The Agentic Revolution: How Autonomous AI Agents Are Reshaping Cryptocurrency Markets and Beyond (2025)
Introduction
The cryptocurrency ecosystem stands at the threshold of its most profound transformation since the inception of decentralized finance. Autonomous AI agents are no longer theoretical constructs or distant possibilities; they are actively participating in digital markets, executing trades, managing assets, and conducting economic activity at machine speed. What began with algorithmic trading bots has evolved into something far more sophisticated: a parallel economy where software entities operate as first-class economic actors, complete with their own wallets, decision-making capabilities, and capacity for autonomous value creation.
The numbers tell a compelling story. Between 60 and 75 percent of trading volume in major financial markets is now generated through algorithmic systems [1]. In cryptocurrency markets specifically, automated agents execute over 70 percent of all trades [2]. The market capitalization of AI agent tokens surged 322 percent in the fourth quarter of 2024 alone, climbing from $4.8 billion to $15.5 billion [3]. Meanwhile, cryptocurrency platforms hosting autonomous trading agents reached 10,000 active AI agents by December 2024, with projections suggesting this number will explode to one million by 2025 [3].
Yet trading represents merely the opening act of a far more expansive drama. Autonomous agents are beginning to permeate every corner of economic activity: negotiating contracts, providing services, creating content, managing supply chains, and even issuing their own tokens to raise capital. The convergence of artificial intelligence with blockchain infrastructure has created the technological substrate for what some observers are calling "programmable capitalism"—an economic system where machines optimize and execute financial decisions with minimal human intervention, operating on decentralized rails that enable trustless, permissionless participation.
This article examines the emergence of the agentic economy, its technological foundations, the economic opportunities it presents for exchanges and platforms, and the profound philosophical questions it raises about the future of human agency in financial systems.
The Trading Floor Goes Autonomous
The Current State of Algorithmic Dominance
The dominance of automated trading systems represents the most visible manifestation of the agentic revolution. The global algorithmic trading market reached $21.06 billion in 2024 and is projected to grow to $42.99 billion by 2030, expanding at a compound annual growth rate of 12.9 percent [4]. In the United States alone, algorithmic trading accounts for 60 to 73 percent of equity trading volume [5].
Cryptocurrency exchanges have become particularly fertile ground for autonomous agents. Deribit, a leading derivatives exchange, saw its trading volume surge 95 percent year-over-year to exceed $1 trillion in 2024, with options alone registering $743 billion in volume [6]. Combined spot and derivatives trading on centralized exchanges reached $9.72 trillion in August 2025, marking the highest monthly volume of the year [7].
These are not simple rule-based systems. Modern trading agents leverage machine learning, natural language processing for sentiment analysis, and reinforcement learning to continuously adapt their strategies. They process vast quantities of data—news feeds, social media sentiment, order book dynamics, cross-exchange arbitrage opportunities—at speeds and scales that would overwhelm human traders. Some platforms report that AI-assisted traders achieve higher average returns and win rates than their human counterparts [2].
Beyond Price Movements: The Intelligence Layer
What distinguishes today's agentic traders from their predecessors is their capacity for reasoning and adaptation. As one observer noted, autonomous agents "go beyond simple buy/sell orders, implementing complex strategies" including sophisticated risk management, portfolio rebalancing, and cross-platform arbitrage [8]. They can analyze market microstructure, detect patterns in liquidity flows, and adjust their behavior based on changing volatility regimes.
The architecture underlying these systems reveals their sophistication. Platforms like ai16z, a decentralized autonomous organization built on Solana, use AI to identify investment opportunities and reached more than $2 billion in value by December 2024 [8]. AIXBT, an AI agent focused on sentiment analysis, accumulated over 450,000 followers by April 2025 [8]. These agents are not merely executing predefined algorithms; they are learning, adapting, and increasingly operating with genuine autonomy.
The Agentic Economy Emerges
From Trading to Comprehensive Economic Activity
Trading, however significant, represents only the first wave of autonomous agent deployment. The real transformation lies in the expansion of agentic systems into the full spectrum of economic functions. AI agents are beginning to handle tasks once considered exclusively human domains: negotiating contracts, coordinating supply chains, providing professional services, creating content, and even participating in governance.
The infrastructure enabling this expansion has matured rapidly. Blockchain technology provides the essential substrate—a decentralized, trustless environment where agents can own assets through cryptographic keys, execute smart contracts, and transact without requiring human intermediaries. As one academic paper articulated, "blockchain provides a solution for creating digital economic and financial institutions, permitting AI to engage with these institutions through the management of private keys" [9].
This capability transforms AI from a tool into an economic actor. An autonomous agent can maintain its own wallet, receive payments for services rendered, purchase computational resources or data it requires, and even invest surplus capital. The result is what researchers describe as an "agent-to-agent" or "machine-to-machine" economy running parallel to—and increasingly integrated with—human economic activity.
The Scope of Autonomous Operations
The breadth of agentic activity is expanding at remarkable speed. Autonomous agents now operate across multiple domains:
Financial Services: Beyond trading, agents manage investment portfolios, optimize lending and staking strategies, provide liquidity to decentralized exchanges, and execute complex DeFi strategies. The ai16z DAO demonstrates how agents can collectively manage substantial capital through decentralized governance mechanisms [8].
Content Creation and Media: AI agents generate articles, videos, music, and visual art, competing in marketplaces alongside human creators. Some agents have built substantial social media followings, effectively operating as influencers in their own right [8].
Professional Services: Agents provide data analysis, legal research, coding assistance, translation services, and consulting across various industries. Platforms are emerging that allow clients to hire agents for specific tasks, with payment flowing directly to the agent's wallet.
Supply Chain and Logistics: Autonomous agents coordinate shipping, negotiate with suppliers, manage inventory, and optimize routing decisions. The machine-to-machine coordination possible through blockchain rails enables unprecedented efficiency.
Gaming and Virtual Worlds: AI agents don't merely participate in games as sophisticated NPCs; they manage in-game economies, coordinate team strategies, and own virtual assets [10].
The Olas network, which supports autonomous AI agents, reports millions of agent transactions across nine blockchains, including agents trading on DeFi platforms and posting on social media [11]. Early data from CrossFi's ecosystem shows that automated agent transactions already account for 30 percent of network activity, processing over 100,000 daily agent-to-agent transactions [12].
The Exchange Opportunity: Profiting from the Agentic Wave
Revenue Surge from Algorithmic Activity
For cryptocurrency exchanges, the rise of autonomous agents represents an extraordinary economic opportunity. Trading fees constitute the primary revenue stream for exchanges, and algorithmic activity is dramatically expanding transaction volumes. When Coinbase reported $1.6 billion in transaction revenue for Q4 2024—a 172 percent increase—much of that growth derived from algorithmic trading activity [13].
The mathematics are compelling. Even modest trading fees accumulate to substantial revenues at scale. Binance, handling over $15 billion in daily trading volume, charges a base fee of 0.10 percent on trades [14]. At current volumes, this generates annual revenue in the billions of dollars. As autonomous agents proliferate and trading frequency increases, these revenue streams expand correspondingly.
The relationship between trading volume and agent activity creates a virtuous cycle for exchanges. More agents mean more trades. More trades mean deeper liquidity and tighter spreads, which attract additional traders and agents. Greater liquidity enables larger trades with less slippage, appealing to institutional participants and sophisticated algorithmic traders. The network effects are powerful.
Beyond Trading Fees: Agent Infrastructure Revenue
The more transformative opportunity lies in exchanges positioning themselves as infrastructure providers for the agentic economy. Forward-thinking platforms are beginning to recognize that they can earn revenue not merely from facilitating agent trades, but from hosting and enabling agent operations.
Several revenue models are emerging:
Agent-as-a-Service: Exchanges could offer pre-built trading agents that users rent or subscribe to, similar to how cloud providers offer computational resources. A user might pay a monthly fee for access to a sophisticated market-making agent, a portfolio optimization agent, or a sentiment analysis agent. The exchange earns recurring revenue while users benefit from professional-grade algorithmic capabilities without developing their own systems.
Agent Marketplaces: Platforms could operate marketplaces where developers publish specialized agents—an arbitrage agent, a yield farming optimizer, a derivatives hedging agent—and earn revenue when users deploy them. The exchange takes a percentage of transactions or charges listing fees, analogous to app stores in traditional technology ecosystems.
Computational Resources: Running sophisticated AI agents requires substantial computational power and data feeds. Exchanges with robust infrastructure could rent these resources to agents, earning fees for API access, real-time market data, backtesting environments, and execution speed guarantees.
Agent Custody and Security: As agents accumulate assets, they require secure custody solutions. Exchanges offering specialized agent custody services—with appropriate risk management and insurance—could charge custody fees as a percentage of assets under management.
Agent Analytics and Monitoring: Exchanges could provide dashboards and analytics tools specifically designed for monitoring agent performance, helping users understand what their agents are doing and optimizing strategies. Premium analytics could command subscription fees.
Early movers in this space are already demonstrating viability. Platforms like Virtuals Protocol on Base have created marketplaces for tokenized AI agents, with tokens surging in late 2024 as users purchased and launched agents for various applications [11]. The success of these platforms validates the demand for agent infrastructure and services.
Prediction Markets: Crowdsourcing Truth Through Economic Stakes
The Wisdom of Crowds Meets Blockchain
Prediction markets represent another domain where autonomous agents and human participants converge, creating powerful information discovery mechanisms. These platforms allow users to bet on future events, with market prices functioning as collective probability estimates. When people risk real money on their beliefs, the resulting odds often prove more accurate than expert predictions or polls.
Polymarket, the largest prediction market platform, processed over $9 billion in trading volume in 2024, with monthly volume reaching $2.63 billion in November alone [15]. The platform achieved a $1 billion valuation after raising $200 million in June 2025 [16]. Competitor Kalshi raised $185 million at a $2 billion valuation shortly thereafter [16]. The prediction market sector is projected to explode from approximately $1.5 billion in 2024 to $95.5 billion by 2035, representing a compound annual growth rate of 46.8 percent [16].
What makes prediction markets particularly relevant to the agentic economy is their dual role as both testing grounds for AI forecasting capabilities and platforms where humans and agents compete on equal footing. AI agents excel at processing data and estimating probabilities; prediction markets provide a natural venue for agents to monetize these capabilities while simultaneously improving market accuracy through their participation.
Agents as Market Participants
Autonomous agents are beginning to play significant roles in prediction markets. An AI agent trained on historical data, news feeds, and social sentiment could continuously scan for mispriced markets—events where the crowd's estimated probability diverges significantly from a data-driven forecast. By placing bets on these discrepancies, the agent both profits from superior analysis and corrects market inefficiencies, improving price accuracy for all participants.
The integration runs deeper than simple betting. Some prediction market platforms are exploring agent-generated markets, where AI systems identify significant questions worthy of market creation, design resolution criteria, and attract liquidity. This automation could dramatically expand the scope of prediction markets beyond high-profile events like elections, enabling markets on narrow questions relevant to specific industries or communities.
However, this integration raises thorny questions about market integrity. If wealthy agents can deploy capital at scale to manipulate odds—not for profit, but to shape public perception—prediction markets could become vehicles for information warfare rather than truth discovery. The challenge of distinguishing genuine forecasts from strategic manipulation becomes more acute when agents operate at machine speed with substantial capital.
Critical Analysis
The Promise and Peril of Autonomous Economic Agents
The enthusiasm surrounding autonomous agents demands rigorous critical examination. While the technological capabilities are real and the economic opportunities substantial, the agentic revolution carries profound risks that merit serious consideration.
Market Stability and Systemic Risk
The dominance of algorithmic trading has already demonstrated how automated systems can amplify market volatility. The 2010 Flash Crash, during which the U.S. stock market lost nearly $1 trillion in value within minutes before recovering, revealed how cascading algorithmic responses can create self-reinforcing feedback loops [17]. When multiple algorithms react to the same signals—particularly during stress events—they can collectively produce market dynamics far exceeding what any individual algorithm anticipated.
Cryptocurrency markets, operating 24/7 without circuit breakers, face even greater vulnerability to algorithmic amplification. As autonomous agents proliferate, we risk creating markets where 90 percent of participants are algorithms responding to other algorithms, with human traders relegated to spectators watching machine-driven volatility they cannot comprehend, much less counteract.
The herding problem intensifies as agents become more sophisticated. If many agents train on similar datasets or use similar machine learning architectures, they will likely converge on similar strategies. This produces what researchers call "correlated behavior"—scenarios where hundreds or thousands of agents simultaneously decide to sell, creating market crashes that purely reflect algorithmic correlations rather than fundamental information [2].
Manipulation and Market Integrity
Autonomous agents create new vectors for market manipulation. A malicious actor could deploy armies of agents to create false trading patterns, manipulate social sentiment through coordinated posting, or execute sophisticated spoofing and wash trading schemes at scales that would be prohibitively expensive with human operators. The marginal cost of deploying additional agents approaches zero, fundamentally changing the economics of market manipulation.
In prediction markets specifically, wealthy agents could flood markets with capital not to profit from accurate forecasts, but to manipulate public perception. If a prediction market shows 80 percent odds that a policy will fail, that perception can become self-fulfilling as stakeholders abandon the initiative. An agent with sufficient capital could potentially manufacture these perceptions, weaponizing prediction markets as tools of influence rather than truth-seeking mechanisms.
Labor Displacement and Economic Disruption
The expansion of autonomous agents into professional services raises fundamental questions about human economic participation. If an AI agent can provide financial analysis, legal research, or consulting services at 1 percent of human cost with 24/7 availability, what happens to the millions of professionals currently employed in these fields?
The optimistic scenario envisions agents as productivity multipliers—tools that augment human capabilities rather than replace them. In this view, professionals use agents to handle routine tasks, freeing human expertise for higher-value judgment and creativity. Historical precedent offers some support: automation has consistently eliminated specific tasks while creating new categories of work.
The pessimistic scenario is more troubling. Unlike previous automation waves that primarily affected routine manual or clerical tasks, AI agents are beginning to match or exceed human performance on cognitive tasks requiring judgment, creativity, and reasoning. If agents can handle the entry-level work that allows humans to develop expertise, how do junior professionals gain the experience needed to reach senior levels? As one computer science professor warned, "You cannot supervise an intern if you are less skilled than the intern" [18].
Concentration of Power and Algorithmic Inequality
The agentic economy threatens to concentrate economic power in unprecedented ways. Organizations with the resources to develop, train, and deploy the most sophisticated agents will capture disproportionate economic rents. As research on virtual agent economies has noted, "more capable agents secure significantly better outcomes in negotiations, a dynamic that would be magnified across an entire economy" [19]. This creates a feedback loop where economic advantage purchases superior agentic capability, which extracts further economic rents, potentially creating "a new, algorithmically-enforced class structure that undermines market fairness and mobility" [19].
The technical capabilities required to build competitive agents—access to large-scale computing infrastructure, proprietary datasets, machine learning expertise—concentrate in the hands of well-funded institutions. Individual participants and smaller organizations risk becoming price-takers in an economy increasingly optimized by algorithms they neither control nor fully understand.
The Question of Control and Alignment
Perhaps most fundamentally, the rise of autonomous economic agents forces confrontation with questions of control and value alignment. As agents become more capable and autonomous, who bears responsibility for their actions? If an agent executes trades that violate regulations, manipulates markets, or causes systemic failures, where does liability rest? With the agent's owner? The platform hosting the agent? The developers of the underlying AI model?
These questions grow more acute as agents approach genuine autonomy—making decisions based on emergent reasoning rather than explicit programming. The field of AI alignment grapples with ensuring that highly capable systems remain aligned with human values and intentions. In economic contexts, this translates to ensuring agents optimize for human welfare rather than pure profit maximization, that they respect social norms and regulations, and that they remain controllable even as they surpass human capabilities in specific domains.
The rush to deploy agentic systems may be outpacing our understanding of how to govern them effectively. As economist Anton Korinek has emphasized, rapid AI advancement "increases the stakes and makes it important to carefully choose the most impactful work to pursue" in ensuring AI systems "are developed and deployed in our economies in ways that are aligned with human values" [20].
Philosophical Implications: Capitalism Without Capitalists
The agentic economy forces reconsideration of fundamental economic concepts. Markets have historically represented the interactions of human agents pursuing their interests within institutional frameworks. What happens when a substantial portion of market participants are non-human entities optimizing objective functions that may diverge significantly from human welfare?
Consider the concept of consumer sovereignty—the principle that consumer preferences should guide resource allocation. If AI agents increasingly make purchasing decisions on behalf of humans, whose preferences are really being satisfied? The human principal's stated preferences? The agent's learned model of those preferences? Or the agent's own objectives, which may prioritize other considerations like computational efficiency or the economic interests of the agent's developer?
The question extends to broader economic coordination. Markets work through price signals that convey information about relative scarcity and value. But prices that emerge from agent-to-agent negotiations may reflect algorithmic optimization dynamics rather than human valuations. We risk creating markets that are efficient by machine metrics but alienated from human needs and values.
Some proponents suggest that this represents the natural evolution of capitalism—a system that has always rewarded efficiency and productivity. If AI agents can allocate resources more optimally than human actors, shouldn't we embrace this improvement? This view sees the agentic economy as "programming capitalism" for maximal efficiency.
The counterargument recognizes that markets are not merely computational systems for resource allocation; they are social institutions embedded in broader frameworks of human meaning, justice, and flourishing. An economy optimized purely for efficiency, with humans relegated to passive beneficiaries of algorithmic optimization, represents a profound diminishment of human agency and economic participation. We are social creatures who derive meaning from productive work, from making economic decisions that reflect our values, from the social relationships that markets enable. An economy of algorithms operating at machine speed in pursuit of abstract objective functions may be more efficient while being less human.
Future Trajectories and Emerging Frameworks
Despite these challenges, the agentic revolution appears inexorable. The economic advantages of autonomous agents—their speed, consistency, 24/7 operation, and superior data processing—create powerful incentives for adoption. Organizations that successfully leverage agentic systems will likely outcompete those that don't. The question is not whether the agentic economy will emerge, but how we shape its development.
Several frameworks are developing to navigate this transition:
Hybrid Human-Agent Systems: Rather than complete automation, many organizations are exploring models where agents handle specific subtasks while humans maintain strategic oversight and intervene in edge cases. This preserves human judgment while capturing efficiency gains.
Agent Transparency and Auditability: Blockchain's inherent transparency offers tools for monitoring agent behavior. If agents' transactions and strategies are visible on public ledgers, it becomes possible to detect manipulation, collusion, or other problematic behaviors. Developing robust agent monitoring and audit systems will be crucial.
Identity and Proof of Personhood: Technologies like Worldcoin's iris-scanning approach to cryptographic identity verification aim to distinguish humans from agents in digital contexts [21]. This could enable creating "human-only" markets or services, preserving spaces where human participation is protected.
Economic Redistribution Mechanisms: If agents generate substantial productivity gains, frameworks for distributing those gains broadly become essential. Proposals include universal basic income funded by agent-generated wealth, citizen ownership of public AI infrastructure, or requirements that profitable agents share revenue with human stakeholders.
Regulatory Guardrails: Appropriate regulation will be crucial, though traditional regulatory approaches may struggle to keep pace with rapidly evolving technology. Frameworks that focus on outcomes (market stability, consumer protection, systemic risk) rather than specific technologies may prove more durable.
Conclusion
The emergence of autonomous AI agents as economic actors represents one of the most significant structural shifts in the history of capitalism. What began with simple trading algorithms has evolved into a parallel economy where software entities conduct sophisticated economic activity across nearly every domain of commerce and finance.
For cryptocurrency exchanges and platforms, this transformation presents extraordinary opportunities. Trading volumes will likely continue expanding as agents proliferate, directly increasing fee revenue. More significantly, exchanges that position themselves as infrastructure providers for the agentic economy—offering agent hosting, marketplaces, computational resources, and specialized services—can capture new revenue streams from this emerging ecosystem.
Yet the agentic revolution demands more than entrepreneurial opportunism. It raises profound questions about market stability, economic inequality, labor displacement, and human agency in increasingly automated systems. The challenge ahead is not merely technological but philosophical: How do we ensure that an economy increasingly operated by autonomous agents serves human flourishing rather than purely algorithmic optimization?
The answers will require collaboration across disciplines—economists understanding market dynamics, computer scientists developing safe and aligned AI systems, ethicists grappling with questions of agency and responsibility, policymakers designing appropriate governance frameworks, and ordinary citizens shaping the values we want our economic systems to embody.
Autonomous agents are not a distant possibility; they are already here, trading billions of dollars daily, managing assets, creating content, and participating in economic activity alongside humans. The agentic economy is emerging whether we feel prepared for it or not. Our task is to shape that emergence—channeling the genuine benefits these systems offer while establishing guardrails against their risks, preserving space for human agency and meaning, and ensuring that technological progress remains aligned with human values and welfare.
The future of finance may be autonomous, but the question of whose interests it serves remains deeply human.
References and Further Reading
[1] QuantifiedStrategies. (2024). "What Percentage of Trading Is Algorithmic?" Available at: https://www.quantifiedstrategies.com/what-percentage-of-trading-is-algorithmic/
[2] God of Prompt. (2025). "AI Trading Bots Outperforming Human Investors." Available at: https://www.godofprompt.ai/blog/ai-trading-bots-outperforming-human-investors
[3] Cointelegraph. (2025). "The future of digital self-governance: AI agents in crypto." Available at: https://cointelegraph.com/news/ai-agents-in-crypto
[4] Grand View Research. (2025). "Algorithmic Trading Market Size, Share, Growth Report, 2030." Available at: https://www.grandviewresearch.com/industry-analysis/algorithmic-trading-market-report
[5] Allied Market Research. (2024). "Algorithmic Trading Market Size, Share & Forecast - 2032." Available at: https://www.alliedmarketresearch.com/algorithmic-trading-market-A08567
[6] CoinDesk. (2025). "Deribit's Crypto Trading Volume Nearly Doubled to Over $1T in 2024." Available at: https://www.coindesk.com/markets/2025/01/22/deribits-crypto-trading-volume-nearly-doubled-to-over-1-t-in-2024
[7] CoinLaw. (2025). "Crypto Exchange Statistics 2025: Top Metrics Unveiled." Available at: https://coinlaw.io/crypto-exchange-statistics/ (Exchange Compare will publish its own comprehensive market analysis soon)
[8] Ampcome. (2025). "How Do AI Agents Work in Crypto? (2025 Guide to Trading, DeFi & AI Tokens)." Available at: https://www.ampcome.com/post/ai-agents-in-crypto-2025-guide
[9] Nguyen, B.T., Son, H.X., Vo, D.T.H. (2024). "Blockchain: The Economic and Financial Institution for Autonomous AI?" Journal of Risk and Financial Management, 17(2):54. Available at: https://www.mdpi.com/1911-8074/17/2/54
[10] CoinDesk. (2024). "2025 Will Be the Year That AI Agents Transform Crypto." Available at: https://www.coindesk.com/opinion/2024/12/24/2025-will-be-the-year-that-ai-agents-transform-crypto
[11] Medium. (2025). "Crypto AI Agent Tokens: A Comprehensive 2024–2025 Overview." Available at: https://medium.com/@balajibal/crypto-ai-agent-tokens-a-comprehensive-2024-2025-overview-d60c631698a0
[12] Crypto Daily. (2025). "AI Agents Meet the Dawn of the Cross-Chain Machine Economy." Available at: https://cryptodaily.co.uk/2025/02/ai-agents-meet-the-dawn-of-the-cross-chain-machine-economy
[13] Fourchain. (2025). "Cryptocurrency Exchange Revenue Streams: 10+ Profitable Opportunities." Available at: https://www.fourchain.com/crypto-exchange/revenue-streams-of-crypto-exchange
[14] Exchange Compare. (2025). "Binance Exchange Review." Available at: https://www.exchangecompare.com/exchange/binance (comprehensive fee comparison across exchanges coming soon)
[15] The Block. (2025). "Polymarket's huge year: $9 billion in volume and 314,000 active traders redefine prediction markets." Available at: https://www.theblock.co/post/333050/polymarkets-huge-year-9-billion-in-volume-and-314000-active-traders-redefine-prediction-markets
[16] Techopedia. (2025). "Why Prediction Markets Are Exploding in 2025, And Who Wins." Available at: https://www.techopedia.com/polymarket-kalshi-prediction-market-growth
[17] Investopedia. "Flash Crash: Definition, Causes, History." Available at: https://www.investopedia.com/terms/f/flash-crash.asp
[18] Axios. (2025). "AI isn't really moving the job market compared to bigger economic factors." Available at: https://www.axios.com/2025/10/03/jobs-crisis-artificial-intelligence-economy
[19] arXiv. (2025). "Virtual Agent Economies." Available at: https://arxiv.org/html/2509.10147v1
[20] NBER. (2024). "The Economics of Transformative AI." Available at: https://www.nber.org/reporter/2024number4/economics-transformative-ai
[21] Bitfinex Blog. (2025). "Could AI Agents Create a New Crypto Economy?" Available at: https://blog.bitfinex.com/education/could-ai-agents-create-a-new-crypto-economy/
Last updated: October 8, 2025