AI Compute

Sector thesis

AI Compute is the hardware and infrastructure that powers artificial intelligence—the chips, servers, and data centers that train and run AI models. Right now, it's one of the most capital-intensive sectors in tech because companies worldwide are racing to build AI capabilities, and that requires enormous amounts of specialized computing power. The megatrend is simple: AI has moved from research labs into production. Every major tech company, cloud provider, and enterprise is investing heavily in AI infrastructure. This isn't a temporary spike—it's a structural shift in how computing resources are allocated. The demand for chips and data center capacity to support AI is outpacing supply, which creates a multi-year tailwind for companies that build or operate this infrastructure. Within AI Compute, there are three main buckets: semiconductor makers (who design and manufacture the specialized chips used for AI), data center operators (who own and rent out the physical servers and cooling systems), and equipment suppliers (who build the networking and power systems that connect everything). Each has different economics and risks. The biggest risks are real. Chip design cycles are long and expensive—if a company bets wrong on what customers need, they can waste billions. Data center operators face rising electricity costs and environmental scrutiny. And there's always the risk that the current AI investment boom slows faster than expected, leaving excess capacity and lower prices. For a retail portfolio, this sector works as a growth holding if you have a 3-5 year horizon. Watch for signs of actual AI adoption (not just hype), electricity costs in key regions, and whether companies are actually using the infrastructure they're buying. Be honest about your conviction—this sector can be volatile, and it's easy to get caught up in the narrative.