In February 2026, Citrini Research published “The 2028 Global Intelligence Crisis,” a fictional macro memo written from June 2028. It models a scenario where AI delivers on every productivity promise and the result is catastrophic: a negative feedback loop where companies replace white-collar workers, funnel savings into more AI, and repeat. The S&P peaks at 8000, then falls 38%. The consumer economy collapses because machines spend zero on discretionary goods. The $13 trillion mortgage market cracks as prime borrowers lose the income their loans were underwritten against.
Two months later, OpenAI published Industrial Policy for the Intelligence Age, a 13-page policy paper proposing an ambitious new social contract for the AI transition. Buried between safer proposals about workforce retraining and grid expansion is a radical idea: a Public Wealth Fund. AI companies and the government would seed a national investment vehicle, it would hold diversified assets capturing AI-driven growth, and returns would be distributed directly to every citizen. A sovereign wealth fund meets universal basic income, funded by the companies building the technology that’s displacing workers in the first place.
Read side by side, these two documents tell a striking story. Nearly every mechanism Citrini dramatizes as a crisis is something OpenAI quietly acknowledges as a real risk. And the wealth fund OpenAI proposes is their answer to the exact problem Citrini says will be unsolvable in time.
What OpenAI is actually proposing
The paper is deliberately vague on mechanics, which is either intellectual honesty or strategic ambiguity depending on your read. What’s clear is the structure: AI companies contribute capital or equity into a fund. The fund invests broadly, not just in AI stocks but across the economy. Returns flow to citizens as direct payments, giving people who aren’t invested in financial markets a share of the upside.
The logic tracks. If superintelligence delivers even a fraction of the productivity gains OpenAI predicts, corporate profits will surge while labor’s share of income continues to shrink. A wealth fund is a mechanism to redirect some of that surplus back to the public without relying solely on tax policy, which moves slowly and gets captured by lobbying.
Norway’s Government Pension Fund is the obvious precedent. Alaska’s Permanent Fund is another. Both work. But neither was designed to redistribute gains from a technology that could restructure the entire economy in a decade.
But look at who benefits from the proposal itself. The paper arrived in April 2026 amid mounting external commentary calling it a “policymercial,” timed as OpenAI navigates IPO pressures, intensifying competition, and growing public backlash against AI-driven displacement. The wealth fund proposal is generous in spirit, but the mechanics tilt favorably toward frontier labs. OpenAI and its peers would seed the fund, likely with equity or equity-linked contributions. That equity would then be held by a government-backed vehicle with a mandate to hold long-term, creating a structurally captive buyer for their stock. The fund’s investment mandate, to capture “AI-driven growth,” would naturally overweight the very companies contributing to it.
Meanwhile, the fiscal burden of the fund’s parallel requirements, the supply-side investments, the adaptive safety nets, the grid expansion, falls on taxpayers and government budgets, not on the labs. The companies contribute equity (which costs them nothing in cash flow) and in return get a government-endorsed buyer of their shares, a public narrative of shared prosperity, and a regulatory environment shaped by their own policy proposals. This doesn’t make the idea bad. But it does mean the altruism and the self-interest are hard to separate, and any serious implementation would need to account for the structural advantages the proposal creates for its authors.
The bear case already has a script
Citrini’s scenario is deliberately extreme. They call it a thought exercise, not a prediction. But read OpenAI’s policy paper against it, because the overlap is uncomfortable.
Citrini’s central concept is “Ghost GDP,” output that shows up in national accounts but never circulates through the real economy because it routes through GPU clusters instead of paychecks. OpenAI’s paper uses softer language, but describes the same dynamic: corporate profits and capital gains expanding while “reliance on labor income and payroll taxes” declines. Their proposal to “modernize the tax base” is a direct response to the fiscal collapse Citrini’s scenario models, where federal receipts run 12% below projections because the government taxes human labor and human labor is shrinking.
The displacement spiral Citrini describes, where AI improves, companies cut headcount, savings fund more AI, AI improves further, is exactly what OpenAI’s “adaptive safety nets” proposal is designed to interrupt. The fact that OpenAI proposes automatic triggers tied to displacement metrics tells you they view the spiral as plausible enough to require a pre-programmed fiscal response. You don’t build circuit breakers for risks you think are theoretical.
The wealth fund itself maps directly to a proposal in Citrini’s timeline. In their 2028 scenario, Congress debates a “Shared AI Prosperity Act” that would establish a public claim on the returns of intelligence infrastructure, with dividends funding household transfers. OpenAI is essentially publishing that proposal two years earlier, from the perspective of the company that would be contributing to it. The difference is that in Citrini’s telling, the proposal arrives too late, mired in partisan gridlock while the feedback loops accelerate.
This is where the comparison gets uncomfortable. OpenAI’s paper reads like a company that’s studied the Citrini scenario (or something like it) and is trying to pre-empt it. Their 32-hour workweek pilots assume labor demand will fall so substantially that the workweek itself needs to shrink. Their “pathways into human-centered work” section, redirecting displaced workers into caregiving and community services, is a dignified policy version of the same downshift Citrini dramatizes as a crisis, where the ex-Salesforce PM ends up driving for Uber at a quarter of her prior salary.
Then there’s this line from the OpenAI paper: “If AI winds up controlled by, and benefiting only a few, while most people lack agency and access to AI-driven opportunity, we will have failed to deliver on its promise.” That’s not a hypothetical. That’s the base case they’re trying to prevent, and it’s the base case Citrini models as already happening.
So would the fund actually cause inflation?
This is where the two frameworks diverge, and where the real analysis begins. The inflation question isn’t simple because the two sources point toward different macro regimes.
Any policy that puts new money into consumers’ hands without a corresponding increase in the supply of goods and services is inflationary. That’s not ideology. It’s accounting. The question with a Public Wealth Fund isn’t whether it could cause inflation. It’s whether the productivity gains from AI would grow the supply side fast enough to absorb the demand.
The optimistic case: AI dramatically reduces the cost of producing goods and services. Healthcare gets cheaper. Energy gets cheaper. Food production gets more efficient. Housing construction gets faster. In that world, distributing fund returns to citizens doesn’t cause inflation because the economy is producing more output per unit of input. Prices fall or stay flat even as people have more money to spend.
The pessimistic case: AI productivity gains concentrate in digital and knowledge sectors while physical-world bottlenecks persist. You can use AI to design a better building, but permitting, materials, and labor still constrain how fast it gets built. Healthcare might benefit from AI diagnostics, but hospital beds, nurses, and drug manufacturing don’t scale with software economics. In this scenario, fund distributions chase the same constrained supply of real-world goods, and prices rise.
And then there’s the Citrini case, which is neither inflation nor simple deflation but something weirder. Consumer spending collapses as displaced workers cut back. Credit tightens as mortgage assumptions break down. The real economy deflates. But the government, desperate to stabilize, responds with massive fiscal transfers funded by deficits. You get falling real incomes and rising financial asset prices simultaneously. The fund would be distributing returns into an economy that’s deflating in the places people actually spend money and inflating in the places they don’t.
The realistic outcome is probably somewhere in this triangle, and the timing matters enormously.
Timing is everything
Inflation isn’t just about totals. It’s about sequences. If a Public Wealth Fund starts distributing returns before AI has materially expanded productive capacity, you get demand-pull inflation in exactly the sectors where people spend most of their income: housing, healthcare, food, and energy.
This is a lesson from pandemic-era stimulus. Direct payments worked as crisis relief, but they hit an economy with constrained supply chains. The result was the sharpest inflation spike in four decades. The mechanisms differ (pandemic stimulus hit a supply-shocked economy, a wealth fund would hit a restructuring one), but the principle holds: timing demand-side transfers to supply-side conditions is the difference between stimulus and inflation.
The counterargument is that fund returns would likely start small and scale gradually as the fund’s assets appreciate. This isn’t helicopter money. It’s investment returns distributed over time. If the fund is structured conservatively, early distributions might be modest enough to be absorbed without price pressure.
But political incentives push in the other direction. Once a fund exists and citizens expect distributions, there’s enormous pressure to increase payouts, accelerate timelines, and expand eligibility. Alaska’s Permanent Fund has faced exactly this dynamic for decades.
In the Citrini timeline, this pressure would be even more acute. If displacement is visible and accelerating, the political demand to pay out early and pay out more would be overwhelming, regardless of what the supply side is doing. The fund becomes a crisis-response tool operating under crisis-response pressure, which is exactly the wrong condition for a long-duration investment vehicle.
Where inflation would hit hardest
Not all prices respond the same way to increased demand. Think of the economy in three buckets.
Digitally scalable stuff (software, AI tools, streaming, information services) has near-zero marginal cost. More demand doesn’t meaningfully increase prices. Fund distributions spent here are basically non-inflationary.
Physically constrained essentials (housing, healthcare, childcare, education) are a different story. These are bottlenecked by regulation, labor shortages, physical infrastructure, and long build cycles. More money chasing these goods pushes prices up. This is where inflation would concentrate, and it’s where lower-income households spend the largest share of their budgets.
Commodities and manufactured goods (food, energy, consumer products) sit somewhere in between. AI could reduce production costs over time, but in the near term, these prices are driven by global supply chains, weather, geopolitics, and energy markets, not domestic demand policy.
The irony is that the people a Public Wealth Fund is designed to help, those not currently participating in financial markets, spend disproportionately on the category most vulnerable to inflation. Without parallel supply-side investment in housing, healthcare, and care infrastructure, fund distributions could erode their own purchasing power.
The supply-side prerequisite
OpenAI’s paper contains the seeds of an answer here, though it doesn’t connect the dots explicitly. Several of their other proposals target exactly the supply constraints that would make a wealth fund inflationary: grid expansion to lower energy costs, care economy investment to expand healthcare and childcare capacity, AI-enabled labs to accelerate drug development and scientific discovery.
If you read the paper as a package rather than a menu, the implicit argument is that the supply-side investments need to come first or at least in parallel. The wealth fund works only if the economy it’s distributing returns from is actually producing cheaper, more abundant goods.
This sequencing problem is the crux of the policy challenge. Building energy infrastructure takes years. Training healthcare workers takes years. Reforming housing policy takes decades. AI capabilities, by OpenAI’s own projections, are advancing on a timeline measured in months.
The gap between when AI concentrates wealth and when supply-side capacity catches up is the window where inflation risk is highest, and where a poorly timed wealth fund could do real damage to the people it’s meant to help.
But this framing has a blind spot. The argument that “the physical world doesn’t yield to software economics” is true in the short run but risks underestimating what happens when AI starts attacking the physical bottlenecks directly. AI-designed buildings that optimize for faster construction. Automated permitting simulations that compress months of regulatory review into days. Robotic construction that breaks the labor constraint entirely. AI-trained healthcare workers who reach competency faster. Synthetic biology that reshapes food and energy production at the molecular level. None of this is here yet at scale, but the compounding effects of agentic systems and robotics could close the supply gap faster than a linear projection would suggest.
Every previous technology shift, from electrification to the internet, ultimately expanded supply more than it destroyed demand. The pessimistic supply-side framing assumes AI is different because it displaces cognitive labor. But if AI also displaces the physical and regulatory bottlenecks that constrain supply, the inflationary window could be shorter than it appears. The honest answer is that nobody knows the timeline. The wealth fund’s design should account for both possibilities: a long supply-side lag where distributions are inflationary, and a faster one where they aren’t. The worst outcome is designing the fund for one scenario and getting the other.
What this means for the wealth fund’s viability
The Citrini scenario exposes the core vulnerability of the Public Wealth Fund as a policy tool. The fund assumes it can capture AI-driven growth and redistribute it. But if Citrini’s feedback loops play out, the growth the fund is capturing is a mirage, corporate profits inflated by cost-cutting that is simultaneously destroying the consumer economy those profits depend on.
In Citrini’s timeline, the S&P peaks at 8000 in October 2026, then falls 38% as the displacement spiral reaches the financial system. A wealth fund seeded during the euphoria phase and invested in equities would be distributing returns from an asset base that’s about to crater. Citizens expecting dividends from the fund would instead watch their national investment vehicle lose a third of its value, precisely when they need support most.
This is the timing problem scaled up. The fund needs the financial assets it holds to reflect real productive capacity rather than a labor-arbitrage bubble. If AI stocks are surging because companies are cutting headcount, not because the economy is genuinely more productive in a way that supports broad-based consumption, then the fund is investing in the problem, not the solution.
What a non-inflationary version looks like
A Public Wealth Fund doesn’t have to be inflationary if it’s designed with supply-side discipline.
Tie distributions to productivity metrics, not political cycles. Payouts should scale with measured increases in productive capacity, not with fund asset values alone. If the economy isn’t producing more, distributing more is just redistribution with extra steps and price pressure.
Invest fund capital in supply-side assets. Rather than a pure financial portfolio, allocate a significant share to infrastructure, housing, energy, and healthcare capacity. The fund’s returns should partly come from expanding the supply of the goods its beneficiaries need most.
Start with in-kind benefits before cash. Early distributions could take the form of subsidized access to AI tools, healthcare, education, or energy, things that expand capability without creating pure demand-side pressure. Cash distributions can scale up as the supply side catches up.
Build in automatic stabilizers. If inflation in essential categories exceeds target thresholds, distributions should automatically reduce or pause. This removes the political pressure to maintain payouts during inflationary periods.
Whether any of these design principles survive contact with the political process is a separate question, and one that both OpenAI and Citrini, in their own ways, answer pessimistically.
The political obstacles deserve more than a hand-wave. A U.S. sovereign wealth fund would be the largest peacetime intervention in capital markets in American history. Every design parameter becomes a political battleground: who’s eligible, how much they receive, which assets the fund holds, which industries it avoids, who sits on the board, and whether distributions are universal or means-tested. The partisan dynamics are predictable. The right frames it as socialism with a Bloomberg terminal. The left fights over whether distributions should be flat or progressive. Both sides try to use the fund’s holdings as leverage, pressuring boards of companies the fund owns stakes in to align with whatever the current administration wants.
The Alaska Permanent Fund is instructive here not just for payout pressure but for the full spectrum of governance dysfunction. Alaskans have repeatedly rejected attempts to use fund earnings for public services, preferring direct checks. Any proposal to invest in supply-side assets rather than maximizing financial returns would face the same resistance, multiplied by 330 million stakeholders instead of 700,000. The fund’s mandate would be pulled between “maximize returns for distribution” and “invest in the supply-side capacity that prevents your distributions from being inflationary,” and those objectives conflict in the short term.
OpenAI’s vagueness on mechanics could be strategic ambiguity designed to avoid committing to positions that would generate opposition before the conversation even starts. Or it could be an honest admission that the details must emerge through democratic process because no single company should dictate the structure of a national wealth fund. Both readings are plausible. The charitable interpretation is that OpenAI is offering the concept while leaving the architecture to the institutions that would be accountable for running it. The skeptical interpretation is that vagueness lets the proposal function as a political signal without requiring the hard trade-offs that would make it real.
What this means for asset prices
A Public Wealth Fund also reshapes capital markets. The fund itself would be a massive new buyer of assets, and the policy signals around it would shift how markets price risk, inflation expectations, and the role of alternative stores of value. The Citrini scenario complicates every asset class, because the macro regime the fund operates in determines which way prices move.
Stocks: the obvious winner, with a catch
Equities are the most direct beneficiary. A nationally seeded fund buying diversified, long-term assets is, by definition, a persistent new source of demand for stocks. Think of the upward pressure Japan’s Government Pension Investment Fund or Norway’s sovereign fund puts on global equities, then scale it for the U.S. market.
AI-exposed companies benefit twice: once from the productivity gains driving their earnings, and again from a government-backed fund allocated to capture exactly that growth. The largest tech firms, the ones building and deploying frontier AI, become structurally bid.
The catch is valuation risk, and this is where Citrini’s scenario bites. If markets front-run the fund’s creation, pricing in a permanent new buyer before it actually deploys capital, you get a speculative run-up disconnected from earnings. In Citrini’s timeline, that run-up peaks at S&P 8000 before the displacement spiral pulls the floor out. A wealth fund seeded at those valuations is buying at the top of a labor-arbitrage bubble. The correction doesn’t just hurt Wall Street. It hits the fund’s beneficiaries, the citizens expecting returns from an asset base that just lost a third of its value.
There’s also a governance problem. If the fund holds significant stakes in major corporations, who votes those shares? How do you prevent political influence over corporate boards? Norway navigates this with strict ethical guidelines and passive management. An American version would face far more political pressure to use its holdings as leverage, which introduces a whole category of distortion the market isn’t currently pricing.
Gold: the inflation hedge that prices the fund’s credibility
Gold’s response to a Public Wealth Fund is really a referendum on whether markets believe the supply-side story. If investors trust that AI productivity gains will keep inflation contained even as distributions flow to citizens, gold stays range-bound. It’s not needed as a hedge because the deflationary force of AI offsets the inflationary force of redistribution.
If investors don’t trust that story, gold runs. The logic is straightforward: a government creating a new redistribution vehicle funded partly by monetary expansion or deficit spending is exactly the macro regime gold is designed to price. It doesn’t matter whether the fund is technically funded by corporate contributions or tax revenue. What matters is the fiscal posture around it, whether government spending rises to support the parallel supply-side investments the fund depends on, and whether the Fed accommodates that spending.
Gold also benefits from a subtler dynamic. A Public Wealth Fund signals that the government views AI-driven wealth concentration as a structural problem requiring permanent intervention. That’s a regime change in fiscal policy. Markets that price in larger, more persistent government involvement in capital allocation tend to bid up hard assets as a hedge against policy error.
The Citrini scenario complicates this. If the displacement spiral produces real-economy deflation while the government responds with massive fiscal transfers, gold rallies on the fiscal response even as the underlying economy contracts. Gold becomes a hedge not against consumer inflation but against fiscal desperation, a bet that the government will keep spending into the void regardless of whether it’s working.
Bitcoin: the most interesting case
Bitcoin’s reaction to a Public Wealth Fund would be layered and partly contradictory.
The inflationary scenario is bullish for bitcoin in the same way it’s bullish for gold, but more so. Bitcoin’s fixed supply makes it a purer expression of the “governments will debase currency to fund redistribution” thesis. If the fund’s creation coincides with expanded fiscal deficits, money supply growth, or any perception that the Fed is accommodating a new spending regime, bitcoin absorbs that narrative faster and more aggressively than gold.
But there’s a second, more structural bull case. A Public Wealth Fund is a centralized vehicle controlling a significant share of national wealth, managed by government appointees, subject to political cycles. For the segment of the population that already distrusts institutional control over capital, the fund’s existence validates the case for a decentralized, permissionless alternative. Bitcoin hedges not just inflation in this framing but the concentration of financial power in a single government-managed entity.
Citrini’s scenario adds a third angle. Their timeline includes agentic commerce routing around card interchange fees via stablecoins on Solana and Ethereum L2s. If that payments shift happens, crypto infrastructure gets embedded in the real economy independent of the macro speculation narrative. Bitcoin benefits from that ecosystem expansion regardless of whether the broader crisis plays out.
The bear case is less obvious but real. If a Public Wealth Fund actually works, if it delivers broad-based returns, reduces inequality, and operates transparently, it undermines the narrative that government-managed financial systems are inherently captured or incompetent. A well-functioning fund reduces the urgency of the “exit to crypto” thesis. Most people don’t want to be their own central bank. They want institutions that work. A successful wealth fund is exactly that.
There’s also a regulatory dimension. A government managing a multi-trillion dollar fund has strong incentives to maintain control over monetary policy and capital flows. That could translate into tighter crypto regulation, not out of hostility but out of a desire to keep the fund’s macro environment stable and predictable. Bitcoin thrives in policy uncertainty. A confident, well-capitalized fiscal state is a less favorable environment than a panicking one.
Putting it together
The Public Wealth Fund’s impact on asset prices comes down to the same variable that determines its inflationary impact: whether AI delivers real supply-side abundance or just financial returns concentrated in equity markets.
If abundance is real, stocks grind higher on earnings growth, gold stays flat, and bitcoin trades on its own adoption cycle rather than macro fear. If abundance is overpromised and the fund becomes a transfer mechanism in a supply-constrained economy, all three inflate in nominal terms, but only gold and bitcoin preserve real purchasing power. Stocks rise in price while the currency they’re denominated in loses value.
If the Citrini scenario plays out, the sequencing is different again. Stocks crash first as the displacement spiral hits earnings, gold and bitcoin rally on the fiscal response, and the fund itself becomes a casualty of the crisis it was designed to prevent.
The most likely outcome is a messy middle. Some sectors experience genuine AI-driven deflation while others stay constrained. The fund delivers modest returns that help but don’t transform. And asset markets spend a decade trying to price the difference between the two, which means elevated volatility across all three asset classes as the market repeatedly re-evaluates whether the abundance story is real.
The velocity mismatch
Neither OpenAI nor Citrini says this outright, but both demonstrate it: the speed of AI capability improvement is structurally faster than the speed of institutional response.
OpenAI’s paper proposes legislative action, institutional design, new tax frameworks, international coordination, auditing regimes, and a sovereign wealth fund. All of it requires political consensus, bureaucratic implementation, and years of iteration. Their own technology, by their own admission, is advancing from tasks that take hours to projects that take months, on a timeline measured in quarters.
Citrini’s entire narrative is built on this gap. The canary dies not because the risks are unforeseeable, but because the policy response can’t keep pace. OpenAI publishing a wish list doesn’t close that gap. It documents it.
A Public Wealth Fund that takes five years to legislate, structure, capitalize, and deploy is not a response to a displacement spiral that compounds quarterly. It’s a memorial.
The bigger question
The Public Wealth Fund idea reveals a tension at the center of OpenAI’s policy vision. They’re simultaneously arguing that AI will create unprecedented abundance (lower costs, higher productivity, scientific breakthroughs) and that the government needs to build massive new redistribution infrastructure to make sure people benefit from it.
If the abundance story is true, prices should fall and living standards should rise without a wealth fund. If the redistribution story is necessary, it’s because the abundance isn’t reaching people through markets, which means the supply-side transformation is either slower, less evenly distributed, or more captured by capital than the optimistic pitch suggests.
Both things can be true at once. But the policy design has to account for the gap between them. A Public Wealth Fund that assumes abundance while operating in scarcity isn’t a safety net. It’s an inflation engine. And if Citrini’s scenario is even directionally correct, the fund could be seeded at the top of a bubble it helped inflate, holding assets whose valuations depend on the very displacement it was created to offset.
The conversation OpenAI wants to start is the right one. The answer just has to be more honest about sequencing, trade-offs, and the stubborn physical constraints that don’t bend to software economics, no matter how intelligent the software gets. The canary is still alive. But the people building the mine are the ones telling you to buy canary insurance.
