AI wealth rhetoric doesn't empower ordinary workers

Ramaswamy starts from a bold proposition: workers can build wealth in the AI era. That claim is half a prediction and half a plea. It assumes the economic architecture around AI will tilt toward broad participation rather than concentrated ownership. That’s possible. It’s also a political and institutional choice, not a natural law.

Before getting to the problems, it’s worth saying why this optimism has a point. AI really does create new ways to add value. New tools reduce entry costs for entrepreneurship; automation can free experienced workers to do higher-value tasks; platforms can scale individual creators in ways that would have looked like science fiction back when dial‑up was still screaming in our living rooms.

I’ve covered enough cycles in Silicon Valley to know that when tooling improves, talented, well‑positioned people can turn those tools into businesses and equity. Think of how inexpensive cloud infrastructure turned a PowerPoint and a couple of engineers into something that could credibly chase incumbents. In science fiction terms, this isn’t a jump to a post‑scarcity Culture novel — it’s closer to early Charles Stross: intense opportunity in pockets, paired with serious dislocation elsewhere.

But agreeing with the headline doesn’t mean signing on to the whole argument. Ramaswamy’s optimism is an argument about what could happen. It’s much quieter about who actually gets to participate and on what terms.

Start with the ownership problem.

Wealth-building requires a stake in the upside. Salaries buy consumption; equity and capital gains build lasting wealth. If AI mostly increases productivity while ownership of the AI stack — models, infrastructure, and distribution — remains concentrated in a small circle of investors and executives, then the bulk of gains will accrue to that circle. We’ve seen this movie across the industrial age; the twist this time is the speed and reach of software platforms, which can concentrate returns far faster than railroads or steel mills ever did.

Ramaswamy sketches a hopeful pathway for workers. Sure, but he glides past the hard barriers between a good idea and durable equity ownership: access to capital, the know‑how to pitch investors, proximity to funding networks, and legal structures that make employee ownership meaningful rather than decorative. If policy and corporate practice stay as they are, AI wealth looks likely to repeat the pattern of earlier tech booms: huge, but tightly held.

History backs that concern. During the first wave of industrialization, factory workers got higher wages but almost no ownership; the real fortunes went to the mill owners and financiers. It took decades of legal and political fights to carve out basic labor protections, let alone profit-sharing arrangements. Expecting AI to naturally produce a broad investor class without similar institutional work is like expecting a Victorian textile baron to wake up one morning and invent the credit union out of sheer generosity.

This is why “just train the workforce” rings incomplete. Training matters — a lot. But training without a pathway to ownership is a half‑measure. A worker who learns to operate or augment AI systems still needs mechanisms to capture long-term value: employee equity programs that vest sensibly, co‑op or shared‑ownership structures, profit‑sharing, easier small‑business financing, or secondary markets where workers can actually sell their stakes.

You can see hints of these ideas in today’s companies. Employee‑ownership advocates often point to firms that have experimented with broad‑based stock grants or profit pools that meaningfully change workers’ net worth instead of tossing them token shares. These examples are still the exception. The question Ramaswamy doesn’t really confront is how you turn those isolated experiments into standard practice rather than PR gloss.

This is where policy design and corporate strategy intersect. Employers could redesign compensation to tilt a bit more toward equity and profit participation, not just bonuses. Cities and states could expand access to startup incubators, shared AI infrastructure, and low‑cost capital so that “AI founder” isn’t a role reserved for people who already know a partner at a coastal venture firm. Public policy could encourage more substantive employee ownership structures instead of box‑checking plans with negligible allocations. None of that is starry‑eyed optimism; it’s unglamorous institutional plumbing.

There’s a respectable counter‑argument here: let markets breathe and they’ll sort this out. New startups emerge, creating ownership opportunities; labor mobility lets skilled workers jump to places with better upside. That’s not wrong. Entrepreneurs do find unexpected niches, and talent has proved surprisingly willing to walk when equity looks better elsewhere.

But markets alone won’t erase geographic disparities or close the initial access gap. Capital networks still cluster. Legal and financial literacy are not evenly distributed. Without deliberate interventions, “self‑correction” often just means a new generation of winners who look a lot like the last one, right down to their ZIP codes. I’m not saying markets won’t help — they will — but they won’t, by default, democratize wealth.

Sector differences complicate this even further. Not every worker will be a founder, and not every industry easily translates labor into equity. Manufacturing and health care run on different capital cycles and regulatory constraints than software platforms. You can’t hand a nurse or a machinist the same startup playbook you’d give a cloud‑native engineer and expect identical results. If AI is woven into those sectors purely as a cost‑cutting tool rather than a way to share upside, you get efficiency without wealth-building.

Ramaswamy’s Wall Street Journal column does nudge the debate in a useful direction: it asks how we might design institutions so that AI amplifies workers’ economic stakes instead of erasing them. Where it feels thin is in grappling with which specific tools — equity plans, shared-ownership models, public AI infrastructure, local capital pools — actually scale across very different regions and industries.

Call it techno‑optimism with a civic add‑on: the tools are getting powerful enough that workers could build wealth from them, but unless someone rewires the ownership ladders, AI will mostly accelerate whoever already lives near the top.