Updated · Marcus on AI | Gary Marcus | Substack · Jun 7
AI Generates More Output but Little ROI, Burning $1,000 for Every $100 of Revenue
Updated
Updated · Marcus on AI | Gary Marcus | Substack · Jun 7
AI Generates More Output but Little ROI, Burning $1,000 for Every $100 of Revenue
2 articles · Updated · Marcus on AI | Gary Marcus | Substack · Jun 7
Summary
Studies cited from MIT, McKinsey and Bain say generative AI is producing far more apps, books, papers and web content without materially lifting GDP or delivering clear returns for many companies.
Book output illustrates the gap: available titles have surged, while book sales have edged down, echoing similar jumps in music uploads, self-filed lawsuits and scientific submissions with little evidence of better quality.
Math researchers have warned the flood is now straining verification itself; the Leiden Declaration says automated systems can generate plausible but unreliable proofs that are hard to distinguish from correct work.
Coding remains the strongest candidate for real productivity gains, but it is also the most cash-intensive use case, with one analysis arguing Anthropic and OpenAI may spend $1,000 for every $100 customers pay.
That cost imbalance suggests AI's biggest apparent productivity win could weaken if providers raise prices enough to cover losses, potentially making automation costlier than the human labor it replaces.
AI is driving GDP but not productivity. Is this tech boom just a high-cost illusion of progress?
With most firms failing to profit from AI, is it destined to only enrich a handful of tech giants?
Trillions Invested, Returns Uncertain: The High-Stakes Reality of AI’s Productivity Paradox
Overview
The report highlights a dramatic surge in AI investment, with industry leaders projecting capital expenditures for major tech companies to reach up to $4 trillion annually by decade’s end. Despite this unprecedented spending, many organizations are struggling to achieve measurable returns on their AI investments. While forecasts for AI revenue are rising, actual profitability remains elusive due to high operational costs and challenges in scaling AI projects. This disconnect is driving increased scrutiny from investors and a shift toward demanding clear, tangible value from AI initiatives, signaling a new era of accountability and strategic focus in the AI sector.