Questioning the AI Wealth Pitch for Workers

Vivek Ramaswamy says workers can build wealth in the AI era. Here's the thing: that's not wrong — it's incomplete.

There’s a version of his Wall Street Journal argument that I genuinely like: AI doesn’t have to be a one-way conveyor belt from workers to shareholders. New tools can raise productivity, and when productivity rises, societies can choose to share the gains. That’s the part techno-optimists get right. I’ve watched enough platform shifts — the mobile boom, the SaaS gold rush, the crypto winter-that-wouldn’t-end — to know that technology really does mint new fortunes.

But optimism without mechanics is just vibes.

Ramaswamy’s core claim is that workers can ride the AI wave to build wealth. Sure. The question is: who gets the ladder, who gets the wave, and who’s standing on the beach watching someone else’s yacht.

Who actually captures the upside?

Most AI value is going to cling to capital and control, not just labor. AI systems are classic “scale without bodies” machines: once the model is trained and deployed, serving an extra million users doesn’t require hiring a million people. The upside accrues to whoever owns the model, the data, and the distribution.

That means boardrooms, investor cap tables, and founders with access to compute and talent. You can argue that entrepreneurship is the way out — quit your job, start an AI-powered niche service, go get your slice. Look, I’m very pro-startup; I’ve watched tiny teams turn slide decks into serious companies. But “start a company” is not a realistic prescription if you’re juggling two jobs, caring for kids, and living in a town where the best “ecosystem” is a strip mall and a shuttered Sears.

There’s an unspoken assumption in Ramaswamy’s column that opportunity is evenly distributed as long as markets are left to do their thing. I’ll be honest: that’s not economics; that’s political design. Worker access depends on stuff his piece barely touches — portable financing so you can leave a job without losing your safety net, local ecosystems that support tiny firms instead of just trophy startups, and retraining programs linked to actual employers instead of one-size-fits-none certificates.

Without those, “workers can build wealth” starts to sound like “lottery winners can get rich.”

Ownership isn’t just a stock ticker

If workers are going to share AI’s upside, ownership structures have to change in ways that feel real at ground level. The U.S. already has employee stock-option plans and ESOPs. On paper, they’re a bridge from worker to owner. In practice, they often concentrate most of the meaningful equity in executive suites and early hires with Ivy League LinkedIn pages.

For non-technical workers, equity has to do more than decorate a paycheck stub.

Equity that vests over time is nice, but if you’re living paycheck to paycheck, future upside doesn’t fix this month’s rent. We need models that translate some of that future gain into present stability — think equity-like compensation that can actually underwrite health care, housing, or childcare, not just a line in an HR portal.

The predictable pushback: if you make broad-based ownership too aggressive or too mandatory, you scare off investors and slow innovation. There’s a kernel of truth there. Investors do chase higher-return environments, and founders don’t wake up excited to navigate a 400-page ownership mandate.

But the world isn’t binary between hard caps and free-for-all.

We’ve seen workable middle ground in other contexts: profit-sharing tied to productivity gains, tax incentives for companies that spread equity broadly instead of hoarding it at the top, or co-op-style structures in specific sectors. Some European manufacturers have experimented with works councils and shared governance; they didn’t suddenly forget how to make cars. In the U.S., companies like Publix have quietly built worker-ownership models that, while not perfect, show you can scale and still cut employees into the upside in a systematic way.

The point isn’t that every AI firm must become a co-op. It’s that “ownership” cannot be limited to people who already own plenty.

Retraining: necessary, but hilariously oversold

Ramaswamy’s optimism about retraining and new skills isn’t unique; it’s practically mandatory in any tech op-ed. Workers, we’re told, just need to “upskill” into roles that AI can’t easily automate.

That sounds great from a conference stage.

On the ground, reskilling programs often fail for two very boring reasons: they’re misaligned with local labor demand, and they pretend workers have unlimited time and support. If your training course is 20 miles away, you don’t have a car, and you’re covering childcare on your own, the “opportunity” may as well be on Mars.

The stuff that tends to work looks less glamorous and more contractual: apprenticeships where employers commit to hiring trainees on the other side, employer-sponsored training that’s mapped to real job ladders, and public funding that follows actual placement rates instead of brochure copy. Put some skin in the game for both training providers and employers, and suddenly “retraining” stops being a punchline.

Geography is destiny — unless you rewrite the map

Here’s the part that’s usually missing in these wealth-building arguments: place. If AI-related growth keeps clustering in the same familiar ZIP codes — the San Franciscos, Bostons, and Austins of the world — then a huge chunk of the country is left reading op-eds about prosperity it will never see.

Urban tech hubs generate intangible assets: investor networks, mentors, shared office space, the casual collisions that turn ideas into funded companies. You can’t download any of that over a 5G connection.

Policy that intentionally seeds regional innovation centers — and ties support to concrete hiring and training commitments — matters more for worker wealth than another TED-ready speech about disruption. Otherwise, we’ll get what we always get: a few cities flush with AI windfalls and a lot of places stuck managing the downside.

A few levers that actually point at workers

If we’re serious about workers sharing AI gains, the tools exist; we just rarely use them at scale:

• Make profit-sharing automatic when firms deploy AI that clearly displaces labor — not as charity, but as a negotiated cut of productivity gains.
• Tie public training dollars to verifiable placement outcomes; fund apprenticeships with private-sector matching instead of just subsidizing classrooms.
• Encourage portable benefits and small-business-friendly financing so workers who do take the risk of starting something can survive the early, ugly months.

Yeah, no, none of this is tidy. But saying “workers can build wealth” without grappling with governance and access is like telling sailors to learn to swim after you’ve already blown a hole in the hull.

Science fiction has been running this simulation for decades. In Heinlein’s lunar colonies, it wasn’t technology that decided who prospered; it was who controlled the banks, the air, the rocks. AI is less dramatic than a moon revolt, but the principle holds: tools concentrate power unless institutions drag that power back into some kind of balance.

If Ramaswamy is right that AI can make workers wealthier, the test won’t be the technology — it’ll be whether we’re willing to redesign ownership, training, and geography so those workers aren’t just spectators in someone else’s AI success story.