Signal discovery
Explore market structure, alternative data and behavioural patterns while controlling for leakage, instability and overfitting.
Platform / Overview
Markets. Compute.
Digital rails.
Three investment engines connected by one research discipline. We build and back systems where intelligence can move from model to infrastructure to capital decision.
01 / MARKETSAI quantitative
AI quantitative
trading systems
We treat trading as an applied research environment: hypotheses are expressed as models, tested against uncertainty and evaluated through controlled feedback.
Systematic construction
Translate multiple uncertain signals into portfolios designed around liquidity, concentration, correlation and tail behaviour.
Measured deployment
Connect models to execution through observability, explicit controls and post-trade learning rather than opaque automation.
02 / AI infrastructure
Compute is a full-stack investment.
Accelerators are only one part of productive AI capacity. We assess the connected system: power availability, site design, networking, storage, cooling, software orchestration and utilisation economics.
Our focus is infrastructure that turns scarce physical resources into reliable, measurable intelligence—not capacity purchased for its own sake.
Read the compute thesis ↗LAYER 05Workloads & utilisation
LAYER 04Orchestration & software
LAYER 03Accelerated compute
LAYER 02Network & storage
LAYER 01Power & thermal systems
03 / DIGITAL RAILSFrom tokens
From tokens
to useful systems
Issuance is easy. Durable utility is harder. Our token foundry thesis starts with the operating problem, the rights being represented and the infrastructure required for real use.
Rights before representation
Define what a token actually conveys—access, ownership, settlement, governance or another enforceable utility.
Programmable rails
Assess custody, identity, compliance, interoperability and lifecycle operations as one connected product system.
Utility before attention
Look for value creation grounded in lower coordination cost, better access or improved operational efficiency.
04 / OPERATING SYSTEMControls are part
Controls are part
of the architecture.
Risk, governance and review are not gates added at the end. They shape how a system is researched, funded, deployed and monitored from the beginning.
Explicit assumptions
Every decision begins with what must be true, what can be measured and what would invalidate the thesis.
Stage-gated capital
Exposure should grow with evidence, operational readiness and the ability to observe outcomes.
Independent challenge
Models and investment cases improve when failure modes are actively sought rather than quietly accepted.
Continuous postmortems
Unexpected outcomes become inputs to process improvement, whether or not they created a financial loss.
Explore the thinking
See the research
behind the system.
Read the questions shaping our work across markets, infrastructure and programmable assets.
Research library ↗