The article discusses the problem of assessing the investment attractiveness of risk projects for developing artificial intelligence, the methods of such assessment and their features. It is shown that due to the lack of relevant statistical, financial, operational information, the models and methods of investment valuation are, for the most part, subjective. The use of only one model or method of assessing investment attractiveness in the field of the development of artificial intelligence projects is insufficient, while the complex use without taking into account systemic aspects is likewise not sufficiently substantiated. To solve the existing problem, it is proposed to comprehensively use the available capabilities of the method of functional cost analysis (FSA), the essence of which is that the development project is decomposed into separate functions, and the necessary resources are measured and fixed for each function. Analysis of the functions of the object and the costs of the implementation of the functions makes it possible to identify the most acceptable variant of the object from the position of its functional content. At the same time, the article considers the possibility of using the functional-cost analysis method in the evaluation, the essence of which is that the development project is decomposed into separate functions, and for each function, all necessary resources are measured and fixed. An analysis of the object’s functions and their costs will help to identify the most economical version of a risky investment project from its functional content. It is reasonably noted that the main resources to support and promote the development of innovative projects are venture companies that invest considerable funds both at the initial stages and at the stages of development and expansion of projects. The amount of financial resources coming from business angels, crowdfunding and business accelerators is much smaller and goes mainly to the initial stages of project implementation.