A new working paper from the Levy Economics Institute, "Plutonomy — the AI Edition — and the Coming Crisis"1 by Yeva Nersisyan and L. Randall Wray, arrives as a serious and sobering diagnosis of the American financial condition in 2026. Drawing on Galbraith, Minsky, and FDR's New Deal, the authors construct a historically grounded argument that today's AI-driven equity bubble mirrors the investment trust mania of the late 1920s with uncomfortable precision, and that the institutional capacity to absorb the coming shock is far weaker than most assume.
A Plutonomy Runs on Asset Prices, Not Wages
The paper revives a term coined in a 2005 Citigroup memo: the plutonomy, an economy driven not by broad consumer demand but by the spending and asset accumulation of the ultra-wealthy. Nersisyan and Wray note that labor's share of national income has fallen from roughly 52 percent in the early postwar period to below 43 percent today. The implication is structural: "our labor force — the majority of income earners — receives less than half of national income," depressing consumption, production, and employment. The economy compensates by relying on rising asset prices to sustain spending at the top, where the top 0.1 percent of Americans own nearly double the stock held by the bottom 90 percent combined.
Circular Finance, Then and Now
The authors draw a direct and troubling analogy between the investment trust pyramid schemes of the 1920s and today's circular finance in the tech sector. Galbraith described how newly formed trusts would issue shares to other trusts, generating artificial valuations disconnected from any underlying business revenue. Today's equivalent, Nersisyan and Wray argue, is Big Tech investing in AI startups using cloud computing credits rather than cash, with those startups then purchasing services from the same firms. "This circular finance," they write, "artificially inflates the value of both companies." The structural mechanics are near-identical. So is the endpoint.
The Bezzle Builds
Galbraith's concept of the "bezzle" — undiscovered embezzlement that expands in booms and is revealed only in crashes — runs as a thread through the paper. The authors apply it directly to the AI sector's financing practices: private credit funds lending to tech companies through payment-in-kind structures, which allow borrowers to roll interest into principal rather than service it, are described plainly as Ponzi finance in Minsky's taxonomy. The same circular exposure runs through pension funds, bank credit lines, and retail investors newly admitted to mega-IPOs through lowered Fidelity thresholds. "Much like the subprime mortgage-backed securities of the 2000s or the investment trusts of the 1920s," Nersisyan and Wray write, "these IPOs represent the worst excesses of financial capitalism."
SpaceX, Nasdaq, and the Fast-Track
The Magnificent Seven IPO cycle is treated as a pressure-release mechanism for the bezzle, not a sign of health. Nasdaq's relaxation of listing standards — creating a fast-entry track for newly public stocks reaching top-40 market cap within 15 trading days — ensures that passive index funds become forced buyers. Worker pension capital flows automatically into the bubble whether or not the underlying businesses generate revenue. SpaceX's dual-class share structure, in which Musk retains 40 percent of shares and 80 percent of votes, is emblematic: public investors fund the asset, but the plutocrat controls it.
What Comes Next, and What Should
The BIS 2026 Annual Report, cited by the authors, warns that a reversal of AI optimism "could have major financial consequences, given AI firms' rising leverage and growing footprint in credit markets." Nersisyan and Wray are direct about the policy fork ahead: Hoover's liquidationism, the post-GFC Wall Street bailout model, or FDR's tougher restructuring path. Their preference is the third, paired with a Keynesian-Minskyan fiscal approach that reorients the economy toward mass consumption, constrains speculative finance, and rebuilds the institutional safeguards dismantled over the past 45 years. Without that, the crash resolves nothing structurally.
Five Key Takeaways for Advisors and Investors
1 The AI bubble is structurally similar to the 1920s investment trust mania. Circular finance between hyperscalers, chip makers, and AI labs means that liquidation is not a tail risk — it is embedded in the architecture.
2 Private credit is not a safe haven. PIK loan structures, CLO tranching, and deep tech-sector exposure mean private credit funds carry Minsky-classified Ponzi finance risk that flows through pension and 401k exposure.
3 Passive index inclusion amplifies the bust. Fast-track Nasdaq rules turn index funds into forced buyers at peak valuations. Any correction in top-40 AI names hits broadly held portfolios automatically.
4 Labor's declining income share suppresses the demand base. An economy relying on plutocratic consumption and asset appreciation for growth is inherently fragile. Any sustained correction in asset prices removes the primary engine.
5 The policy response will determine the damage. Unlike the New Deal, the post-GFC response imposed minimal constraints on finance. If that model repeats, Main Street absorbs the pain while the financial architecture that caused the crisis remains intact.
Footnote:
1 Nersisyan, Yeva, and L. Randall Wray. "Plutonomy — the AI Edition — and the Coming Crisis." Levy Economics Institute Working Paper No. 1122, July 2026. https://www.levyinstitute.org/wp-content/uploads/2026/07/wp_1122.pdf
