Everyone seems to be excited about the agentic economy due to OpenClaw.

AI agents negotiating contracts. Optimizing procurement. Autonomously allocating capital. All commerce at machine speed.

The narrative is elegant. Intelligence becomes autonomous, commerce becomes frictionless, and GDP expands without humans in the loop.

But this framing misses the real story. The agentic economy is not primarily an economic revolution, however much the blockchain kids on the block want to say it is.

It is a sovereignty problem. A jurisdiction problem. A cybersecurity problem.

And in a multipolar world with no coherent rule-based order, autonomous AI agents do not neatly fit inside nation-state boundaries.

That’s where this gets interesting.

The optimistic vision assumes a few things:

But we don’t live in that world anymore. We’ve moved on, in the words of Canadian Prime Minister Mark Carney. We live in a world of:

So when we talk about AI agents transacting across borders, we should actually ask a simple question:

Who actually governs them?

If an AI agent based in Singapore negotiates a contract with a counterparty in Brazil, executes via infrastructure hosted in the US, clears through a European bank, and the transaction is exploited by a malicious actor operating elsewhere, then…

Who has jurisdiction? Who assigns liability? Who enforces?

Autonomous agents move at millisecond speed, but legal systems do not.

And that mismatch will define the next decade.

AI agents:

Jurisdiction is territorial. Law is political. Enforcement is fragmented.

In essence, autonomous agents compress time and geography. But nation-states do not.

This creates friction in the real world not because the agents themselves are intelligent, but because they are borderless.

In a unipolar world, this might have been manageable. There was implicit alignment across enforcement systems.

But in a a multipolar world, alignment is eroding.