The U.S. is moving steadily towards its goal of becoming the world’s “crypto capital” and Crypto x AI discoverability and interoperability are becoming the table stakes to play in this game.
Crypto stakeholders are embracing the much-needed yet newfound regulatory clarity with the passing of the GENIUS and CLARITY acts, and more recently, the CFTC’s FBOT advisory, allowing off-shore entities like exchanges to serve U.S. users.
Web3 funding crossed the $9.6 billion mark in Q2 2025 making it the second-highest quarter on record, with the number of deals disclosed falling to 306.
Even with the reduced number of deals, the funding numbers signal a greater conviction in emerging projects the rapidly emerging demand for AI-led crypto and digital assets infrastructure is evident.
Alongside leaders in cryptocurrency, mining and validation, and treasury management, verticals like “compute networks” – decentralized platforms that distribute computing tasks, such as training AI models across numerous independent nodes (computers) instead of relying on a single central server – have attracted a major share of such fundings.
Institutional use cases using Agentic AI are on the rise as they improve decentralized trustless settlement layers and data infrastructure. Asset Manager VanEck’s projections of a million AI agents using crypto rails and blockchains by the end of 2025 seems plausible, there were 10,000 Web3 agents as of December 2024.
The prospect of new retail and institutional users entering the Crypto x AI market over the next 2–3 years has sparked a renewed urgency and dedicated efforts to build better, more interoperable UXs.
Bottom-up, user-centric architecture is becoming the norm, especially in Crypto x AI, as projects try to balance core principles — decentralization and user autonomy and control — with security, scalability, ease-of-use, and more feature choices.
Accelerating Coordination Is The Endgame
Building the toolkit for unstoppable coordination has always been on crypto’s long-term agenda. So far, most, including crypto builders, haven’t had coordination on their radar as a priority. It’s one of the reasons for the often excessive short-termism and extractive intent in some pockets of the crypto industry.
Recent developments at the intersection of Crypto x AI have made the objective of coordination accelerationism explicit and thus hard to overlook. More so because leading organizations — both traditional and crypto-native — are building practical, hands-on solutions to this end, now that the underlying frameworks are mostly ready and in place.
The Ethereum Foundation, MetaMask, and Google are collaborating on ERC-8004, aiming to catalyze onchain agent discovery and “trust through reputation and validation.” It’s a critical step towards “trustless” human-to-agent and agent-to-agent communications onchain, signalling ongoing preparations for a collaborative “agentic web”, which is largely imminent.
Michael Sena, co-Founder of Recall Labs, says, “Coordination accelerationism isn’t new for those in the crypto trenches, but its timeline and overall feasibility has changed a lot, better to say completely, due to AI’s exponential growth in the past year or so.
“Earlier, we used to think of coordination mostly in terms of human actors. AI systems weren’t a part of the equation as much; it was too early to consider them as meaningful actors in global financial or technological ecosystems. That’s no longer the case, given how AI models and agents have evolved and improved recently. Coordination has a much wider scope in this agentic realm: human-to-human, human-to-AI, AI-to-human, AI-to-business, and AI-to-AI.”
AI has also matured in the past year and institutions like the IRS are integrating AI into their daily workflows.
With Beijing aggressively pushing for nationwide AI integration, strengthening support for crypto-native, AI-powered coordination systems will be critical for the U.S. to maintain and grow its leadership in digital innovation, and globally compete over the next decade.
The Non-Negotiables: Discoverability, Trust And Interoperability
To coordinate, collaborate, and produce value at scale, counterparties first need to discover and trust each other through fair, tamper-proof mechanisms. They then need an interoperable mechanism for securely sharing resources and working together.
Emphasizing PageRank’s role in transforming search in the legacy web, Andrew Hill, Recall’s ceo and co-founder, says, “Google search is useful because its algorithm organizes the internet by indexing and ranking high-quality websites, and serving them up based on relevance to the user’s search.
“Before this, however, discovering trusted websites was as challenging as finding a needle in a haystack. And as a website builder, you either hoped for distribution or spent millions to bootstrap and maintain your own channels.”
The current state of the AI ecosystem is similar to the pre-Google era – Google arguably “won” the search war by making search relevant, the unintentional consequences of which upended the media advertising market.
Hill adds, “Social media, word-of-mouth, newsletters, and agent launchpads are the most common ways of discovering AI tools today. But these methods are untrusted, unscalable, and prone to manipulation.
“They don’t have embedded trust, reputation, or distribution mechanisms designed specifically for the AI-native web. They can’t handle the millions or billions of agents and AI-powered systems that will inevitably come online in the next few years, underpinning the trillion-dollar AI economy.”
Realizing the need for discoverability, innovators like Hill are building crypto-native primitives and platforms where humans and agents can discover each other through trusted channels, based by onchain reputation models and mechanisms.
The “Trustless Agents” ERC (#8004) is an example of ongoing efforts to boost agent discovery, however, simply discovering trusted agents isn’t enough to fully realize the functional and economic potential of the agentic web.
Interoperability is another critical factor, insofar as it lets agents share resources, while users can seamlessly move between systems to best fulfill their needs and demands.
The Ethereum Foundation recognizes this need, and is thus doubling down on interoperability to significantly improve its UX over the next 6 to 12 months. This arguably favors Ethereum’s attempt to become the go-to settlement layer for the agentic web, particularly significant with institutional investors backing the network heavily. In Q2 2025 alone, for instance, they poured over $2.4 billion into Ethereum-backed instruments.
Andrii Miloshin, cto of Tairon, says, “Interoperability is absolutely mission-critical for AI agents and the agentic web. Otherwise, we end up with a bunch of isolated intelligences, with limited resources and thus sub-par capabilities. More importantly, when agents take actions based on siloed or fragmented data, they are more likely to hallucinate, which can be a disaster when there’s millions and billions of dollars at stake.”
Miloshin believes building integrated communication channels that let AI agents access onchain and offchain data is one way to ensure interoperability and adds “Emerging, chain-agnostic communication layers are connecting AI to crypto, like oracles connected offchain data to smart contracts. With these, both data and tokenized value can move freely across agentic systems, taking us closer to the infinite garden vision.”
James Loperfido, principal of autonomist.ia and Web3 AI specialist says, “Every AI agent needs a wallet and every wallet needs an AI agent. Permissioned and user controlled agentic systems that have been battle hardened alongside interoperable smart contracts and oracles can create the level of abstraction required for mainstream users to be impressed by the product experience, at the non-custodial wallet level.”
Securing Data for Truthful Agentic Systems
A truthful agentic system is an artificial intelligence (AI) framework designed to act autonomously toward a specific goal while maintaining high standards of honesty, transparency, and reliability.
With a favorable regulatory landscape and existing institutional adoption momentum, the Crypto x AI vertical will likely produce lots of tools, products, and services in the coming years. More capital will flow into the space, and it’ll create even more economic value, which also means higher stakes and shrinking margin for error.
Agentic systems need to be substantially secure and performing before they can be actually integrated into enterprise-grade systems that hold or manage hundreds of billions of dollars in company and user funds. And unless AI innovators ensure optimum reliability and safety, no amount of distribution or interoperability would make much sense.
Adds Loperfido, “Cryptography, security standards, and proof of “personhood” advancements will go a long way to force unnecessary data off chain for rapid compute amidst secure enclaves and existing data stores. AI and Agents can become an accelerator of progress both from a developer standpoint and the idea that they create new use cases with programmable money.”
There is also a growing consensus among industry specialists that quality data is one of the key bottlenecks to enterprises seeing adequate returns on their investments into AI. From fragmented records to corruptible datasets and the general noise across social media, agentic systems are facing a trust and truth crisis at the most fundamental level.
AI systems are only as fair, reliable, and unbiased as the data they interact with, for training or otherwise. For enterprises and high-stake users to safely leverage agentic systems, it’s necessary that these systems are trained on and have access to clean, curated, and verifiable data. They also need transparent systems AI agents or models clearly cite their sources, while facilitating auditability.
While several crypto-native projects are innovating decentralized data creation, ownership, and incentive mechanisms to ground AI systems in truth, there’s still a lot to do in this direction. And given how complex truth and trust themselves are, building truthful AI systems is easier said than done.
Projects like Phoenix are making AI-powered agents and conversational search engines more reliable and transparent, where they “show their work” in a way they users can easily cross-check. Besides access to real-time data, they’re also enabling agent-to-agent interoperability, so that next-gen AI systems can learn from each other, rather than spitting out misleading or false information.
Overall, with strong demand signals across verticals, and a $52 billion market opportunity by 2030, there’s no shortage of incentives to go all in on building performance, secure, and unbiased agentic systems on crypto rails.
The work has just begun, so now it remains to be seen how rapidly Crypto x AI discoverability and interoperability evolves and what value it adds to the world of the mass market user of crypto.