Opinion | The Hallucinatory AI Math

2026-05-31 · Source: Technology - WSJ.com · Field: Finance & Economics — Capital Markets & Investment Management, Economic Analysis & Policy · Depth: Fundamental Awareness, quick

Summary

The article critiques inflated market projections for AI-related companies, labeling them "hallucinatory AI math." SpaceX's recent S-1 filing claims a \$28.5 trillion "actionable total addressable market," with \$23 trillion attributed to enterprise artificial intelligence applications, a figure nearly matching the entire U.S. economy's \$31 trillion annual output. This trend follows AI chip maker Cerebras's May IPO, where its stock nearly doubled. Cerebras CEO Andrew Feldman projected a nearly \$5 trillion market from a single use case, estimating 47 million software engineers each spending \$100,000 on AI tokens annually. These projections are driving significant investor enthusiasm, but the analysis warns of an inevitable price drop.

Key takeaway

For investors evaluating AI-related IPOs or growth stocks, scrutinize the underlying market projections. Wildly optimistic total addressable market (TAM) figures, like \$23 trillion for enterprise AI or \$5 trillion from a single use case, often signal unsustainable valuations. You should compare these projections against established economic realities, such as the \$31 trillion U.S. economy, to identify potential "hallucinatory math" and mitigate risk before an inevitable price correction.

Key insights

The article highlights highly speculative and potentially unrealistic market valuations for AI-related ventures.

Principles

• Inflated TAMs often precede market corrections.
• Bottoms-up projections can mask market realities.
• Investor enthusiasm can override economic fundamentals.

Method

The article describes a method of market projection: multiplying a large user base (e.g., 47 million software engineers) by a high per-user spend (e.g., \$100,000 in tokens) to generate massive total addressable market figures.

In practice

• Scrutinize AI market size claims.
• Compare TAMs to broader economic output.
• Evaluate per-user spend assumptions critically.

Topics

• AI Market Valuation
• Initial Public Offerings
• Total Addressable Market
• Investment Risk
• Cerebras Systems
• SpaceX

Best for: Investor, Entrepreneur, Executive

Related on AIssential

• AI Weekly Issue #494: SpaceX wants $80 billion. OpenAI wants a trillion. — Public markets are now scrutinizing the high valuations of major AI companies, shifting from private to public pricing.
• The AI bubble is all over now, baby blue — The AI investment bubble is collapsing due to LLM's inherent technical limitations and unsustainable economics.
• $700 Billion in Capex. $50 Billion in Revenue. AI’s Math Is Broken. — Despite exponential growth, major AI labs struggle with profitability due to costs scaling faster than revenue and rapid commoditization.
• OpenAI's Historic US$1tn IPO Plan and the Looming AI Bubble — OpenAI's US\$1 trillion IPO plan highlights both immense AI market potential and significant bubble concerns.
• Tokenomics Is Dead. The Lie Of Affordable Inference — Current "cheap inference" is a VC and hyperscaler-subsidized illusion, masking significant AI industry losses.
• Sam Altman and the day Nvidia’s meteoric rise came to an end — The "scaling = AGI" myth's collapse led to a tech stock plateau, revealing LLMs' commodity nature.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Technology - WSJ.com.