There's a version of the next decade where GDP grows faster than it has in generations. Double-digit growth in some parts of the world. Entire industries restructured. A productivity explosion that makes the internet era look modest.

Jeremy Allaire — co-founder of Circle, the company behind USDC — laid this out on a recent episode of No Priors. He wasn't being promotional. He was being candid about something most people building in AI aren't saying out loud: the gains are coming, but they might not land where you think.

The specific phrase he used is worth sitting with: capital capturing capital.

When AI agents start running financial workflows, executing trades, managing supply chains, filing contracts, and negotiating deals — the productivity gains are real. But they accrue first, and possibly permanently, to whoever owns the infrastructure underneath them.

That's not a bug. That's how capital works. It compounds. It concentrates. And AI doesn't change that dynamic — it accelerates it.

Allaire identified three specific mechanisms worth understanding:

Compute and infrastructure ownership. The people who own the validators, the data centers, the GPU clusters — they capture returns at the base layer. You don't have to build a single product to profit enormously from an agentic economy. You just have to own the pipes the agents run through.

Financial rails and tokenized assets. As money moves on-chain — and it is moving on-chain faster than most people realize — the intermediaries and large token holders who control those rails collect value passively. Every transaction, every settlement, every cross-border payment that touches infrastructure they built or own produces a return. At scale, that's not a business model. That's the tax.

Automation replacing labor. This is the oldest mechanism in the story, but it's about to hit knowledge work in ways that manufacturing workers experienced decades ago. When agents handle the cognitive tasks that previously required humans, returns shift from wages to capital. Unless that shift is deliberately offset — through redistribution, through new ownership structures, through policy — the people doing the work won't capture the value of the productivity they helped create.

It doesn't matter what your politics are bro, this is capitalism in its purest form.

None of this is inevitable. Allaire's point — and what makes it worth taking seriously — isn't that abundance is bad. It's that abundance without institutional design defaults to concentration.

The legal and regulatory infrastructure we have was built for a different economy. Corporate structures assume human principals. Labor law assumes employment relationships. Tax policy assumes income flows through wages and business profits in roughly familiar patterns. None of those assumptions hold cleanly in an economy where AI agents are acting as economic actors, managing assets, and generating value that has no obvious human owner.

The lag matters. When technology moves faster than governance, the people with the most capital and the best lawyers define the defaults. By the time policy catches up, the defaults are entrenched. We've witnessed policymakers struggle with this these last 25 years.

What would need to change? Legal recognition for on-chain corporate forms — so there's accountability when agents act badly or extract value unfairly. Mechanisms for broader economic participation — fractional ownership, distributed token grants, platform designs that build in distributional intent rather than bolting it on later. New labor-welfare frameworks calibrated to displacement rather than unemployment — because what's coming isn't just job loss, it's a fundamental restructuring of how value creation maps to income.

And governance. Transparent, auditable governance of agentic systems — so humans can verify what agents are doing and enforce rights when something goes wrong — without those audit trails becoming surveillance infrastructure.

Here's the honest version of where this goes:

In the first scenario, the GDP growth is real, the gains are concentrated, policy lags, and by the time anyone gets serious about redistribution the winners are already too entrenched to move. High growth. High inequality. The abundance exists but it's not yours.

In the second scenario, the right people are paying attention right now — designing platforms with participatory governance, advocating for tax structures on capital rents from automation, building income stabilization mechanisms tied to displacement. The growth is still large. But more of it lands with more people.

In the third scenario — probably the most realistic — the response is fragmented. Some regions get it right. Others don't. The policy patchwork creates winners and losers at the regional level, not just the individual level. Where you live when this unfolds matters enormously.

That third scenario — fragmented response, regional winners and losers — is the one that should matter most to anyone in the Coachella Valley.

This is not a region with existing infrastructure dominance. There are no data centers to own, no GPU clusters, no token allocation tables that local stakeholders wrote. Which means the default outcome, if nobody is deliberate about it, is that the agentic economy happens to this place rather than for it. Productivity gains flow out. Value accrues elsewhere.

The region remains a place where people come to spend money that was made somewhere else. And that's just status quo with more horsepower.

The alternative requires showing up now — before the defaults are set — with clear institutional intent. What kinds of companies get recruited here. What the anchor institutions are doing with AI. Whether the people training on new tools are building ownership in the systems they're learning to use, or just becoming more efficient at jobs that are two years from automation.

Palm Springs had promising conversations — genuinely promising — and then executed exactly like you'd expect: same rooms, same people, same circling. Palm Desert put AI at the top of their agenda in January 2025. By April 2026 it has quietly disappeared from the list entirely.

I started these conversations with local leaders, institutions, and city councils almost three years ago. The good news is they're no longer confused about what ChatGPT is.

GDP is not a social contract. Growth doesn't distribute itself. An agentic economy that produces historic levels of abundance while concentrating it in the hands of infrastructure owners is not a success story — it's a warning.

Allaire's framing isn't pessimistic. It's an engineering prompt. The institutional design that determines who benefits from this economy hasn't been written yet. It's being written now, in product decisions and regulatory filings and token distribution structures and governance models that most people aren't paying attention to.

The window he is describing is not abstract. It's local. It's now. It won't stay open long. The people who shape what comes next aren't waiting for consensus. They're already building.

Sat Singh is a founder building AI apps and agentic tools in the Coachella Valley. He has been trying to explain this for three years. It's going great.

If this article resonates, pass it along to someone in the valley who'd benefit. This is a community project — it grows the same way communities do, one action at a time.