Executive Summary
Morgan Stanley's Global Head of Thematic Research identifies a critical structural shift: compute demand will meaningfully exceed supply in 2026, even as software and hardware efficiency improves. This represents a variant perception against current market pricing that assumes supply will scale linearly with demand. The firm's thematic categories outperformed MSCI World by 16% and S&P 500 by 27% in 2025, validating their framework. The key insight is that AI use cases are multiplying faster than infrastructure can scale, creating a bottleneck that extends beyond semiconductors into power grid capacity. Energy demand in developed markets, flat for decades, is now reflecting upward due to AI infrastructure and data centers. This creates what Morgan Stanley calls 'the politics of energy' - policymakers must prioritize low-cost, reliable power without destabilizing grids or increasing household bills. The convergence of these themes creates non-linear effects: AI accelerates energy demand, energy costs shape politics, and politics influence supply chains. However, heavy insider selling at NVDA ($329M in 90 days) suggests management may not share the same conviction about sustained compute constraints. The thematic framework identifies four interconnected forces, but the compute constraint thesis faces validation challenges given current semiconductor inventory cycles and efficiency gains.
Key Insights
what The Hosts said“compute demand is likely to exceed supply, meaningfully, even a software and hardware become more efficient”
This is a preview. Log in to see the full analysis including investment opportunities, risks, catalysts, and detailed insights.