Key Takeaways
- The AI infrastructure layer is still in early innings despite the surface-level saturation.
- Climate tech and infrastructure are creating investment opportunities that resemble early internet infrastructure.
- The founding teams that combine domain depth and AI fluency are the highest-priority targets.
Every year, Saim Abbasi and the Iron Key Capital team do a formal review of the fund's investment thesis against what is actually happening in the markets they track. The 2026 version of that review produced specific conclusions about where to focus and what the next 12 to 24 months are likely to look like for early-stage investing.
AI Infrastructure, Not AI Applications
The AI application layer, the products that use large language models to solve specific business problems, has become genuinely crowded. The infrastructure layer, the tools, platforms, and services required to deploy AI reliably in enterprise environments, is less crowded and better positioned to capture value over a longer cycle. Iron Key's current thesis prioritizes AI infrastructure over AI applications in most sectors.
Climate and Physical Infrastructure
The energy and infrastructure requirements of the AI economy are creating investment opportunities in sectors that have not traditionally been early-stage VC territory. Grid modernization, energy storage, data center cooling, and the software that optimizes physical infrastructure are all areas where Saim sees investment opportunities that are early relative to the commercial cycle they will ultimately address.
The Founder Profile That Matters Most
The specific founder characteristic Iron Key Capital is prioritizing in 2026 is the combination of deep domain expertise in a specific sector and genuine fluency with AI tools as building blocks. This combination, a healthcare operator who builds AI-native tools for clinical workflows, or a logistics professional who builds AI-optimized supply chain software, produces companies that are hard for purely technical or purely domain-focused teams to replicate. The domain knowledge creates the problem understanding. The AI fluency creates the solution.
"The most interesting opportunities are never in the most crowded conversations. They are one layer below the obvious."