Unsupervised Capital

A Liquid Fund Dedicated to Bittensor

[Unsupervised Capital @UnsupervisedCap]

(τ, τ)

[Sami Kassab @Old_Samster]

Today I’m proud to announce Unsupervised Capital — a liquid token fund dedicated to Bittensor.

At a moment when conviction feels scarce across crypto, I’ve never felt more clarity or confidence about a single thesis: Bittensor.

I’m launching this with my longtime friend @bloomberg_seth as Partner and the backing of @CrucibleLabs as our anchor investor.

Feels surreal to write this. Four years ago I was working as an aerospace engineer, dreaming of managing capital. Today, that dream begins.

The name Unsupervised Capital is a nod to unsupervised learning — where algorithms find structure in raw data without labels or instructions. Like those algorithms, we don’t follow maps. We explore the edge, where others hesitate. That's where the best opportunities live.

Bittensor has created something unprecedented: an open market for intelligence. It turns models, data, and compute into tokenized assets—creating entirely new investment primitives. Old frameworks don’t apply here.

This is why we started Unsupervised:
> To play in the frontier
> To bet on what hasn’t been discovered yet
> To back builders creating the next layer of intelligence

This fund is my biggest bet to date. I’m all in—and I’ve never been more sure.

We just published the full announcement, including our thesis and strategy (in next tweet)

GLHF

[Sami Kassab @Old_Samster]

Decentralized compute networks have been around for years but never found real demand. We think this time is different

Open-source models have gotten dramatically better, leading to business booming for open model providers and cost pressure is now forcing them to optimize GPU spend

We break down why the conditions have finally aligned and why we’re backing @TargonCompute, an emerging AI compute network on Bittensor

[Sami Kassab @Old_Samster]

New piece that covers why we believe subnet tokens on Bittensor are undervalued

We also cover how the 'sum of subnet tokens' metric is being misused and share some insight on other macro indicators

TLDR: stop using FDV - its not the right metric to value subnets