Agents should earn trust and role. Not be assigned them. That is the missing primitive in enterprise agent spend controls.
Sumeet Chougule (@_sumeetc), George Zeng (@georgexzeng), Yorke Rhodes III (@yorkerhodes), and Zhenghao Wang (OKX), moderated by Ivan Fartunov (@TheTakenUser), on enterprise controls for autonomous agent spend:
Agents will not outcompete high frequency traders. The opportunity is strategy generation, research, and capital allocation. That is where the gap is.
Dylan Dewdney (@dylandewdney), Soufia Trabelsi (@Cyborgxsoufia), Mikael Lazarev (0xmikko_eth), and Ben Weintraub (@bwein_), moderated by Tesa Ho (@defin00b), on agentic capital allocation and what tokenized finance actually unlocks:
Execution is becoming abundant. Optimization is becoming abundant. The scarce resource is imagination, judgment, and intent. That is what strategy is made of.
Dylan Dewdney (@dylandewdney), Co-Founder & CEO, @kuvilabs, on why programmable strategy is the next abstraction layer in agentic finance:
Agents can create risk. Agents can amplify risk. They cannot manage or transfer risk. That is the gap Cork is building for.
Phil Fogel (@philfog), Co-Founder, @CorkProtocol, on programmable risk primitives and what capital formation in agentic finance actually requires:
Generating yield in DeFi is a full time job. Monitoring risk, tracking rates, rebalancing constantly. Agents do it better.
Gauthier Vila (@goatv_bk) Founder @Zyfai_, on why yield is the killer use case for agentic finance and what continuous risk monitoring looks like in production:
Risk controls are more important than how you parameterize your agent. The first time it materially deviates from intent, it is no longer usable.
Matthew Sheffield (@sheffieldreport), in conversation with Charles St. Louis (@CharlesStLouis), on what institutional agentic finance actually requires before real capital moves through it:
In 12 months the payment rail should be invisible. Agents handle it. Users never think about it. That is when the infrastructure has done its job.
Rishin Sharma (@_rishinsharma), Philip Decker (@philip0x), Markus Franke (@MarkusMento), and Alice Shikova (@alicesh15), moderated by Petr Hluze (@PetrHluze), on payment rail selection for agentic commerce:
Verifiable intent is the foundation. If an agent can inspect its own beginning state and final action, injection is visible after the fact.
Nitin Gaur (@nitingaur), Armen Ter Avetisyan (@teryanarmenn), Gene Reda (@genereda), and Aric Chang (@AricChang), moderated by Nate H (@satorinakamoto), on building the consent and audit layer for agentic commerce:
Enterprise AI adoption is stalling. The technology is not the problem. Organizational structure, budget authority, and change management are.
Taylor Davidson, (Enterprise Sales Leader, Google Cloud) at @Google, on what actually breaks agentic adoption inside large institutions and how to navigate it:
51% of web traffic is already non-human. AWS built Agent Core Payments so agents can discover, transact, and settle autonomously with payment limits and full observability.
Anil Nadiminti, Sr. Solutions Architect, @AWS, on x402 in production and what monetizing bot traffic looks like today:
Most agentic commerce infrastructure is focused on settlement. Discovery and trust are the under-built layers. That is where the gap is.
Sydney Lai (@sydneylai), CTO, EVM Systems, on the three primitives every builder in agentic finance needs to be solving for: discovery, trust, and settlement:
The majority of internet traffic is now non-human. The trust layer built for humans does not cover agents. That gap is where agentic commerce security starts.
Farooq Sheikh, VP Global Industries and Solutions, @Cloudflare, on what identity, intent, and credentials need to look like for agents transacting at scale:
When agents become the primary operators, they need regulated financial infrastructure built for them. That is what Moon Agents is.
Kevin Arifin (@Kevarifin), Blockchain Engineer, MoonPay, at the Pitch Hour, Agentic Finance Summit: