AI: Why the Current Market Is Nothing Like the Dot-Com Bubble

Profits Growing Faster Than Stock Prices

The beginning of 2025 was marked by a sharp rebound in the US market: the S&P 500 surged by 36% since its low point on April 8th, bringing the annual performance to around +16%. However, this recovery was very concentrated: the AI giants—the « Magnificent Seven"—captured most of the movement. This concentration naturally fuels fears of a bubble. But Ramify highlights a crucial point: in 2025, the increases are justified by earnings, not by an irrational multiple as seen in 1999-2000.

The numbers are clear. Between September 30, 2021, and September 30, 2025, earnings in the S&P 500 tech sector jumped by 77%, with a sector performance of +121%. Stock prices are rising, but so are profits, and in a proportionate manner. During the internet bubble, the situation was the reverse: from June 30, 1996, to March 31, 2000, profits increased by roughly 79% while prices skyrocketed by +479%. The price-to-fundamentals disconnect was glaring.

Another structural element is the breakdown of overall performance. From December 31, 2024, to September 30, 2025, half of the MSCI USA's performance (+11.75%) came directly from earnings—more than 6%. In contrast, in the eurozone (+13%), less than 3% came from earnings; the bulk came from multiple expansion. In other words, the American stock market enthusiasm is grounded in real growth, not speculative frenzy.

This dynamic is explained by the business model of AI players: unlike the web pioneers who built before monetizing, hyperscalers are already generating revenue as they invest. Demand for cloud services is growing, productivity is improving (software development, advertising, business tools), and monetization is expanding among corporate clients. It is this operational strength that supports stock prices.

AI giants have nothing in common with dotcoms

The other fundamental difference from the year 2000 lies in the state of the balance sheets. The « Magnificent Seven » boast a net negative debt of -22%, which means they hold more cash than financial debt. They can finance their massive investments—such as purchasing GPUs, data centers, or cloud infrastructure—without heavily relying on borrowing. As a result, they are less vulnerable to a sudden tightening of financial conditions.

Their net margin, around 29%, reflects an extraordinary level of profitability. For comparison, the tech leaders during the dot-com bubble had a much more modest margin of 16% and a positive net debt of 4%. The companies dominating the market in 2025—Microsoft, Apple, Nvidia, Amazon, Alphabet, Meta, Tesla—are therefore more liquid, more profitable, and better capitalized than their predecessors from the « dotcom » era. In other words, the market is not paying for promises; it's paying for already tangible profitability.

Valuations should also be put into perspective. Despite their stock market performance, the P/E multiples of AI stocks have decreased in recent years, given the rapid growth in earnings. The anticipated P/E for the « Magnificent Seven » is around 26.81—a far cry from historical excesses. During the internet bubble, the average multiple reached 522, and Japan's bubble in the 1980s soared to 672. The levels in 2025, high but sustainable, remain rational when compared to the achieved profitability.

An external catalyst, not an internal overvaluation

For Ramify, the question isn't whether a bubble already exists, but rather which signals could cause it to emerge or burst. In any major phase of technological innovation, some investment decisions are perfectly rational... and others much less so. The challenge lies in identifying potential catalysts.

Certain macroeconomic factors—such as inflation, unemployment levels, Federal Reserve decisions, and earnings revisions—may trigger temporary corrections but not a generalized explosion. However, two elements are closely monitored:

1. A sudden monetary tightening. A rapid increase in interest rates would raise the cost of capital, limit corporate leverage, and potentially trigger forced sales. At this point, such a scenario remains unlikely: current trajectories suggest stabilization or even a decrease.

2. Tax pressure on financial assets. Given the high level of US federal debt, certain proposals to tax wealth or unrealized gains could force investors to sell positions to meet their obligations. This risk, though political and still uncertain, is considered more plausible.

Nonetheless, none of these triggers have materialized. The current market, boosted by AI, remains in an unusual configuration: high but supported valuations, a sector in real growth, and balance sheets of a strength rarely seen in the history of equity markets.