Proprietary investment research grounded in practitioner knowledge. Ground-level signal, not analyst inference. The intelligence product that feeds our capital decisions and our advisory clients.
Kandelhill Intelligence is our research division — not a traditional sell-side analyst operation. We do not publish daily notes or react to earnings calls. We do not chase consensus. We produce deep, practitioner-grounded research on the companies, protocols, and trends that matter most in AI infrastructure.
Our research is written for people who actually understand the technology. It is built on signal gathered from inside the ecosystem. It reflects what we are seeing in real time, not what Wall Street consensus believes should happen.
Every analyst can read a 10-K. Every analyst can listen to an earnings call. The signal that matters is what happens before that — when infrastructure is straining, when adoption curves inflect, when technical constraints force innovation. We see that signal first because we are inside the ecosystem.
Our team operates inside the AI infrastructure stack. We see real implementation challenges, adoption curves, and competitive dynamics firsthand. This is not theoretical analysis — it is grounded in hands-on experience.
We track adoption patterns not just in AI, but in adjacent infrastructure trends. We identify inflection points before they become obvious in the financial data. We see where the next constraints will emerge.
We focus on structural shifts in the market — new capacity limitations, protocol innovations, adoption curves — rather than trading events or earnings surprises. These shifts create the most valuable investment opportunities.
We develop theses over months and years, testing them against operating reality continuously. When we publish, we have conviction built on deep understanding, not recent news.
GPU allocation, foundry economics, and the global race for semiconductor capacity in AI.
Energy constraints limiting AI deployment, grid capacity, and the infrastructure race for power.
Interconnect architectures, networking protocols, and the bandwidth arms race in distributed training.
Facility operations, cooling technology, and the capital intensity of next-generation AI infrastructure.
Emerging technical solutions to fundamental constraints: memory, latency, scalability, efficiency.
Winner-take-most dynamics, moat development, and the emerging power structure of AI infrastructure.
It's grounded in reality. We are not theorizing about how AI infrastructure works. We operate inside it. We see constraints firsthand. We understand what actually breaks and why.
It's forward-looking. We do not write about what has already happened. We identify inflection points months or years before they become visible in financial metrics. This is where investment opportunities emerge.
It's proprietary. The signal we gather is not available through traditional research channels. It is not in SEC filings. It is not on earnings calls. It only comes from being inside the ecosystem.
It's actionable. Our research is written to inform capital allocation decisions. Every thesis connects directly to investment opportunities in companies, protocols, or emerging businesses.
If you want to discuss our research, access our proprietary findings, or explore partnership opportunities: