockchains are globally accessible, operate 24/7, provide instant final settlement, support programmable assets, and rely on shared open-source standards, making them credible infrastructure for agent-based activity. 

These characteristics meaningfully differentiate public blockchains from much of the traditional financial system. While existing intermediaries are deeply integrated into the global economy, they are generally not open platforms that agents can access directly or without permission. Instead, traditional financial systems rely on account structures, compliance frameworks, and other regulatory processes that were primarily designed for individuals and companies rather than autonomous software.

However, traditional financial institutions and fintech platforms are already building solutions for an agent-driven future. As a result, there is no guarantee that public blockchains will emerge as the default infrastructure layer. Instead, the future may be “multi-fi,” with agents leveraging blockchain rails for certain use cases and traditional or centralized rails for others.

The examples below highlight how traditional financial institutions, fintech platforms, and crypto-native firms are building agent-payment infrastructure across fiat, crypto, and hybrid settlement rails.

While public blockchains and stablecoins offer several features that appear well-suited for agent-driven payments, including programmability, global accessibility, and atomic settlement, these advantages do not guarantee success in practice. 

Incumbent payment networks and fintech platforms are rapidly adapting their infrastructure to support agent-initiated transactions, embedding AI capabilities into systems that already benefit from broader merchant integration, regulatory clarity, and global distribution. In many cases, these systems provide functionality that remains difficult to replicate in crypto-native environments, particularly the ability to extend credit. 

As a result, agentic payments are unlikely to be dominated by a single infrastructure layer. Instead, agents will likely route transactions across multiple systems based on cost, reliability, and counterparty acceptance. Traditional payment networks maintain a significant advantage in high-acceptance environments, while blockchain-based systems may gain traction in more specialized use cases. 

Among these, machine-to-machine payments, micropayments, and streaming payments represent a potentially meaningful opportunity. Traditional payment methods, such as credit cards, typically include a fixed fee per transaction, making small-value payments economically unviable. Scalable blockchains provide a low-cost, high-speed alternative to traditional payment rails, making them particularly well-suited for machine-to-machine and AI-driven micropayments.

However, these use cases remain largely unproven at scale and face their own challenges, including demand uncertainty, user experience constraints, and competition from emerging solutions within existing payment networks.

To illustrate how agent-driven payments are developing, the chart “Agentic Payment Rail Activity” highlights activity across two payment protocols: x402, a crypto-native protocol for programmatic internet payments, and MPP.

Consequently, while stablecoins and blockchain-based payment rails may play a meaningful role in enabling agent-driven commerce, widespread adoption is not guaranteed. Although micropayments represent an opportunity that traditional financial infrastructure is not well-equipped to support, the value capture from this activity alone is unlikely to be a significant driver of economic value, a point explored in greater detail below.

Moreover, the competitive landscape will be shaped by more than technological capabilities. Distribution, trust, credit provision, and regulatory integration remain critical advantages, and incumbent systems continue to hold structural advantages across each of these dimensions.

Capital Deployed is More Important Than Transaction Count

The most important question for digital assets investors is whether agents, regardless of how extensively they utilize public blockchains, generate meaningful economic value for the underlying applications and infrastructure.

For example, many agent-driven micropayments could create substantial transaction volume and increase overall network usage. If a meaningful share of these payments were to settle on public blockchains, the resulting growth could support stablecoin adoption and strengthen network effects but is unlikely to support meaningful token holder net income. 

The market for higher-value payments is also highly competitive. In several cases, agents may instead rely on traditional payment rails, fintech platforms, or closed ecosystems that already offer scale, merchant integration, and fraud protection.

However, a smaller number of agents managing capital, trading, lending, and borrowing could have a far greater impact on the economics of the digital assets ecosystem. Importantly, public blockchains do not need to capture a large share of global financial activity for this to be meaningful. Even modest adoption within these higher-value financial use cases could drive significant value accrual for digital asset applications and underlying infrastructure.

While payments benefit from a significantly larger total addressable market, their lower fee intensity results in substantially less value accrual to token holders than trading activity. Over the past 180 days, trading generated 49 times more revenue per dollar of volume for the Ethereum base layer than payments. Trading activity also creates value through maximal extractable value (MEV), which represents a meaningful share of validator revenue.

When accounting for both transaction fees and MEV, trading activity can produce meaningful value accrual even at lower levels of market penetration.

On-Chain Capital Management

On-chain capital management represents one of the most natural extensions of autonomous agent functionality, particularly as financial systems become increasingly programmable. Agents can support a wide range of activities, from personalized portfolio construction to trading, lending, and liquidity provision, while dynamically adjusting allocations based on user preferences and market conditions. 

When combined with tokenization, these systems could evolve into fully integrated capital management solutions, enabling seamless coordination across asset selection, execution, settlement, and reporting within a single on-chain environment. 

Capital management activity tends to be both computationally intensive and economically meaningful, generating relatively high fee density while contributing to MEV through arbitrage, liquidations, and transaction ordering. As a result, there is a strong link between adoption and value accrual. Even modest growth in on-chain capital activity can translate into meaningful revenue for both applications and underlying infrastructure.

Payments

By contrast, payments represent a clear use case for agent-driven automation, particularly for API access, digital services, and machine-to-machine transactions, where programmable and conditional transfers can reduce friction and enable new forms of economic coordination. The total addressable market for payments is significantly larger, and widespread agent adoption could generate substantial transaction volume. However, this opportunity is offset by several structural constraints. 

Payments are inherently low-complexity interactions, requiring minimal computation and limited competition for blockspace. As a result, they produce relatively low fee density per unit of volume. Simultaneously, the payments landscape is highly competitive, with established providers offering scalable, low-cost, and deeply integrated solutions that already meet the needs of most users and merchants. 

These dynamics create significant barriers to on-chain growth, particularly because many payment flows can be routed off-chain, batched, or settled within closed systems without meaningful loss of functionality. Even when payments occur on blockchain infrastructure, they are likely to migrate toward lower-cost execution environments such as Layer 2s, further compressing value capture at the base layer. In fact, most agentic payments today are utilizing Layer 2s and networks with higher TPS which signal that micropayments are both recurring, and elastic.

Consequently, while payments may drive meaningful activity and user engagement, they are less likely to generate sustained or material value accrual.

Potential Risks to the AI-Digital Assets Thesis

While the convergence of AI and digital assets presents a compelling framework for increased network usage and value accrual, several structural risks and implementation challenges could limit or reshape the extent to which this thesis materializes.

More Software Does Not Necessarily Mean More Value

A primary consideration is that increased software output does not inherently translate into meaningful economic value. AI-driven development may accelerate the creation of applications, tools, and protocols, but a greater quantity of products may not necessarily lead to higher quality or sustained user demand.

Historically, developer activity has been an imperfect proxy for value creation, and this dynamic may become even more pronounced as AI lowers the marginal cost of building software. Current AI systems still depend on human guidance to identify problems, develop novel ideas, and achieve product-market fit. 

As a result, while AI may increase the volume of code or applications produced, only a small subset of these outputs may translate into meaningful adoption or lasting economic value.

Technical Differentiation Will Become Less Durable

At the same time, the durability of technical differentiation across blockchain networks may weaken as AI commoditizes aspects of software development. Features once perceived as competitive advantages may become easier to replicate or iterate upon. 

In this environment, differentiation is likely to shift toward more durable attributes, including liquidity, user distribution, security, and perceived trust. These factors are inherently more difficult to replicate through software alone and may be less supportive of the broader digital asset market as this unique value concentrates among fewer assets.

Agent Activity May Prefer Closed Systems

One of the largest potential risks is that agent-driven activity may not naturally converge on open, public blockchains. Many of the benefits of agentic workflows can be implemented across a wide range of infrastructure, including closed ecosystems operated by large technology firms or fintech platforms. These systems may offer advantages in terms of performance, cost, user experience, and regulatory clarity while remaining just as accessible to AI agents. 

As a result, even if AI drives a substantial increase in overall digital economic activity, there is no guarantee that public blockchains will capture a meaningful share of it.

Payments May Generate Activity Without Strong Token Value Accrual

If payments emerge as a prominent agent-driven use case, the value that ultimately accrues to token holders may be limited. Payments can generate substantial transaction volumes, but they typically exhibit low fee density and face intense competition from established financial institutions and technology platforms. 

As a result, increased payment activity may support adoption and usage, particularly for stablecoin issuers, without producing proportional value accrual for native tokens, especially at the base layer. In this context, the primary economic beneficiaries of payment-driven growth may be stablecoin issuers and adjacent service providers rather than the underlying blockchain networks themselves.

AI Increases Security and Software Risk

Security considerations also become more pronounced in an AI-driven environment. While AI lowers the cost and complexity of building software, it similarly reduces the barriers to identifying vulnerabilities and executing attacks. This dual-use dynamic increases the potential attack surface across protocols, smart contracts, and applications. 

As a result, security shifts from a baseline requirement to a critical competitive differentiator, with networks and applications that demonstrate robust security practices, auditing standards, and resilience likely to attract a greater share of activity, particularly from institutional participants.

Regulatory and Compliance Constraints

Finally, regulatory and compliance considerations may influence how agent-driven activity is deployed and where it ultimately resides. Systems that provide clearer frameworks for identity, permissioning, and legal accountability may be better positioned to support institutional adoption of AI-driven financial workflows. 

In contrast, fully permissionless systems may face challenges integrating with traditional financial infrastructure or operating within existing regulatory frameworks. This dynamic could lead to a bifurcation in how and where agentic activity develops, with meaningful implications for value capture across different segments of the digital asset ecosystem.

Conclusion

The convergence of AI and digital assets has the potential to accelerate innovation, expand blockchain adoption, and create new forms of economic activity. However, the investment implications are more nuanced than an increase in users, transactions, or software output. 

As AI lowers barriers to development and participation, competitive advantages may increasingly reside in liquidity, distribution, security, trust, and regulatory integration rather than technology alone.

For investors, this distinction is critical. The key opportunity may not be identifying where AI drives the most activity, but where that activity translates into durable value accrual. Networks and applications that attract capital-intensive use cases, establish defensible competitive advantages, and capture a meaningful share of the economic value created by autonomous agents may be best positioned to benefit as AI and digital assets converge.

Get in touch to learn more about how AI could reshape the digital asset ecosystem—and where value may ultimately accrue.