
Thesis Intelligence
Living Intelligence
Every investment thesis is a set of beliefs about the world. Most funds state theirs once and move on. We treat ours as a living system — six hypotheses, each continuously tested against incoming evidence, external research, and identified risks.
The graph below maps this system. Evidence from McKinsey, IBM, EQT, and other tier-1 sources flows in and either strengthens or challenges our hypotheses. Risks are tracked explicitly. Source links point to the original research. This is not a static diagram — it is an automated intelligence layer that evolves as new data arrives.
Core Hypotheses
What We Believe — and How We Test It
Six hypotheses form the backbone of the Ekohe Fund thesis. Each carries a confidence score reflecting the weight of supporting evidence against identified risks. We validate continuously — not once at fundraise, but throughout the fund lifecycle.
AI Transforms SMB Economics
AI produces measurable EBITDA expansion, cost compression, and revenue growth in operationally manual SMBs within 12–24 months.
Japan SMBs Are Mispriced
Japan's SMBs trade at 3–6x EBITDA reflecting pre-AI productivity. Low adoption + succession crisis = exploitable valuation gap.
Deep Deal Supply
Enough quality targets in Japan at 3–6x EBITDA to fill a $75M fund. Supply is demographic (persistent), not cyclical.
Exits at Transformed Multiples
AI-transformed Japanese SMBs exit at 10–15x EBITDA via strategic sales, secondary PE, TSE Growth Market IPOs.
AI Capability Accelerates
Each AI generation makes transformation faster and cheaper. Playbooks compound. The window for pre-AI valuations is closing.
GP-as-Operator Edge
Vertical integration of GP + AI transformation capability eliminates 6–12 month integration lag. Ekohe already builds operational AI for PE and other institutional clients.
Knowledge Graph
Explore the Evidence Network
Click any node to inspect it. Hover to highlight connections. Drag to rearrange. Filter by type using the toolbar. Source and entity nodes link to original references.