Artificial Intelligence: A Lasting Revolution or a Speculative Bubble?

The recent explosion in valuations of companies linked to artificial intelligence is stirring debate in financial markets. Some fear a repeat of the late 1990s Internet bubble, while others, including experts at Citigroup, Goldman Sachs, and Morgan Stanley, see it as a structural bull market with long-term growth potential.

Unprecedented Concentration in Indexes

AI-related companies now make up a significant portion of the S&P 500, generating about 35% of the index's total profits and nearly half its market capitalization. This concentration is largely centered around what analysts now call the "Magnificent Seven": NVIDIA, Microsoft, Apple, Alphabet, Amazon, Meta, and Tesla.

NVIDIA perfectly illustrates this trend with a valuation now exceeding $1.8 trillion, driven by explosive demand for its H100 and A100 GPU chips, which have become essential for artificial intelligence data centers. Microsoft, boasting a market cap of over $2.8 trillion, benefits from its strategic partnership with OpenAI and the growth of its Azure AI services. This concentration has made major stock indices heavily reliant on the performance of these tech giants, with the NASDAQ-100 now having a technology weighting above 50%.

Fundamentals Holding Up to Scrutiny

Contrary to a purely speculative bubble, Citigroup analysts note that these companies' valuations remain somewhat consistent when examining their fundamentals. The PEG ratio, which measures valuation relative to expected growth, is often below the market average for industry leaders. NVIDIA has a PEG of 0.8, Microsoft 1.2, and Alphabet 0.9, compared to an S&P 500 average of 1.4.

Return-on-equity ratios also highlight the exceptional performance of these businesses. Apple shows an ROE of 147%, Microsoft 36%, and NVIDIA 115%, strikingly higher than the S&P 500 average of 18%. Operating margins are equally impressive, with Microsoft Azure AI generating margins of 42%, Google Cloud AI 35%, and NVIDIA's Data Center division reaching 73%.

Goldman Sachs points to the Sharpe ratio of leading AI stocks, standing at 1.8 for a weighted portfolio versus 1.2 for the traditional S&P 500, suggesting better risk-adjusted performance.

The Shift to Widespread Adoption

McKinsey & Company’s research teams identify a crucial shift in the AI market, moving from an innovation phase dominated by technology creators to one of widespread adoption by traditional companies. This transition should allow a wide range of sectors to boost productivity through artificial intelligence tools.

In finance, JPMorgan Chase has announced $150 million in annual savings thanks to AI solutions, while Visa now uses artificial intelligence to detect fraud in real time with 99.9% accuracy. The healthcare sector is also advancing, with Johnson & Johnson claiming a 30% acceleration in new drug development via machine learning algorithms.

Traditional industry is following suit, with General Electric deploying predictive maintenance solutions for its turbines and Caterpillar developing autonomous management systems for its construction equipment. This steady diffusion expands the investment landscape beyond just tech giants to more diverse players.

Investments of Historic Proportions

To support this growth, major tech companies are investing sums that exceed certain key public budgets. Microsoft is dedicating $50 billion to AI data centers, Google is investing $48 billion in its cloud infrastructure, Amazon is committing $75 billion over fifteen years to AWS, and Meta is allocating $37 billion in 2024, mainly for AI and the metaverse.

These investments now represent 2.1% of U.S. GDP in AI-related direct expenditures and have led to the creation of 3.2 million jobs in the broader tech sector, according to Bureau of Labor Statistics estimates. The impact is also felt among semiconductor suppliers, who have received $280 billion in orders related to AI infrastructure.

The geography of these investments is shaping a new economic map. While Silicon Valley still accounts for 45% of global AI investments, Texas is emerging as a major hub with new megadata centers in Austin and Dallas. Internationally, Singapore stands out as Asia's nerve center, while Ireland hosts Microsoft and Google’s European hubs.

Persistent Red Flags

Despite this overall optimistic picture, several financial institutions express reservations. Some AI-focused startups have valuations that raise eyebrows: OpenAI is valued at $86 billion despite only $2 billion in revenues; Anthropic reaches a $15 billion valuation with an uncertain business model. These figures recall the excesses of the Internet bubble, when some tech firms had price/earnings ratios above 200.

Environmental impact is another major challenge. An AI data center consumes between 30 and 50 megawatts—the equivalent of a city of 50,000 inhabitants. ChatGPT alone consumes 564 MWh per day, and training a model like GPT-4 requires 1,287 MWh, equal to 550 tons of CO2. The International Energy Agency estimates that AI data centers could account for 4% of global emissions by 2030.

The regulatory framework is also evolving rapidly. The European Union has implemented the AI Act, imposing fines of up to 7% of global revenue for companies breaking the new rules. In the U.S., the Executive Order signed by President Biden in October 2023 requires tighter oversight of the most powerful AI models, while China is developing its own regulatory approach, focusing on training data control.

A Structural Transformation Underway

According to Citigroup analysts, artificial intelligence represents a transformation comparable to the rise of the Internet, but with stronger economic fundamentals. Unlike the 2000 bubble, leading sector companies generate substantial profits and maintain defensible competitive positions due to massive investments in research and development.

Morgan Stanley teams highlight that this tech revolution comes with a transformation of traditional business models. The adoption of AI by conventional companies should result in significant productivity gains over the coming years, thereby justifying current valuations through long-term earnings growth.

The question is no longer whether artificial intelligence will transform the global economy, but how quickly this transformation will occur, and who will be the winners and losers of this technological revolution. In this climate of uncertainty, vigilance is essential to analyze each company's fundamentals and distinguish durable opportunities from fleeting hypes.