The CEO-ROI Paradox of Artificial Intelligence
Davos 2026 confirmed a structural condition: AI has become a CEO-owned strategic variable while, in aggregate, failing to deliver stable financial returns. The mismatch sits between capital deployment, institutional absorption and infrastructure throughput.
"AI is no longer a discretionary innovation initiative. It is a general-purpose platform transition with asymmetric downside risk."
The signal from Davos
The 2026 discourse confirms a persistent misalignment between the rate of capital deployment, the institutional capacity for organisational absorption, and the physical and regulatory constraints governing AI scalability. Training a single frontier model now requires 10²³-10²⁵ floating-point operations, a 10,000x increase over models from the early 2010s.
AI has moved to the CEO desk
- →72% of CEOs now identify themselves as the primary decision-maker on AI strategy - a 2x year-on-year increase (BCG AI Radar 2026).
- →50% say their own tenure depends on getting AI right by 2026.
- →94% will keep investing in AI even without immediate returns.
- →Corporate AI spend is set to rise from 0.8% to 1.7% of annual revenues in one planning cycle.
Capital is outrunning absorption
Across 4,454 CEOs in 95 countries, 56% report AI has produced neither revenue growth nor cost reductions in the past 12 months. Only 12% report achieving both. Global corporate AI spend is on track to exceed $300bn in 2026, up from ~$120bn in 2022 (McKinsey; IDC).
This is not anomaly. It is the classic early-diffusion pattern of a general-purpose technology: deployment outpaces integration, and integration outpaces financial performance.
Infrastructure and legitimacy as binding constraints
Global data-centre electricity demand is projected to grow 15-20% per year through 2030 (IEA). Satya Nadella framed the risk directly: AI can lose "social permission" if energy intensity is not matched by visible societal value. Competitive advantage is now about throughput - the ability to convert capital into operational systems within physical and regulatory limits.
What this means for operators and allocators
The CEO-ROI gap is structural, not cyclical. Governance centralisation will precede ROI stabilisation. The winners in the next 24 months will be those who close the integration gap and secure throughput - not those who simply deploy more capital.
The full brief - with all 20 Davos 2026 AI numbers and sources - is published on LinkedIn.
The CEO-ROI Paradox of Artificial Intelligence - full analysis on LinkedIn
Includes the Top 20 AI numbers from Davos 2026 with sources. Subscribe to The Game Changer Intelligence to get every brief the moment it is published.