Investors are understandably questioning whether the extraordinary capital being poured into artificial intelligence will generate adequate returns. A less obvious risk, however, is underestimating the immense computing power an economy increasingly run by AI agents and robots could consume. Unlike human counterparts, AI agents operate around the clock and interact continuously, potentially driving demand for computing infrastructure on an entirely different scale. This escalating need could be a boon for data centres as automation integrates into every aspect of economic activity. Many forecasts, including those from the technology sector, suggest productivity gains will ultimately justify today’s significant investments.
An example of this vast potential is the MatrAIx initiative from Harvard University researchers. This project seeks to create a model of 8.3 billion virtual personas, aiming to simulate and predict real-world outcomes, such as estimating product efficacy or market response. A smaller, one-million-persona model demonstrates meaningful predictive power, hinting at the immense commercial applications. The computing capacity required for such systems, especially if organisations test numerous future scenarios simultaneously or delegate tasks to autonomous agents, would be equally vast. This highlights automation’s inevitability, positioning computing power as critical infrastructure rather than a discretionary expense.
This paradigm shift strengthens the argument for continued investment in capacity across the AI value chain. Yet, history shows that transformational technologies rarely reward capital evenly. Not every early participant in a technological revolution achieves profitability. With global debt levels remaining high and interest costs rising, the traditional anchors of investing are already strained. Investors face the complex task of discerning where profits will ultimately accrue and identifying the specific companies poised to capture value from automation. The challenge lies in determining which businesses will become the bedrock for future innovation and which will simply fail to deliver adequate returns.
