AI Alone Won’t Fix Growth If Wealth Stagnates

Awash in wealth but starving for productivity — fine line, loud claim. Fortune’s headline captures a real anxiety: global assets look inflated while output growth limps, and artificial intelligence is cast as the cavalry. Frankly, the worry is right and the homework is thin.

Start with the part Fortune gets directionally correct. There is a gap between how rich the world looks on paper and how fast living standards are improving. Anyone staring at asset prices and tepid growth can feel that tension. The headline works as a provocation because it forces a basic question: if we’re so wealthy, why does growth feel so constrained and opportunity so uneven?

Then the article just…vibes. No definitions, no metrics, no sense of scale.

Counting money, not output

Fortune never pins down what “awash in wealth” actually means. Are we talking about paper valuations in public markets? Central bank balance sheets? Household net worth? Sovereign wealth funds? The phrase does a lot of work, but without a balance sheet attached to it, it’s just a mood.

Same problem on the other side of the ledger. “Starved for productivity” can mean GDP per hour worked, total factor productivity, or sector-by-sector output gaps. Each tells a different story. The article keeps it vague, which lets the thesis sound bold without ever becoming falsifiable. The math doesn’t lie — except when you never bother to write down the math.

When I was building models at Goldman, this is the kind of slide that would get torn apart in the first five minutes of an investment committee. Narratives without anchors are fun for speeches, useless for capital allocation. If you’re going to claim a distortion between wealth, growth, debt, and opportunity, you owe the reader a view of the accounts: where the wealth sits, where the debt sits, where the output stalls.

AI as deus ex machina?

To its credit, the article does more than hand-wring. It points to AI and says, explicitly: “We need AI to come through.” That framing captures a real mood in boardrooms right now. AI is being positioned as the productivity engine that will finally justify lofty valuations and strained public finances.

And yes, the potential channels are plausible. Automation of routine tasks, better decision support, new products, more scalable services — all of that could raise measured output per worker if implemented well. There’s a solid economic story here about general-purpose technologies that diffuse through the system and lift the frontier.

But the piece treats AI like a switch you flip, not a system shock you have to manage. Technology doesn’t raise productivity just by existing in a lab demo or a vendor deck. It needs retooled workflows, retrained labor, rewired IT stacks, and governance that lets firms take risk without blowing themselves up. None of that is incidental.

Right now you can see this disconnect in real companies. Look at big firms racing to bolt generative AI onto customer service or internal tooling — a lot of pilots, a lot of press releases, and not much validated productivity data yet. That’s not failure; that’s the lag between “cool capability” and “actual process redesign.” Fortune glides over those frictions. Legacy systems, misaligned incentives, CFOs who still measure success on blunt cost cuts instead of long-horizon productivity gains — these are exactly the frictions that will determine whether AI moves the needle or just fattens the software line item.

Let’s be real: “We need AI to come through” sounds less like a strategy and more like hoping the curve eventually bends in your favor.

Distributional first, efficiency later

The bigger blind spot is distribution. The headline tosses “opportunity” in with growth and debt, but the article never asks the obvious question: opportunity for whom.

Even if AI lifts aggregate productivity, that doesn’t tell you who captures the gains. Capital owners can take the margin expansion, incumbents can deepen their moats, and regions or workers on the wrong side of the skills divide can be left watching asset prices rise while their bargaining power erodes. You can have rising output and stagnant median wages. You can have a booming AI sector and hollowed-out local labor markets.

This is the part that can’t be outsourced to “the market will sort it out.” Markets are good at channeling capital to high-expected-return projects for those already positioned to benefit. They are not good at funding broad-based reskilling, cushioning displaced workers, or correcting for concentrated economic power. Ignoring those gaps doesn’t make them go away; it just means they show up later as political risk and social instability.

History rhymes here. Think back to earlier waves of automation: manufacturing automation did lift productivity and lower prices, but it also hit specific communities hard when policy and training didn’t keep up. The Fortune piece nods to imbalance without pressing on the mechanics of who wins, who loses, and what guardrails are needed.

The comforting counter-argument

One plausible rebuttal to all this is familiar: if there really is a wall of capital chasing returns and AI really does unlock higher productivity, markets will find a way. Investment will flow to the most efficient uses, firms will adopt AI where it boosts profits, and, over time, society reaps the benefits like it has with past technological shifts.

There is some truth in that story. Ignoring markets is as naïve as worshipping them.

But AI’s economics may be more skewed than earlier technologies. Network effects and data advantages can produce platform dominance quickly. A few firms can embed themselves deep in the stack and capture most of the upside, while everyone else rents access on their terms. Reskilling, meanwhile, shows all the hallmarks of a classic underprovided public good — diffuse benefits, concentrated costs, unclear short-term payoff. Assuming these issues “work themselves out” is how you end up with impressive national accounts and a furious electorate.

What would make the argument actually useful

If the author had done three things, the piece would go from provocative to genuinely informative.

First, define the terms with even a basic set of metrics: what kind of wealth is piling up, what kind of productivity is lagging, and in which major regions. Even rough contours would sharpen the thesis.

Second, map where AI is most likely to lift measured productivity — for example, which sectors have high information intensity, lots of repeatable processes, and enough margin to invest in retooling. Not every industry has the same upside or the same adoption friction.

Third, outline the policy levers and corporate choices that influence distribution: tax treatment of AI-driven profits, support for worker transitions, antitrust posture toward AI platforms, and standards for data and model governance. None of these are about blocking innovation; they’re about making sure the gains don’t bottleneck in a handful of balance sheets.

A sober risk checklist would also help: mismeasurement of productivity, short-term job displacement, concentration of economic power, regulatory drag. Those aren’t reasons to fear AI; they’re the constraints any serious strategy has to incorporate.

So yes, the Fortune headline taps into something real about inflated wealth and thin productivity — and it’s probably right that AI will sit at the center of how that tension resolves. But unless we start talking in ledgers, sectors, and specific trade-offs, “we need AI to come through” will stay what it currently is: a punchy line, not a plan.