Artificial Intelligence – Bubble or Breakthrough?

What is a Bubble?

To assess whether artificial intelligence (AI) is experiencing a bubble, we first need to understand what a bubble actually is.

An asset bubble is an economic cycle in which the price of an asset, such as stocks, real estate, or commodities, rises rapidly and significantly beyond its intrinsic value. This unsustainable surge is typically driven by speculative demand, amplified by exaggerated market sentiment and herd-like investor behaviour. The bubble “bursts” when momentum shifts, leading to a sharp drop in prices and widespread financial losses.

Early optimism drives prices upward (start), speculative euphoria sustains peak valuations (middle), and reality triggers a sharp correction (end).

Historic Examples of Asset Bubbles:

Tulipmania (1630s): The earliest recorded asset bubble occurred in Holland, where tulip prices soared twentyfold between November 1636 and February 1637, only to collapse by May 1637.^[1]

Japan’s Economic Bubble (1980s): A 50% surge in the yen triggered a recession in 1986, prompting aggressive stimulus. Land and stock prices tripled within four years. At its peak, the Imperial Palace grounds were worth more than all the real estate in California.^[2] The bubble burst in 1991, ushering in Japan’s “Lost Decade.”^[3]

The Dot-Com Bubble (1990–2000): The internet boom led to massive speculation in tech startups. ^[4] The NASDAQ rose from 750 to over 5,000 by early 2000, before crashing by 78%. It took 15 years to reach a new high.^[5]

The U.S. Housing Bubble (1996–2008): Following the dot-com crash, investors turned to real estate. U.S. house prices doubled between 1996 and 2006, with two-thirds of the growth occurring between 2002 and 2006. Rising interest rates and falling prices triggered a collapse in mortgage-backed securities, leading to the worst global downturn since the Great Depression.

What Happens When a Bubble Bursts?

When a bubble bursts, prices collapse rapidly, leaving investors with steep losses and triggering panic selling. Confidence erodes, credit tightens, and businesses that relied on inflated valuations may struggle or fail. Unemployment rises as companies downsize, and consumer spending falls, deepening the economic slowdown.

Is AI a Bubble?

Now that we understand how bubbles form and burst, we can examine whether the current growth in AI and tech stocks fits the pattern.

In 2024, AI was responsible for over 60% of the S&P 500’s total gains,^[6] with companies like Nvidia, Microsoft, and Alphabet driving the bulk of market growth.^[7] Nvidia alone contributed nearly one-third of the index’s return, despite concerns that its valuation far outpaced its earnings.^[8]

This surge has been fuelled by massive capital inflows. Global spending on generative AI is expected to reach $644 billion in 2025, a 76.4% increase from the previous year.^[9] Yet many AI companies remain unprofitable: over 1,300 are valued above $100 million, and nearly 500 are unicorns worth over $1 billion.^[10]

The most striking example is OpenAI. According to reports, OpenAI generated $4.3 billion^[11] in revenue in the first half of 2025 but posted a net loss of $13.5 billion.^[12] It’s on track to lose $27 billion by year-end.^[13][14] Its valuation-to-revenue ratio is 24:1, compared to 35:1 for Nvidia and 14:1 for Microsoft.^[15][16]

These elevated ratios echo the Nifty Fifty era of the 1960s and 70s, when a group of high-growth US companies, like Xerox, IBM, and Polaroid, were considered “one-decision stocks” that investors could buy and hold forever. Many traded at P/E ratios above 50, justified by their perceived invincibility. But when market sentiment shifted in the mid-70s, even the strongest names saw sharp corrections.

A Bank of America survey found that 54% of investors already believe AI is in bubble territory. ^[17]The surge in AI mentions on earnings calls, the flood of capital into startups, and speculative investor behaviour all point to irrational exuberance^[18], a hallmark of bubbles.^[19]

But Is It Justified?

Despite these concerns, AI is not just hype, it’s delivering real-world impact. Goldman Sachs estimates that generative AI could raise global GDP by 7% over the next decade,^[20] while McKinsey projects it could add $4.4 trillion annually to the global economy.^[21] These gains support long-term investment and justify elevated valuations.

Companies like Nvidia aren’t just riding the wave, they’re selling the “picks and shovels” of the AI gold rush. Nvidia’s chips are essential for training and running large AI models, and its revenue has grown accordingly. This is infrastructure investment, not speculation.

Unlike the dot-com era, today’s AI leaders are profitable, diversified, and strategically positioned. Microsoft’s integration of OpenAI into Azure and Office 365 is already reshaping enterprise software.^[22]  These firms aren’t betting on unproven ideas, they’re embedding AI into products used by billions.

Final Thoughts:

Yes, AI is booming, but that doesn’t automatically make it a bubble. The difference lies in underlying value. Today’s AI leaders aren’t just promising the future; they’re already delivering it. While corrections may come, the long-term trajectory of AI appears grounded in real-world impact, not fantasy.