Ramaswamy's AI Wealth Pitch Misses Working Realities

He says workers can build wealth in the AI era. He doesn't say who gives them capital. He doesn't say how winners are chosen. That contrast matters.

Ramaswamy's Wall Street Journal piece sells a familiar equation: personal effort plus technology equals upward mobility. I don't disagree with the instinct. Workers should have paths to ownership; technological shifts really can open new rungs on the ladder. But the column skips past the hard questions about who owns what, who sets the rules, and who actually captures the gains. Those aren't philosophical gaps. They're balance-sheet gaps.

Let me start with something he gets right. Pushing workers to think like owners is better than telling them to brace for obsolescence. That mindset shift matters in any transition. But insisting on “ownership” without defining the asset, the terms, or the dilution schedule? That’s how you turn economics into vibes.

Who owns the machines?
Short paragraph. Big gap.

Ramaswamy gestures at workers “building wealth” but never defines what counts as wealth, and he mostly dodges the mechanics of ownership. Equity in a single firm is different from a claim on an ecosystem dominated by a few platforms that intermediate supply, demand, and data. Employees can be granted stock. They can get options. They can get bonus units with inspiring tickers. None of that changes the destination of most of the cash flows if AI scale mostly benefits the entities that already own the AI stacks, the infrastructure, and the customer relationships. So unless new equity models directly confront concentration — who holds the valuable IP, who controls access, who sets the take rate — “ownership” becomes an HR talking point, not a structural change to where the money goes. The math doesn't lie: if the capital base is fixed in a few hands, spreading around slivers of exposure does not suddenly democratize the surplus.

We've seen a version of this movie before. During the first dot-com wave, plenty of employees got stock options that looked like a path to prosperity; when markets turned, that “ownership” mostly translated into underwater grants and concentrated downside. The founders and early capital holders, who controlled the actual levers, walked away with the durable wealth. AI-era equity programs risk repeating that pattern if they don’t address who actually owns the underlying assets and data.

Training is necessary. It's not sufficient.
Short sentence. Then go long.

Right, reskilling is the part of his argument that sounds concrete. Train workers to operate AI tools, and they'll ride the wave instead of being crushed by it. Nice story. The problem is that training silently assumes that demand for that retrained labor will exist at wages that replace, or improve on, what was lost. Sometimes that happens. Often it doesn't. Markets reallocate, but markets also consolidate. When a single AI system can perform a task across many geographies and time zones, the bargaining power of any one worker drops. You can be the best AI prompt engineer in your county and still have minimal pricing power if the tool itself centralizes output, sets standards, and makes your individual contribution interchangeable.

My Goldman days taught me to watch margins and market structure, not just headcount. You can boost an individual’s productivity and still see their share of the surplus fall if the buyer side concentrates or if a few platforms intermediate most transactions. That’s not a failure of grit. That's what happens when the gains from efficiency show up as higher margins for intermediaries instead of higher pay for the people using the tools. So reskilling, without changes in market power or bargaining frameworks, is an incomplete strategy dressed up as a solution.

Policy and power are not optional
Short paragraph.

Ramaswamy leans hard on a narrative of individual agency. Fine. I'd prefer that to fatalism too. But the institutional levers that decide who actually captures gains — tax rules, antitrust enforcement, corporate governance norms, labor law, securities regulation — barely show up in his vision. Policies that nudge broader ownership could involve tax preferences for genuine employee ownership, different treatment for long-term worker-held equity, or real scrutiny of platforms that extract rents from both users and suppliers. None of that appears in his piece. That omission matters because, like it or not, rule design will decide whether AI widens the dispersion in outcomes or dampens it. Let's be frank: technology is neutral; incentives and enforcement are not.

If you want a corporate example of how design choices matter, look at how Costco treats its workforce versus a typical low-cost retailer. Same broad sector, similar tools, very different approaches to wages, benefits, and internal promotion. The result is a different distribution of who captures value along the chain. AI won’t automatically mimic one model or the other; the rules and incentives will tilt it.

Deep dive: what “building wealth” would actually require
Long sentence filled with specifics and analysis.

If “building wealth” is supposed to mean durable, meaningful claims on the economic surplus AI generates, then workers need at least three things: access to equity that isn’t constantly diluted or contingent on short-term performance, competitive markets that leave room for real bargaining, and public policy that limits value extraction by digital gatekeepers. Those are structural moves, not motivational memos. Employee ownership schemes that dump volatility onto workers without protections turn them into shock absorbers for business risk, not empowered co-owners. Entrepreneurship is great, but individual hustle will not offset platforms that control distribution, data, and discovery. Inside boardrooms, compensation committees and governance rules quietly determine who sees the upside when AI boosts productivity — senior executives, shareholders, or a broader base of employees. Pretending this is just about mindset misses the governance layer entirely.

Counter-argument, then rebuttal
Short. Then long.

You could say: workers have always adapted to new tools, and new entrepreneurs will emerge to exploit the gaps AI leaves. Some will. History does show that new technologies create entire categories that nobody predicted in advance. The honest counter is that the set of people able to pivot successfully is bounded by access to capital, networks, education, and time. Those inputs are not handed out evenly. Celebrating individual grit while ignoring those constraints turns a structural challenge into a morality play. And if the background rules favor scale and extraction, a boom in entrepreneurship can just mean more small firms competing on hostile terrain while dominant platforms levy their tolls on everyone in the ecosystem. So yes, agency matters — but it's constrained by ownership patterns, platform dynamics, and how much bargaining power is left at the edge.

A practical note from someone who traded markets
One-sentence paragraph.

From my years on the trading desk I can tell you: incentives change behavior faster than manifestos ever will.

Ramaswamy nudges the debate in a better direction by refusing to cast workers as spectators in the AI transition. That’s useful framing, but without concrete mechanisms on equity, power, and policy, “how workers can build wealth” reads less like a plan and more like an aspiration that assumes the hard parts away. Frankly, unless his optimism grows some teeth in the form of actual structures, AI will just amplify the ownership patterns we already have, not the ones his headline imagines.