Is This Time Truly Different?

“The four most expensive words in the English language are, ‘This time it’s different.’”

Sir John Templeton

The US stock market has entered uncharted territory. With a total market capitalisation of $75 trillion, it now sits at an unparalleled 240–260% of US Gross Domestic Product. For context, the historical baseline ranges from just 75% to 100%. Even during the dot-com bubble, this metric reached only 140%, and at the 2021 market peak, it hit 200%. Another warning sign is the Cyclically Adjusted Price-to-Earnings (CAPE) ratio. Over its 150-year history, the CAPE ratio has crossed into the 40s only twice: during the late-1990s dot-com boom, and right now. Against this historical backdrop, many believe a stock market crash is no longer a question of if, but when.

While many believe the stock market is severely overvalued, we must ask a critical question: Are historical valuation methods still appropriate for the current market?

Traditional valuation relies heavily on the CAPE ratio, which weighs current stock prices against ten years of past earnings. But in a market driven by technology which seeks to fundamentally alter productivity, looking in the rear-view mirror would appear imprudent.

Far from a speculative bubble, compelling evidence suggests today’s market valuations are not merely justified - they may actually be understated. Decoding this phenomenon requires looking beyond traditional financial metrics and examining the economy through a new lens: post-labour economics.

Throughout history, human innovation has been driven by a single goal: to minimise physical toil and maximise productivity. In the fifteenth century, the printing press mechanised the reproduction of knowledge, enabling a printing shop to produce in a day what a scribe might copy in a year. During the Industrial Revolution, the steam engine, power loom and locomotive substituted machine power for human and animal strength, transforming manufacturing and transport. Twentieth-century computing extended the same process from muscle to information, automating calculation and data processing at a scale no human workforce could match. For millennia, we have been steadily converging toward an economic singularity - a future where human labour is ultimately made entirely redundant.

To understand the threat to the modern workforce, we must categorise human value into four areas: strength, dexterity, cognition, and empathy. Mechanical systems conquered strength centuries ago. Today, AI threatens cognition, while AI-driven robotics increasingly threatens dexterity. With AI models already outperforming humans across many domains of complex problem-solving and knowledge work, the white-collar sector is facing generational disruption. Meanwhile, the rapid advancement of AI-driven humanoid robots threatens to replace the skilled physical labour of tradesmen, builders, and logistics workers.

Time and time again, history has proven that superior technology always wins. If a system is better, faster, cheaper, and - where applicable - safer, it is economically unviable to retain human labour. The mechanisation of agriculture provides perhaps the clearest precedent. In 1900, agriculture employed 41% of the US workforce; by 2000, that figure had fallen to just 1.9%. Between 1948 and 2017, US farm output nearly tripled even as the number of labour hours worked fell by more than 80%. Tractors, combine harvesters and automated machinery did not merely make agricultural workers more productive - they permanently eliminated the need for much of their labour.

The optimal number of employees is zero.

We are approaching a threshold where AI and robotics will be fully deployable across the economy. When that happens, mass labour displacement will inevitably follow, driven by a simple corporate reality: the optimal number of employees is zero. Biological workers come with strict biological and financial constraints: they need rest and are prone to human error, yet must be compensated with salaries, healthcare, and benefits. By eliminating these friction points, corporations will unlock a level of profitability previously thought impossible. Ultimately, it is this forthcoming explosion in corporate earnings that justifies today’s historic market valuations.

AI’s effect on corporate profits will depend not simply on the productivity it creates, but on where those gains ultimately accrue. Competitive pressure will transfer some of this value to consumers, yet the S&P 500 is increasingly concentrated in firms that control the foundations of the digital economy - from advanced chips and data centres to cloud platforms, software ecosystems and proprietary data. These economic moats preserve pricing power, allowing a greater share of AI-driven cost savings to be retained as profit.

Falling prices offer a second route to profit. Henry Ford demonstrated this with the Model T, using assembly-line efficiency to reduce its price from $850 to as little as $260. Rather than diminishing the commercial opportunity, the lower price unlocked a vast mass market, with more than 15 million Model T’s eventually sold.

AI could reproduce this pattern across the entire economy. As the cost of intelligence falls, it will evolve from a premium technology into a general-purpose input embedded throughout software, services, manufacturing and logistics. Even if competition drives prices lower, demand may expand faster still. The resulting value would be captured through two channels: margin expansion among the companies controlling essential infrastructure, and explosive volume growth among those using it to serve newly viable markets.

This is the crucial connection between post-labour economics and today’s historic valuations. The CAPE ratio compares current prices with ten years of earnings generated by an economy in which output remained closely tied to human labour. If AI and robotics begin to weaken that relationship, the comparison ceases to be like-for-like: current prices reflect post-labour expectations, while the earnings base remains rooted in a labour-constrained economy.

Templeton’s warning nevertheless retains its force. A genuine technological revolution can still be overvalued, and structural change cannot justify any price. The question, therefore, is not whether AI and robotics can transform corporate earning power, but whether today’s prices already reflect too much of that transformation. In my view, today’s historic multiples do not indicate a fundamentally overvalued market or an imminent crash. Rather, they reflect an anticipated expansion in corporate earning power that backward-looking measures cannot capture. In that crucial respect, this time truly is different.