The disconnect between escalating AI capital expenditure and tangible EBITDA margin expansion represents a critical fiduciary risk for boards overseeing digital transformation portfolios.
IMD Business School’s analysis, led by Karl Schmedders, highlights a pervasive inefficiency in how enterprises deploy artificial intelligence, noting that while spending is ubiquitous, value capture remains elusive. From a cost structure & efficiency perspective, the primary driver of this gap is the misalignment between technical implementation and operational unit economics. Companies are often treating AI as a pure capex event, ignoring the substantial opex implications of model maintenance, data hygiene, and integration. This approach inflates the run-rate of the technology stack without delivering the promised process improvement or lean operational gains, thereby compressing gross margin rather than expanding it.
The concept of "vibe coding" and rapid prototyping, while accelerating development cycles, introduces significant enterprise risk and governance blind spots. Without rigorous AI governance frameworks, organizations face heightened exposure to cyber risk and compliance failures, which can erode credit rating stability and increase liquidity costs. For the CFO, this translates to a distorted view of free cash flow, as the payback period for these initiatives extends well beyond initial projections. The ROI calculation is further complicated by the lack of standardized metrics for productivity gain, making it difficult to distinguish between genuine value creation and mere cost shifting within the P&L.
From a capital allocation standpoint, the current trend suggests a market mispricing of AI-driven growth. Investors are applying multiple expansion to firms with high AI spend, yet the underlying NPV of these projects often fails to materialize due to integration friction and skill gaps. This divergence creates a valuation trap where market sizing assumptions outpace actual operating margin realization. IMD’s findings indicate that the hidden costs of AI—specifically the labor hours required to manage and validate outputs—often negate the initial efficiency gains, resulting in a net negative impact on EPS in the short-to-medium term.
The strategic implication is that AI is currently functioning more as a cost center than a profit driver for the majority of enterprises. The trade-off between speed of deployment and fiduciary prudence is becoming a central board-level debate. As AI maturity varies across the portfolio, the risk of competitive position erosion grows for firms that over-invest without a clear path to margin expansion. The broader context suggests that the next phase of AI adoption will be defined not by technological capability, but by the discipline of capital efficiency and the ability to integrate AI into the core supply chain and logistics optimization processes where OTIF and inventory optimization directly impact the bottom line. Whether this shift from experimentation to industrialization can occur without a significant correction in market valuations remains an open question, but the pressure on governance to enforce cost discipline is now undeniable.