Early tech giants like Google and AWS built monopolies because their potential wasn't widely understood, allowing them to grow without intense competition. In contrast, because everyone knows AI will be massive, the resulting competition and capital influx make it difficult for any single player to establish a monopoly.

Contrary to expectations, analysis shows that sectors with low profit per employee, such as healthcare and consumer staples, stand to gain the most from AI. High-tech firms already have very high profit per employee, so the relative impact of AI-driven efficiency is smaller.

Traditional metrics like GDP fail to capture the value of intangibles from the digital economy. Profit margins, which reflect real-world productivity gains from technology, provide a more accurate and immediate measure of its true economic impact.

History shows that transformative innovations like airlines, vaccines, and PCs, while beneficial to society, often fail to create sustained, concentrated shareholder value as they become commoditized. This suggests the massive valuations in AI may be misplaced, with the technology's benefits accruing more to users than investors in the long run.

During major platform shifts like AI, it's tempting to project that companies will capture all the value they create. However, competitive forces ensure the vast majority of productivity gains (the "surplus") flows to end-users, not the technology creators.

The most profound innovations in history, like vaccines, PCs, and air travel, distributed value broadly to society rather than being captured by a few corporations. AI could follow this pattern, benefiting the public more than a handful of tech giants, especially with geopolitical pressures forcing commoditization.

Marks warns against conflating a technology's societal impact with its investment potential. Fierce competition among AI service providers or their customers could pass all productivity gains to consumers through lower prices. This would result in little to no profit for the underlying companies, echoing a similar warning from Warren Buffett during the dot-com era.

JP Morgan's analysis that AI needs to generate '$34/month from every iPhone user' to see a return is a flawed framing. Like cloud computing, the cost and value of AI will be embedded into thousands of different products and services, not borne as a direct consumer subscription. This indirect value capture makes direct per-user ROI calculations misleading.

Marks questions whether companies will use AI-driven cost savings to boost profit margins or if competition will force them into price wars. If the latter occurs, the primary beneficiaries of AI's efficiency will be customers, not shareholders, limiting the technology's impact on corporate profitability.