Is AI a Bubble, or Are Today’s Valuations Justified?

The rapid acceleration of artificial intelligence has made it the defining investment theme of the current market cycle. Giants such as NVIDIA, Microsoft, and other AI-driven companies have collectively added trillions in market value, prompting debate about whether the AI boom represents sustainable growth or the early stages of a bubble.

Some analysts compare today’s environment to 1996–98, when the internet was in its early but rationally optimistic phase. Others see signs reminiscent of the 1999–2000 dot-com bubble. In reality, the truth lies somewhere in between.

Parallels With 1996–98: Why AI Optimism Appears Justified

1. AI Adoption Is Real and Growing

Unlike speculative technology cycles, demand for AI is supported by genuine adoption. Cloud providers are rapidly expanding data-centre capacity, enterprise software companies are embedding generative AI into their products, and GPU orders remain far above available supply. This reflects actual revenue growth, not just forward-looking hype.

NVIDIA’s earnings, for example, have scaled at an extraordinary pace for a company of its size – a clear sign that AI spending is translating into real financial performance.

2. The Market Leaders Are Profitable and Established

This AI cycle is driven by companies with resilient business models, strong cash generation, and well-established customer bases. This differs sharply from the dot-com era, where many high-growth names had minimal revenue and no sustainable path to profitability.

Tech giants today also have the balance-sheet strength to finance long-term AI development, reducing the risk associated with early-stage innovation.

3. Early Productivity Gains Are Already Visible

AI is already delivering measurable productivity improvements, particularly in software development, customer support, marketing, and operational efficiency. These benefits emerged far earlier in the AI cycle than they did during the early years of the internet, reinforcing the belief that long-term earnings potential can justify elevated valuations.

Taken together, these factors make today’s AI boom similar to the fundamentals-driven optimism of 1996–98.

Parallels With 1999–2000: Signs of Speculation at the Edges of the AI Market

1. Smaller AI Firms Show Bubble-Like Behaviour

While the largest AI companies justify investor enthusiasm, many smaller, newly public or early-stage AI businesses appear far riskier. Some have minimal revenue, unproven business models, and unclear paths to profitability — yet their share prices have surged simply due to their association with artificial intelligence.

This mirrors the late 1990s, when countless dot-com companies soared in value despite weak fundamentals.

2. Risk of Excess Capacity and Over-Investment

Tech giants are spending enormous sums on AI chips, data-centre infrastructure, and model development. If enterprise or consumer demand grows more slowly than expected, the sector could face a period of excess capacity — a dynamic that also unfolded in the late 1990s as telecoms firms overbuilt internet infrastructure.

3. Market Concentration Increases Vulnerability

A small group of AI-heavyweight companies is responsible for the majority of stock market gains. This level of concentration has historical parallels with the dominance of Cisco, Intel, and Microsoft in the late 1990s. High concentration increases downside risk: a single disappointing earnings report from a major AI leader could trigger broader market volatility.

These elements suggest that while the core of the AI market is strong, speculative excess is building at the margins, much like in 1999–2000.

Conclusion: AI Is Not a Classic Bubble — But Parts of the Market Are Overheating

The AI market today is a blend of both rational growth and speculative enthusiasm:

• At the core, leading AI firms are profitable, fast-growing, and selling products with clear commercial demand. This environment closely resembles the rational optimism of 1996–98.
• At the edges, high-valuation startups and less-established AI companies show signs of over-excitement, echoing 1999–2000 dynamics.

Overall, AI valuations are partially justified, but the market is increasingly uneven. The most likely risk is not a catastrophic collapse across the entire sector, but a selective correction in the more speculative areas if expectations get too far ahead of reality.