AI research is EXAM INVESTMENT’s fifth business field. Our focus is on the relationship between technical capability, the resources needed to develop it and a credible route to practical use.
Research is not the same as revenue
A demonstration can show potential without proving that a system is reliable, economical or useful at scale. Assessment questions include what has actually been tested, under which conditions, against which baseline and with what failure modes. Claims about a model should be distinguished from evidence of customer adoption.
Teams, data and intellectual property
Important questions concern the team’s relevant capabilities, rights to training and evaluation data, software licences, ownership of research outputs and dependence on third-party models or infrastructure. A business built on external services must understand the terms and costs of that dependence.
Economics and financing
Compute, data preparation, evaluation and deployment can all require significant resources. A financing assessment should consider the next milestone, available capital, additional funding needs, ownership dilution and realistic exit constraints. Progress in AI as a sector does not guarantee a return from a particular company or financing structure.
Governance is part of the assessment
Safety, privacy, security, reliability and human oversight belong in the discussion of an AI system. NIST’s voluntary AI Risk Management Framework provides a public reference for considering trustworthiness across development and use. It is not an investment-rating framework or evidence that a company has been certified.
This page describes the field and our assessment perspective. It does not claim partnerships with particular laboratories, ownership of proprietary models or a fixed yield from AI research.