Fernando de Frutos, Chief Investment Officer of Mora Capital Group, publishes an in-depth analysis in Funds Society on how to evaluate opportunities and risks in the technological stack of artificial intelligence.
Artificial intelligence is not a sector. It's a value chain with eight distinct layers, each with different business models, monetization cycles, and risk profiles. From frontier models at the core to physical enablers at the periphery (data centers, cooling, energy), opportunities are not distributed evenly. Understanding this structure is key to navigating current market valuations.
THE AI STACK: EIGHT LAYERS WITH DIFFERENT FUNCTIONS
At the center lies inference: models used in workflows and learning. Surrounding it are cloud platforms that distribute AI and finance much of the investment. Then come semiconductors, computing, and finally the physical enablers: data centers, cooling, electrical equipment, and energy generation.
The inner layers are closer to monetization and customer relationships. The outer layers may benefit from bottlenecks, but their economic profile tends to be more cyclical. The key question is not only how to position within AI, but what exposure each layer offers and what the market is paying for.
OPTIONALITY: THE OVERLOOKED COMPONENT
A stock exposed to AI can be viewed as the sum of two elements: cash flows and an option on future benefits linked to AI. The cash flow component is what the business can generate without heroic assumptions. The option is what investors pay for potential market expansion, margin elevation, or shifts in growth profile.
Many AI stocks behave like convertible bonds on future AI benefits: part cash flow, part participation in a higher-potential scenario. The best version is a solid business with added AI potential. The most fragile is a valuation that only works if that option materializes quickly.
THE MAIN RISKS: CONCENTRATION, CIRCULARITY, AND THE "WALL"
The first risk is concentration. AI spending depends on a small number of large buyers. Their scale is a strength, but it also creates dependency: if they change budgets or priorities, the entire value chain feels it.
The second is circularity. Parts of the ecosystem sell to each other, finance each other, or validate their own growth hypotheses. The chain becomes fragile when the loop depends more on confidence than on actual returns for the end user.
The third is technological: the "wall." Will frontier models continue to improve at a pace that justifies the scale of investment? If they begin to exhaust the most accessible parts of human-generated knowledge, and reinforcement learning and synthetic data fail to provide sufficient additional momentum, artificial general intelligence remains out of reach. Models would improve, but with smaller gains and larger costs. AI would remain useful, but the curve would flatten.
POSITIONING BY LAYERS
Different levels serve different functions. Central layers may offer better downside protection when AI optionality rests on diversified businesses and healthy cash flows. Intermediate layers offer more direct leverage to the investment cycle, but with greater cyclicality. Peripheral parts may benefit from bottlenecks, but should not be confused with permanent value capture.
The right question is not "Is this an AI stock?" but rather: How much of its valuation is explained by cash flows, how much by its value as an option, and what needs to go right for both elements to be true?
ABOUT THE AUTHOR
Fernando de Frutos - Chief Investment Officer, Mora Capital Management
Fernando de Frutos is responsible for developing the firm's investment policies and their application in designing personalized investment strategies and tactics. With extensive knowledge of investments, he also oversees discretionary portfolio management. Before joining Mora Capital Management Zurich, Fernando was Executive Director at Deloitte Switzerland's Strategic Consulting division, focused on the financial services industry. Previously, he led the Investment Products & Financial Advisory Department at ABN AMRO Private Banking.
He holds a PhD in International Relations and Economic Policy from the University of St Gallen, is a CFA charterholder, and holds an MBA from IMD.
Read the full analysis on Funds Society 👇