Is There An AI Bubble?

3 MIN READ

By RANI URBIS
www.nordis.net

A bubble in finance is when stock prices are driven well above what companies actually earn in profits because investors are betting on future growth.

In simple terms, a financial bubble exists when investor expectations about future profits become disconnected from observable earnings and current cash flow.

You can see this pattern clearly in past episodes, like the dot-com bubble of the late 1990s. During that period, stock prices surged far faster than company profits. The NASDAQ Composite rose dramatically into early 2000 and then fell nearly 78% from its March 2000 peak to its 2002 low. At that time, valuations had run too far ahead of profits. 

The current debate around AI follows that same structure: are markets pricing AI too aggressively?

Real numbers in the AI expansion

There is no question that AI is generating real revenue and driving real capital investment. Recent results through 2025 and into 2026 show continued growth, but also rising expectations built into valuations.

In 2024, NVIDIA reported $60.9 billion in revenue, more than double its 2023 total, driven largely by demand for AI chips. In 2025, quarterly results continued to grow strongly, with second-quarter revenue of $46.7 billion and third-quarter revenue of $57.0 billion, both up significantly year-over-year.

For the full year 2025, Meta reported $200.97 billion in revenue, a 22% year-over-year increase, and $59.89 billion in Q4 revenue alone. Meta’s capital expenditures for 2025 totaled $72.22 billion, reflecting significant investments in AI infrastructure.

Microsoft continues double-digit revenue growth driven by AI and cloud services throughout 2025. Analysts expect continued strength in Azure and cloud segments, as AI contributes a meaningful share of growth.

Across 2026, Alphabet (Google), Microsoft, Amazon, and Meta are on track to spend $650 billion+ collectively on AI and data center infrastructure.

Large rounds and rapid valuation increases are also observable among emerging AI players:

1. Cohere, a Canadian AI startup, raised funding rounds in 2025, pushing its valuation to around $7 billion, with annualized revenue growth rising from millions to tens of millions over the year.
2. Anysphere (creator of Cursor), another emerging AI startup, crossed $100 million in Annual Recurring Revenue (ARR) by mid-2025, with a valuation approaching $10 billion. By November 2025, its value had surged to $29.3 billion after raising $2.3 billion in funding.
3. CoreWeave, a cloud provider focused on graphics processing units (GPUs), raised over $1.5 billion in its 2025 Initial Public Offering (IPO) for its shares/stocks and secured $2 billion more from NVIDIA in 2026 to expand infrastructure.

Why these numbers matter

The revenue figures and spending commitments from 2025 onward confirm that AI demand is real and measurable. Companies are earning profits and reinvesting them in infrastructure and services.

But several data points indicate that a bubble risk is building if current trends continue.

1. Valuation expansion faster than revenue growth. NVIDIA’s revenue more than doubled in one year. Yet its market capitalization expanded by trillions of dollars in a short period. The magnitude of valuation expansion reflects expectations of sustained dominance and unusually high margins for years ahead.
   
   If revenue growth slows while value assumptions remain unchanged, repricing becomes likely.
2. Elevated market multiples. The S&P 500 forward price-to-earnings ratio has traded around 21–23 times expected earnings, above its long-term historical average of roughly 16–18 times.
   
   Higher multiples mean investors are paying a premium today for future growth. If projected earnings do not materialize at the expected pace, those multiples compress.
3. High market concentration. By 2024, the largest companies — many tied closely to AI — accounted for over 30% of total S&P 500 market capitalization, levels near historical highs.
   
   When index performance depends heavily on a few AI-linked firms, the system becomes sensitive to any earnings disappointment from those companies. Concentration amplifies risk.
4. Massive Forward-Looking Capital Commitments. Meta’s estimated $35–40 billion in capital expenditure for 2024 and Microsoft’s anticipated over $50 billion highlight aggressive infrastructure growth driven by expectations of continued demand.
   
   If enterprise AI spending plateaus or pricing competition increases, returns on that capital could fall short of expectations. The larger the capital commitment, the higher the sensitivity to demand shifts.

Emerging pattern

The data shows:

1.  Forward P/E ratios above historical averages 
2.  Top 10 stocks represent 39–40% of index weight 
3. Record-scale AI infrastructure spending 
4. Trillion-dollar valuation expansion concentrated in a few firms

Some can easily argue that these factors can be justified. Some can reason that revenue growth is strong, infrastructure demand is real, and AI adoption is measurable.

Collectively, they indicate that market expectations remain elevated and heavily dependent on continued AI-driven earnings growth.

That is where bubble risk resides. Revenue is real. Infrastructure expansion is measurable. But there is a clear gap between tangible revenue today and the valuation based on future expectations. If that gap continues to widen—if stock prices grow faster than profits for an extended period—the bubble burst will not be far off.#nordis.net