AI Valuations Surge as Market Shapes Reality Through Observation

A new study led by scientists at MIT has reignited the philosophical debate about the role of observation in shaping reality, echoing the long-standing dispute between physicists Albert Einstein and Niels Bohr [1]. By observing the behavior of a single photon bouncing off two atoms, the researchers confirmed that the photon’s state as a wave or particle was not determined until it was measured [1]. This supports Bohr’s theory that observation is not a passive act but an active process that helps create the reality being observed [1].

The implications of this finding extend beyond quantum mechanics and into financial markets, where observation similarly influences outcomes. As noted by George Soros, financial markets are constantly anticipating events, and these anticipations can alter the probability of those events occurring [1]. A case in point is the recent US labor data, where a downward revision of 253,000 jobs over the previous two months significantly shifted market expectations [1]. Following the release, 10-year Treasury yields fell 20 basis points, and the probability of a Federal Reserve rate cut increased by 15 percentage points [1].

The same principle applies to the rapid valuation increases among the four major AI hyperscalers—Microsoft, Alphabet, Amazon, and Meta. Their combined market capitalization has surpassed $10 trillion, largely based on the expectation that they will dominate the next wave of AI advancements [1]. Sam Altman and Mark Zuckerberg have both expressed optimism about the trajectory of AI, with the latter noting that “superintelligence is now in sight” [1]. These predictions are not merely speculative; they are shaping real-world investments. Morgan Stanley estimates that the four hyperscalers will invest $2.9 trillion in AI-related capital expenditures between 2025 and 2028, but will fall short by $1.5 trillion in available funding [1]. The only way to bridge this gap, according to current market dynamics, is through the high valuations assigned by investors [1].

Meanwhile, AI research itself is evolving rapidly. A team of researchers has recently developed an open-source model that enables AI to innovate its own architecture [1]. This development, as experts suggest, is shifting the pace of AI research from a human-limited to a computation-scalable process [1]. As AI systems become more capable of self-improvement, the speed at which they advance could accelerate dramatically, further reinforcing the expectations embedded in current market valuations [1].

The connection between observation and economic outcomes is also evident in the surge of capital expenditures among the hyperscalers. The four companies are projected to collectively spend $400 billion on capex this year alone [1]. This aggressive investment is not only driven by anticipated growth in AI capabilities but is also a result of the market’s recognition of their potential to lead the next technological revolution [1].

However, such rapid development is not without its challenges. The Bureau of Labor Statistics has faced funding cuts that have reduced the quality and quantity of inflation data collected, leading to a greater reliance on imputed data [1]. This has made inflation estimates less reliable, complicating the Federal Reserve’s decision-making process. While the latest jobs report showed a weaker-than-expected figure of 73,000 new jobs in July, the broader inflation picture remains complex, with core goods prices rising at a 3.75% annual rate [1].

The uncertainty is also reflected in the volatility of commodity markets. Copper futures, for example, fell 20% after it was announced that President Trump’s latest tariff measures would exclude metals [1]. This type of unpredictable policymaking creates challenges for businesses reliant on stable commodity prices [1].

At the same time, the overall level of tariffs in the US remains elevated, with a 10 percentage point gap between announced tariffs and those actually paid by importers [1]. This suggests that the full economic impact of the current tariff regime is yet to be felt [1].

As AI investment continues to outpace historical infrastructure booms—such as the railroad and internet telco eras—investors are betting on an AI-driven future where the hyperscalers allocate up to 6% of US GDP to AI development [1]. The market is not just reacting to the potential of AI; it is actively shaping the reality of AI through observation and investment [1].

Source:

[1] https://blockworks.co/news/observation-effect-market-charts