Anthropic forecasts U.S. GDP at $44T by 2030 due to AI (ANTHRO) | Seeking Alpha

TEXT START: Anthropic (ANTHRO) co-founder Jack Clark said on Saturday that one of his company’s internal economic models projected U.S. economic output to reach $44T by 2030 thanks to fast-improving AI capabilities that could transform all fields at a more rapid pace.

The Dissection

The text converts AI’s potential to expand aggregate output into evidence of economic health. It treats a larger GDP number as if it automatically means broader prosperity, stable employment, or continued mass participation. It does not. The forecast is an output claim, not a distribution claim.

Its real function is to present displacement as abundance. AI destroys the wage mechanism while increasing the measured value of what automated systems produce. The machine can make the economy richer while making most humans economically unnecessary.

The Core Fallacy

The central error is confusing production with participation. Under the Discontinuity Thesis, the decisive break occurs when AI severs the mass employment → wage → consumption circuit. A $44T economy does not repair that circuit. It may intensify the break by concentrating productive capacity and income among AI owners, infrastructure controllers, and firms able to deploy automation.

GDP can rise while labor loses bargaining power, wages stagnate or vanish, and ownership becomes the only durable claim on output. The forecast therefore supports the thesis more readily than it refutes it.

Hidden Assumptions

• AI capability will translate into deployable production across “all fields” on the stated timetable.
• GDP measurement will capture machine-generated output cleanly rather than misclassifying, omitting, or lagging behind it.
• Energy, compute, chips, logistics, regulation, and physical maintenance will not become binding constraints.
• The gains will be distributed widely enough to preserve consumption without destabilizing political legitimacy.
• Humans displaced from necessary work will retain purchasing power through wages, transfers, ownership, or some other mechanism.
• Higher output will preserve social cohesion even if productive participation collapses.
• The institutions that currently distribute income will survive the loss of their economic foundation.

The most important assumption is smuggled in silently: that economic growth remains socially equivalent to human viability. That equivalence is exactly what AI automation destroys.

Social Function

Primary classification: elite self-exoneration and ideological anesthetic, with a partial truth underneath.

The partial truth is that AI could generate enormous productivity gains. The anesthetic is the implication that those gains constitute a general solution. The forecast gives capital a clean moral narrative: displacement is not destruction; it is merely the prelude to a larger number on a national accounts spreadsheet.

It also functions as transition management. By advertising abundance before confronting ownership, distribution, and participation, it buys time for incumbents to consolidate the productive assets that generate the abundance.

The Verdict

This is not evidence that the post-WWII economic order survives. It is a projection of a richer machine economy standing on the ruins of the labor economy. If the forecast is even partly correct, the central question is not whether America becomes more productive. It is who owns the systems producing the $44T—and what economic purpose remains for everyone excluded from them.