One of the tasks of applied econometrics and statistics is the estimation of the conditional mean function of the assumed model. The available methods for estimating such a function and the estimation results depend mostly on the a priori assumptions about the population or process that generates the data. The aim of the research presented in this paper is to apply single-index semiparametric econometric model to measure investment risk on the Warsaw Stock Exchange (WSE). Such a model is characterised by some assumptions less restrictive than in the case of other parametric models for the conditional mean function, such as a linear model or a binary probit model. At the same time, a single-index model retains many of the desirable features of a linear model and a least squares method. The presented model was used to measure investment risk for 10 IT companies quoted on the WSE in the period from 2018 to 2021. The data came from the Bloomberg financial service. Rates of return of two stock market indices, WIG20 (Poland) and S&P 500 (USA), were adopted as risk factors. The results indicate that the Polish market determines the volatility of returns of the analysed companies to a much larger extent than is the case with the US market. Furthermore, semiparametric models proved more flexible than the parametric ones regarding theoretical assumptions, which in the event of a large inflow of information might facilitate making correct investment decisions.