Three companies — $GOOGL, OpenAI, Circle — announced this week they’re building a shared infrastructure for agentic payments. At almost the same moment, analysts started warning that SpaceX and OpenAI IPOs could signal a market top. Most people read these as two separate stories. They’re the same one.
The IPO-as-top-signal argument is familiar. When the biggest private companies finally decide to go public, the reasoning goes, they’re timing the window. Liquidity is good, narratives are clean, institutional appetite is peaking. Historically, that’s often when sentiment has already priced in everything good that can be priced in. The concern isn’t wrong. It’s just incomplete.
What it misses is what the money is actually building.
The Google-OpenAI-Circle deal isn’t a headline about three logos aligning. It’s about who controls the rails when an AI agent needs to spend money on your behalf. Agentic payments — systems where software acts, authorizes, and transacts without a human confirming each step — require infrastructure that doesn’t look like a credit card checkout flow. They need programmable, auditable, composable money movement. And blockchains, it turns out, are unusually good at that specific thing.
Matthew Ball wrote in The Metaverse that blockchains “will not become the dominant means for storing data, computing, payments, LLCs” — but that they “will become key to many experiences, applications, and business models.” Jensen Huang called them “a fundamental new form of computing.” That framing is worth sitting with. Not dominant. Not the whole stack. Key. The way TCP/IP is key — you don’t think about it, but everything depends on it. Bitget’s onchain payments matrix this week, connecting banks, card networks, and blockchains into one routing layer, is a bet on exactly that version of the future. Boring infrastructure, extraordinary leverage.
Meanwhile, the question of what an AI agent business model even looks like is finally getting specific. Four models are emerging: open-source infrastructure plays, token distribution (essentially building your own money alongside your product), SaaS, and acquisition. The four-way split isn’t chaos — it’s the terrain being carved before anyone knows which ridge line will hold. Open-source infrastructure builds defensibility through ecosystem lock-in; tokens turn users into stakeholders; SaaS is the known quantity that enterprise procurement can approve; acquisition is what happens when a larger player decides it’s cheaper to buy than build. All four will survive. The question is which one wins at which layer of the stack.
Here’s the thing that isn’t getting enough attention: the agents themselves are becoming better at Java than Erlang. That sentence sounds technical but the implication is broad. In Vibe Coding, Gene Kim, Steve Yegge, and Dario Amodei describe a real enterprise considering migrating from Erlang — a language famous for resilient concurrent systems — to modern Java, specifically because AI coding tools perform better with Java. Not because Java became technically superior. Because the frontier models were trained on more Java than Erlang. Technical decisions are now being shaped by the distribution of the training corpus. The winning language in an AI-first development environment isn’t the most elegant one. It’s the one the model knows best.
Scale that observation out. The winning payment rail in an agentic economy isn’t necessarily the most elegant one. It’s the one the models are trained to use, the one the APIs are standardized on, the one the hackathons teach. OpenAI is throwing an autoresearch hackathon this Saturday with the exact tooling — agents, multi-model orchestration — that will define how the next generation of developers builds. These aren’t academic exercises. They’re curriculum-setting. Whoever writes the tutorials shapes the default stack.
So here’s the harder read on the IPO timing: SpaceX and OpenAI aren’t going public because the market is peaking. They’re going public because the infrastructure phase is ending and the application phase is beginning — and that transition is when you want permanent capital rather than venture money. The window is open not because sentiment is frothy but because the narrative is finally coherent. Agents can transact. The rails are being laid. The business models are resolving. An investor buying into that story now isn’t buying the top of a cycle. They’re buying the beginning of a new cost structure for every industry that moves money.
Whether the valuation is right is a separate question. Whether the infrastructure earns its price depends on what gets built on top of it. And what gets built on top of it is being decided this Saturday, at a hackathon, by people who will never make a market-top list.
The top isn’t where they’re building. The top is what gets built.
Leave a Reply