Executive Summary

The leading US AI companies (Nvidia, Microsoft, Alphabet, Apple, Meta, Tesla, OpenAI and others) are valued in a way that would imply they need to generate $2.4 trillion in additional foreign revenues annually by 2036. This equals the entire current US goods export and is more than double the current US current account deficit. The authors argue that these valuations either presuppose a massive restructuring of the global economy – or represent a bubble. A central tension emerges from the Trump Administration: protectionist trade policy obstructs precisely the market opening that US tech firms need for their projected profits.

People

  • Ricardo Hausmann (Author; Professor Harvard Kennedy School, former Chief Economist of the Inter-American Development Bank)
  • Andrés Velasco (Co-author; former Finance Minister Chile, Dean LSE School of Public Policy)

Topics

  • AI valuations and market bubbles
  • Global economic imbalances
  • US tech monopolies and international dependency
  • Trade policy and protectionism
  • Corporate taxation
  • Geopolitical power shifts

Clarus Lead

The valuations of AI giants pose a provocative question: either the world pays trillions to a handful of American companies – or the market has massively overvalued them. In doing so, a fundamental contradiction of the Trump era reveals itself: Silicon Valley needs global markets that MAGA trade policy shuts down. Furthermore, the tax question is being reframed – returns of $23 million per employee at these conglomerates will provoke governments worldwide to seek redistribution.

Detailed Summary

Hausmann and Velasco model a scenario in which the core AI companies under consideration are trading by 2036 with a price-to-earnings ratio of 20, achieve a net profit margin of 20%, and derive 65% of their additional revenues from abroad. Under these conservative assumptions, an annual foreign revenue growth of $2.4 trillion would result – a sum that implies structural shifts in global division of labor. The US cannot simultaneously force protectionism and expect foreign markets to purchase American technology at an unprecedented scale. Countries whose dollar access is restricted by higher tariffs cannot simultaneously transfer trillions to US corporations.

The authors highlight a second critical element: taxation. The 11 leading AI companies together employ fewer than one million people and have a market value of $23 million per employee. This is not an employment story – it is an entitlement story. National governments, states, and foreign countries will demand shares of these rents. Digital tax debates are merely a prelude; a global tax battle over AI profits will be inevitable. Each additional tax lowers the profits that shareholders receive, and thus the justification for today's valuations.

Geopolitically, the rest of the world will not remain passive. Countries will subsidize domestic AI alternatives, introduce local hosting requirements, favor national champions in procurement, and enforce stronger competition and data protection rules. This signals not the decline of American power, but its transformation: from the 20th-century model (manufacturing scale, military, dollar power) to a new model based on indispensable AI infrastructure. The central tension remains: how do the US sell this infrastructure globally while simultaneously shutting down markets through protectionism?

Key Findings

  • Valuation Logic: Current AI valuations require US corporations to generate additional foreign revenues of $2.4 trillion per year by 2036 – comparable to the entire US goods export today.

  • Political Contradiction: Trump's protectionism and Silicon Valley's business model are in fundamental conflict; global markets cannot simultaneously be shut down and monetized.

  • Tax Trap: Windfall profits ($23 million per employee) will drive governments globally toward taxation, which directly undermines valuations.

  • Geopolitical Counter-Scenario: Countries will reduce dependencies through subsidies, local requirements, and competition measures, not from weakness, but from self-protection.

Critical Questions

  1. Evidence/Data Quality: Are the 65% foreign revenue assumption and 20% profit margin projections based on historical data from these companies, or are they speculative scenarios? How robust are these parameters against sensitivity analyses?

  2. Conflicts of Interest: The authors are academic experts with ties to multilateral organizations – is there a structural bias in favor of redistribution and taxation solutions over tech expansion?

  3. Causality/Counter-hypotheses: Could a significant portion of AI profits stem not from foreign sales but from productivity increases in the US domestic market, which do not require exports?

  4. Feasibility/Risks: If the US increases protectionism while simultaneously needing to maximize market access for AI services – what policy instruments could actually resolve this tension without leading to trade war or disinvestment?

  5. Taxation Realism: Assuming countries implement national AI taxes – would global AI investment then decline or would corporate headquarters shift strategically?

  6. Bubble Indicators: What concrete market signals (profit growth, cost of capital, venture funding) would indicate overvaluation if current prices are already elevated?

Bibliography

Primary Source: Hausmann, Ricardo & Velasco, Andrés: "AI Valuations – The World Must Pay Trillions to US Tech Firms – or It's a Bubble" – Finance and Economics (20.04.2026)
https://www.fuw.ch/ai-boom-wie-die-welt-fuer-us-technologie-zahlen-soll-607245505043

Verification Status: ✓ 20.04.2026

This text was created with the support of an AI model.
Editorial Responsibility: clarus.news | Fact-Check: 20.04.2026