A common theme ran through multiple sessions at WebX 2026 in Tokyo: AI and on-chain finance are moving toward each other, and public blockchains are being framed less as ledgers for assets and more as infrastructure designed to constrain AI. That idea links two areas that have often been discussed separately.
Smart contracts as guardrails and a kill switch for AI agents
The core argument presented at the event was straightforward: if AI starts to control capital and even takes on part of the decision-making process, society will need a neutral set of rules that can actually be enforced. Participants said smart contracts on blockchains could fill that role.
Because smart contracts are auditable, tamper-resistant, and automatically executable, they were described as possible guardrails for AI agents. Speakers also pointed to them as a potential emergency brake, or “kill switch,” if something goes wrong. In that framing, the value of a blockchain is no longer limited to recording assets. It also lies in setting rules for automated AI behavior.
AI agents are moving from conversation to execution
A second theme focused on how AI agents are beginning to move beyond chat and into actual execution. One speaker said on-chain analytics and trading platforms are shifting from tool-based products toward agent-led systems, with conversational interfaces replacing traditional terminals.
Another point raised during the event was that older backtesting methods may need to be rewritten in full for AI agents that can reason on their own. Fast-growing tokenized assets were discussed in the same context. The article pointed to Ondo’s tokenized stock platform, which has surpassed $1 billion in TVL, as part of an early model in which AI agents may directly use on-chain financial assets.
Venture capital is leaning into AI-native finance
The same line of thinking is also showing up in venture capital. According to the report, institutions at the event including a16z argued that the long-term case for tokenization is no longer just about lowering costs. Their focus is shifting toward “AI-native finance.”
Under that thesis, once financial assets are on-chain and turned into programmable building blocks, AI agents could combine, trade, and settle them much like calling an API. That is also why “neutrality” emerged as a key word across the event. Participants said a platform controlled by a single operator would be hard for both institutions and AI systems to trust, while a neutral and open settlement layer would be better suited to support that vision.
Production use cases are still early
The report added one clear caveat: most of these ideas remain forward-looking judgments from people at the event, and real production cases are still at a very early stage.