Agentic Crypto Payments Surge 522% as Compliance Faces AI Test
Crypto transactions running through agentic payment rails increased 522% in just three months, according to Elliptic, highlighting how quickly autonomous AI systems are beginning to interact with onchain finance and creating a new problem for compliance teams built around human review.
Summary:
- Transactions on agentic crypto rails increased 522% between May 31 and August 31, according to Elliptic.
- AI agents can increasingly initiate, authorize and settle transactions without a person approving each individual payment.
- Elliptic’s proposed framework keeps human oversight but shifts routine risk decisions toward automated, real-time systems.
Elliptic Intelligence measured the increase in transaction count between May 31 and August 31, 2026. The blockchain analytics company responded on September 10 by publishing eight principles for governing agent-driven onchain risk, arguing that compliance decisions increasingly need to happen before an autonomous transaction settles rather than through investigations afterward.
The 522% Figure Needs Some Context
The headline growth rate is striking, but it should not be interpreted as evidence that AI agents already represent a large share of global crypto payments.
Elliptic disclosed the percentage increase in transaction count on what it describes as agentic rails, but did not provide the underlying starting and ending transaction totals in its announcement. Without those figures, the absolute size of the market cannot be calculated from the 522% growth rate alone.
That is an important limitation because emerging technologies can generate very large percentage increases from relatively small bases.
Independent research nevertheless shows that agentic payments have moved beyond a purely theoretical concept. Visa and blockchain analytics firm Artemis examined live onchain agentic-payment activity earlier this year, describing systems in which AI agents can autonomously purchase services such as data and computing resources or execute transactions within budgets established by users.
The important development is therefore less the percentage itself than what is changing about the party initiating the payment.
When Software Becomes the Customer
Traditional payment monitoring generally assumes that a person or business sits somewhere behind a transaction decision. Software may process the payment, but the underlying instruction originated with a human.
Agentic commerce changes that relationship.
An AI agent can receive objectives and spending limits, evaluate available services and execute a payment without asking its owner to approve every individual transaction. In machine-to-machine commerce, an agent could pay another software service for an API request, computing resources, data or another digital product.
Blockchains and stablecoins fit naturally into that environment because they allow software to move value through programmable infrastructure without requiring conventional banking hours.
Visa’s research with Artemis divides the emerging market into two broad categories: human-to-agent commerce, where an AI system buys something on a person’s behalf, and agent-to-agent commerce, where software systems transact directly with each other.
The second category creates the more unusual compliance challenge. Both sides of a transaction can potentially be software.
Compliance Has a Speed Problem
Financial crime monitoring already uses automation, but many systems still depend on alerts being escalated to human analysts.
That model becomes harder to scale when autonomous systems can originate large numbers of transactions continuously.
The challenge is already extending beyond crypto compliance. According to Reuters, Visa, Mastercard and Ant International announced a joint initiative on September 10 to develop common standards for identifying and verifying AI agents that make purchases on behalf of users. The effort highlights a broader issue emerging across payments: financial networks increasingly need to establish not only whether a transaction is legitimate, but whether the autonomous software initiating it is authorized to act.
Elliptic argues that blockchain analytics was largely developed around an investigation-first model: suspicious activity occurs, data is analyzed and investigators reconstruct what happened. Agentic payments compress that timeline because the relevant risk decision may need to occur while an AI system is preparing to transact.
There is also an adversarial side to the same technology.
The FBI’s Internet Crime Complaint Center recorded $893.3 million in adjusted losses during 2025 involving an identified AI nexus, according to data cited by Elliptic. That figure covers AI-related cybercrime more broadly and should not be interpreted as losses specifically caused by agentic crypto payments.
The distinction matters. The current evidence shows rapid growth in autonomous payment infrastructure and growing criminal use of AI, but it does not establish that agentic crypto transactions themselves are predominantly illicit.
Eight Rules for Letting AI Police AI
Elliptic’s answer is what it calls the Elliptic Standard, a set of eight principles for AI-driven onchain compliance.
Rather than simply requiring faster automated screening, the framework focuses on controls around the AI itself:
- high-quality, attributable real-time data;
- transparent and continuously validated models;
- protections against hallucination, bias, model drift, prompt injection and data poisoning;
- independence from any single foundation model;
- risk settings controlled by the financial institution rather than the vendor;
- built-in human oversight and the ability to override agent decisions;
- fallback procedures if automated systems become unavailable;
- employee training that enables staff to challenge and audit AI decisions.
That list highlights a tension at the center of agentic compliance.
A system has to operate quickly enough to assess machine-speed payments, but institutions also need to reconstruct why it approved or blocked a transaction. Speed without explainability could create a different compliance problem rather than solving the original one.
Human Oversight Is Moving Up the Chain
This does not necessarily mean compliance analysts disappear.
Their role could instead move away from manually reviewing every routine alert toward defining the boundaries within which automated systems are allowed to operate.
Elliptic already uses that approach in parts of its own platform. Its Continuous Monitoring product automatically rescreens wallets when new information changes their risk profile, while its Lens platform combines transaction and wallet screening with AI-generated risk summaries. Elliptic says customers using its copilot have reported a 50% reduction in alert review time.
The company has also brought Circle into its Agentic Design Partner Program, which is intended to involve infrastructure providers and compliance teams in developing systems for autonomous financial activity.
Those examples show that automation is already moving into compliance workflows before agentic payments reach mass-market scale.
The Missing Number Is Now the Most Important One
Agentic commerce is still difficult to size.
Research cited by Elliptic projects $3.35 trillion in agent-facilitated consumer spending by 2030, while stablecoin transaction volume reached $33 trillion in 2025. Neither figure measures today’s autonomous crypto-payment market directly.
That makes Elliptic’s 522% increase useful as a directional signal rather than a complete measure of adoption.
The next step is greater transparency around the underlying activity: absolute transaction counts, payment value, the networks carrying those payments and the share involving genuinely autonomous decisions rather than conventional automated transfers.
Those numbers will determine whether agentic payments are becoming a material financial rail or simply growing rapidly from an early base.
The compliance question may arrive sooner. If software increasingly gains authority to spend money independently, financial institutions will need systems capable of deciding whether a transaction is safe at roughly the same speed at which the agent decides to make it. The difficult part will be doing that without turning compliance itself into an autonomous black box.
The information presented in this article is intended for informational purposes only and should not be interpreted as financial, investment, or trading advice. Coinspress.com does not promote or advocate for any particular investment strategy, asset, or cryptocurrency project. Cryptocurrency markets are highly volatile and unpredictable – always perform your own research and seek guidance from a qualified financial professional before making any investment decisions.
Alexander Stefanov
Reporter at CoinsPress
Alex is Editor-in-Chief of Coinspress and co-founder of Millennial Media Group, with nearly a decade of experience covering financial markets - crypto first, then everything else. It started in 2016 with Bitcoin. Like most people at the time, he didn't fully understand it - so he kept digging. Blockchain, tokenomics, the projects, the cycles. That curiosity never stopped, and eventually pulled him into traditional markets too: equities, commodities, macro. Not because he left crypto behind, but because you can't properly understand one without the other. What drives him is straightforward: he wants to know why something is happening, not just that it's happening. Most market coverage stops at the headline - price up, price down, here's a chart. Alex finds that kind of reporting actively unhelpful. If you walk away from an article without understanding the mechanism behind the move, what did you actually learn? He holds a degree in Tourism from New Bulgarian University - not the most obvious path into financial markets, but markets have a way of pulling in people who are simply too curious to stay out. He has authored over 200 in-depth analyses and more than 10,000 articles across crypto and traditional finance. He still thinks every day in markets teaches him something new. That's probably why he hasn't stopped.
