Payments are becoming infrastructure. Accountability will define what comes next.

An AI agent buys an age restricted product using stablecoins. The payment settles. The signature is valid. The merchant gets paid. Is a valid payment enough, or should the merchant also verify who’s behind the agent? Curious to hear your thoughts.

AI agents can already pay. The next question is: Who’s behind the payment? That’s the layer we’re building.

HTTP has had a status code sitting unused since the 1990s. 402, "Payment Required." In 2026 it is processing more than 100 million transactions. Here is what actually happens inside it. 🧵

❌ MYTH: Adding identity breaks x402’s minimal design. ✅ FACT: x402 was built to be extended. A merchant can request the proof a transaction requires without changing the payment flow. x402 tells the agent how to pay. Concordium tells the seller who’s paying.

As AI agents become buyers, what’s the biggest unanswered question?

Interesting discussion yesterday with @Ledger around AI agents holding wallets, signing transactions, and moving real value. The conversation made one thing clear: capability is no longer the hard part. Agents can already act. The harder, more important problem is how you create lasting trust when those agents operate autonomously, especially once real money is involved. Most systems today can show that an agent exists and that a transaction was signed. Far fewer can answer the basic questions that matter in practice: Who authorized this agent? Who stands behind its actions? What limits does it operate under? @Concordium’s approach has been to treat these as protocol-level requirements rather than application-level afterthoughts. The Agent Registry links every registered agent to a verified human or business while keeping identity private through zero-knowledge proofs. Protocol-Level Locks add enforceable spending constraints that sit below the application, so even a compromised agent cannot exceed what it was authorized to move. That combination ... verifiable origin plus controlled settlement — is what turns an autonomous agent from an unknown actor into something counterparties can actually reason about. For those who missed the Space with @Ledger, the recording is here: nitter.net/i/spaces/1qKDzzgbXlBJV Worth a listen if you’re following how identity and accountability are being designed into the agent economy.

x402 is making agentic commerce possible. The next layer is giving businesses confidence in who stands behind the agents they’re transacting with. We explored why that matters: 🔗 concordium.com/article/x402-…

x402 answers one question. How does an AI agent pay? Concordium answers another. Who stands behind that payment? The agent economy will need both.

Live now with @Ledger. 🎙️ See you in the Space. 👇

Excited for the conversation with @Ledger in a 30 minutes. AI agents are already holding wallets and moving value. The real question is how we build lasting trust around them. @Concordium identity layer was made for exactly this moment.

One hour. See you at @Ledger. 🔐 👇

For years the internet asked one primary question about users: “Is this a real human?” The rise of autonomous agents has shifted the more important question. When software can spend money, sign permissions, and make decisions at machine speed, the critical issue becomes: who is accountable when something goes wrong? Most current systems can prove an agent exists and show what it has done on-chain. They cannot prove a verified human or business stands behind it. The agent exists. The wallet exists. The responsible party often does not. That gap is no longer theoretical. Real value is already moving through agents, and the first high-profile failures are appearing @Concordium treats identity as a condition of registration, not an optional add-on. Every agent in the Registry is linked to a verified principal. Privacy is preserved through zero-knowledge proofs. Spending can be constrained by protocol-level limits. The result is agents that retain autonomy while remaining answerable. As agentic systems take on greater financial authority, this distinction between “it exists” and “someone is accountable for it” will define which platforms are trusted.

Can I get a GM? 👀

16 years ago today, Satoshi wrote: “If you don’t believe me or don’t get it, I don’t have time to try to convince you, sorry.” The best infrastructure isn’t built for the next headline. It’s built for what comes next.

Tomorrow, our CTO, @PeterMarirosans, joins @Ledger to discuss trust in the agentic economy. Set your reminder 👇

Agentic Finance needs payment infrastructure. It also needs identity and accountability. x402 solves one layer well. Here’s why the next layer matters: 🔗 concordium.com/article/x402-…

Today, our CEO, @BBilowitzki, joins fellow industry leaders at Roundtable Wednesdays, hosted by the Digital Commonwealth Institute. Together, they’ll discuss “The Rise of the Agent: What it means for us mere mortals”, exploring the future of AI agents, trust, and digital infrastructure.

x402 answers one question. How does an AI agent pay? Concordium answers another. Who stands behind that payment? The agent economy will need both.