InfoHub Logo

The idea that Artificial Intelligence (AI) represents a universal 'rising tide' that lifts all boats in the investment world is a persistent narrative. However, a closer look at the data suggests this widely held belief is fundamentally flawed.

The AI Investment Divide: A Closer Look

Analyzing a portfolio of sixty publicly traded companies across the AI value chain in 2025, including semiconductor manufacturers, data center operators, and energy providers, reveals a stark contrast. The top performers saw gains exceeding 140%, while the bottom performers experienced losses between 10% and 30%. This significant disparity, with a gap of approximately 150 percentage points between the top and bottom quintiles within just twelve months, far exceeds historical market variations.

Factors Driving the Divergence

Several structural factors are contributing to this widening gap. Firstly, the immense capital expenditure required for AI infrastructure is a major differentiator. The four major US hyperscalers alone are projected to spend around $700 billion on capital expenditures by 2026, a figure more than double Kazakhstan's annual GDP. This massive investment is leading to unprecedented levels of depreciation, potentially creating significant holes in company balance sheets if AI revenue doesn't keep pace.

Secondly, access to electricity is becoming a critical bottleneck. Unfulfilled cloud power orders, not chip shortages, are now a primary concern for major tech companies. This shifts the economic landscape, favoring companies that control energy generation, particularly those involved in nuclear power, while penalizing those unable to secure timely power access.

Furthermore, the profitability model is skewed. While companies developing AI models may receive promises of future returns, the immediate economic benefits are flowing to the underlying infrastructure providers – GPU manufacturers, data center operators, and energy suppliers. This creates a structural mismatch where those below the model in the value chain are realizing tangible returns today, while model investors are still waiting.

Finally, corporate governance plays a crucial role. Companies that fund growth through stock-based compensation, diluting existing shareholders, are being distinguished from those generating free cash flow and buying back shares. The market's current inability to clearly differentiate these two approaches is a temporary blind spot that is expected to correct itself in the coming years, leading to further divergence in stock performance.

Implications for Investors

For most investors, particularly those in Kazakhstan accessing global markets through ETFs or diversified international portfolios, the current approach offers exposure to an average AI portfolio return. In a market with a 150-point dispersion, this average return is a risky proposition, as it guarantees holding both winners and losers whose performances may cancel each other out. Identifying the true winners and losers in the AI space is no longer optional; it is the primary challenge for long-term investors in developed equity markets.

Looking ahead, the market is still in the early stages of sorting through which entities will truly profit from the AI wave. By April 2028, it is anticipated that at least two of the current top ten AI companies will see their valuations drop by 40% or more, while at least two currently undervalued companies will double their market capitalization. This ongoing selection process will define the economic rent derived from this technological revolution.