Investment Management
REFRACTOR
Firm
Refractor is a private investment firm in Rye, New York. We do original research on technology companies, at whatever stage we find them, and act on very little of it.
We study a small number of companies.
We understand them completely.
We stay with them for years.
That is the entire idea.
Depth over breadth
We would rather understand a handful of companies better than anyone else than hold opinions on a hundred. Depth is what makes a judgement defensible when the consensus moves against it, and it is the whole difference between conviction and a guess.
We build our own instruments
Our research stack was written in house — ingestion, normalization, screening, and the models we underwrite on. Nothing about how we reach a conclusion arrives in a vendor's default configuration, which is why our answers are rarely the same as everyone else's.
Time is the edge
Most capital is compensated on a horizon measured in quarters. Our advantage comes from underwriting outcomes that resolve over years, and from being structurally willing to wait for them.
Sectors
Technology broadly, with genuine depth in the areas where our research compounds and where the layers depend on one another.
Artificial Intelligence
The Intelligence Layer
Model developers, the silicon they train on, and the software businesses whose economics are being rewritten by inference. We study where durable margin accrues once the capability stops being scarce.
Fintech & Financials
The Capital Layer
Exchanges, brokerages, payment networks, banks, and asset managers. Balance sheets and take rates are knowable things — this is where our diligence compounds fastest.
Energy
The Power Layer
Generation, nuclear, storage, and grid. Compute is a claim on electricity. We follow the constraint upstream to the businesses that own it.
Infrastructure
The Physical Layer
Data centers, networking, cooling, industrial systems, and the unglamorous physical assets every software narrative eventually depends on.
Digital Assets
The Programmable Layer
Settlement, custody, and the market structure being rebuilt on programmable rails. We hold it to the same research standard as everything else, and we are unhurried about it.
Approach
How the work actually gets done, and what has to be true before we are willing to act on it.
Underwrite the downside
Every piece of work begins with what would have to be true for a company to be worth nothing. If we cannot describe how we lose, we have not finished the research.
Wait for the work to justify it
Conviction is the only thing that moves us, and it is rare. Most of what we study we never act on, and we are comfortable doing nothing for long stretches.
Follow the constraint
The interesting question is usually one layer away from the obvious one — the input a business depends on, the power it consumes, the network it settles across. We keep following it until the picture is complete.
Team
The firm stays small on purpose. Software does the work that headcount would do elsewhere.
Alec Wilson
Alec worked across Digital Currency Group, BlackRock, and Rockefeller Capital Management before starting the firm — digital assets, institutional asset management, and private wealth. The common thread is the capital layer, which is where Refractor's research runs deepest and where technology and financial infrastructure meet.
We would rather build a system once than staff a process forever. The firm scales through tooling, which is why it stays small — and why the people here spend their time on research rather than on running it.
Contact
We keep a narrow correspondence and answer everything we are sent.
Correspondence
We are easier to reachthan we are to find.
If something here is relevant to what you work on, or you disagree with how we have framed it, write to us — briefly, and tell us how you found us.
Write to us