Goldman's 'Agentic Commerce' Play Is a Wake-Up Call for Crypto Builders

On September 1, Goldman Sachs quietly reshuffled its European Conviction List. Out went Wise, the cross-border payments darling. In came three names that barely moved the ticker: Adyen, RWE, and Talanx. The market yawned. But buried in the note was a phrase that should have stopped every crypto builder mid-scroll: "agentic commerce." Goldman's analyst, Moawalla, didn't just talk about payment processing. He talked about AI agents transacting autonomously—and named Adyen, a centralized Dutch payment company, as the infrastructure best positioned to handle it. That is not a fintech story. That is a sovereignty story. For those who don't track sell-side artifacts, the Conviction List is Goldman's shortlist of buy-rated stocks where the firm is willing to put its own capital and reputation on the line. It is not a casual endorsement. Adyen has had a brutal year—down 28% year-to-date and roughly 37% below its 52-week high. Goldman sees 77% upside, to €1,780 per share. The market is pricing Adyen as a legacy payment processor. Goldman is pricing it as the settlement layer for machine-to-machine commerce. The other two additions are less interesting. RWE is a utility. Talanx is an insurer. They are ballast—defensive names for a list that needs stability. RWE benefits from European energy security, Talanx from rising insurance premiums. They are not technology plays. Their presence makes the list look like a macro hedge. But Adyen is the outlier—the only name with a pure technology thesis. And the removal of Wise is the real signal. Wise built a brilliant cross-border network, but its model is still a pipe. Adyen is a platform. In a world where AI agents need to transact with each other, a pipe is just a connection. A platform is a switchboard. Let's get technical, because this is where the market is missing the point. Adyen's "single platform" is not a marketing slogan. Most payment processors are built on layers of acquisitions—different systems for card routing, risk scoring, settlement, and reconciliation, glued together by APIs. Adyen built one codebase from day one. That codebase handles authorization, clearing, settlement, fraud detection, and merchant onboarding in a single runtime. It holds a European banking license from the Dutch central bank, which means it can connect directly to SEPA, card networks, and local clearing systems without relying on third-party sponsors. That is why Adyen can onboard a global merchant like Toast or Shopify in days, not months. It is also why its net revenue retention has historically stayed above 120%. Once a merchant is on the platform, the switching cost becomes enormous. The platform is the product. I've spent years auditing payment architectures, both centralized and distributed. The structural difference is real. Adyen's integrated stack gives it two things crypto protocols struggle to replicate: deterministic settlement timing and a unified risk engine. When an AI agent initiates a transaction, it doesn't care about "decentralization." It cares about finality, cost, and whether the transaction is allowed. Adyen's API answers all three in milliseconds. The AI partnerships with OpenAI, Google, and Microsoft are not endorsements; they are plumbing. Each of those companies needs a way for its agents to pay for compute, data, and third-party services without human approval. That requires an identity, a balance, and a settlement rule. Adyen provides the last two. This is what "agentic commerce" actually means: not a chatbot recommending a product, but an autonomous agent negotiating, purchasing, and reconciling a transaction. Crypto has been promising this exact future for a decade. We have the identity protocols. We have tokenized balances. We have smart contracts that can settle without a bank. But the market is betting on a centralized Dutch company instead. Why? Because the base layers are not ready. Bitcoin can settle large value with high security, but BRC-20 and Runes are like using a Rolls-Royce to haul cargo—it insults the car and doesn't carry much. For the long tail of micro-transactions AI agents will generate, Bitcoin is structurally wrong. And the Layer 2 landscape? ZK rollups are the most promising path, but the proving costs are absurdly high. Unless gas returns to bull-market levels, operators are bleeding money. Running an AI-agent payment network on today's ZK stack is like paying a six-figure salary for a cashier. None of this means crypto is dead. It means the market is paying for certainty, not ideology. Goldman can model Adyen's revenue because Adyen has a banking license and a balance sheet. Crypto's value proposition is different: it offers neutrality and programmability. But neutrality is hard to sell to a CFO, and programmability is hard to sell when the fees are unpredictable. The crypto ecosystem needs to stop talking about "revolution" and start talking about "settlement assurance." That is the language of the Conviction List. This is the uncomfortable truth. While crypto has been arguing about blockspace, Adyen quietly became the most efficient payment rail for the next generation of commerce. The data network effects are real: every transaction trains its risk models; every new merchant improves the value of the platform. Goldman is not betting on a payments company. It is betting on a data moat wrapped in a banking license. The regulatory dimension is equally important. Adyen's license isn't just a compliance checkbox; it is a distribution advantage. It lets Adyen move money across Europe without asking a correspondent bank for permission. That is the kind of structural integrity crypto claims to offer, but through code rather than paperwork. The irony is that Adyen achieved it with a bank charter. Here's the contrarian part: Goldman's conviction in Adyen might actually be the strongest argument for decentralized alternatives—not against them. Think about what we're about to build. If AI agents are going to transact at machine speed, we are creating the largest honeypot in financial history. A single centralized platform processing millions of autonomous transactions is a single point of failure. It is also a single point of regulatory capture. Adyen can change its fee schedule tomorrow. It can freeze an agent's balance at the request of a government. It can decide which AI actors are allowed to participate. That is not infrastructure; that is gatekeeping. The market doesn't care today because the market values efficiency over resilience. But the more dependent we become on centralized settlement for autonomous commerce, the more fragile the system becomes. We saw this in 2022 when centralized exchanges collapsed. We saw it again when payment processors froze accounts during politically contested events. The same cycle will repeat with AI agents. That is why the crypto opportunity is not "blockchain for payments." It is the neutral trust layer that Adyen cannot build: identity, attestation, reputation, and dispute resolution for machines. Trust is not given; it is compiled, line by line. The code is open, but the vision is ours to build. We do not follow trends; we architect ecosystems. If we spend the next cycle fighting over tokens instead of building the trust layer for agentic commerce, Goldman will be right to ignore us. From the ashes of FUD, we forge true adoption—but only if we stop pretending the answer is a coin. The Goldman note is not a crypto story. It is a signal. The market is willing to pay a huge premium for infrastructure that lets AI agents transact. It just doesn't believe crypto can deliver it. That should be a challenge to every builder. Stop trying to replace Visa. Stop trying to make Bitcoin carry cargo. Start building the protocols that let a machine prove it is trustworthy without asking a bank for permission. Volatility is the tax we pay for freedom. But if we don't start collecting, the agents will pay it to someone else.