This work were requested by Uddeholms AB. Uddeholm is the worldś leading manufacturer of high performance tool steel for industrial tools. The company exists worldwide and is present in over 100 countries. One of the company’s main production processes is manufacturing powder steels with extremely high quality. In this work, one of these steels has been analyzed to evaluate the presence and distribution of non-metallic inclusions. Higher demands on quality and more global competition worldwide means that you always have to strive towards perfection in the manufacturing processes. Non-metallic inclusions have severe effects on the mechanical properties of steels. That's why it's of utmost importance to investigate their presence in the tool steel and especially their size. The reason for this is because it's the largest inclusions that's the most dangerous for the material. To obtain a statistically number of certainty of the largest of inclusions, very vast areas of steel have to be examined. Therefore a theoretical method of extrapolation is often used instead to approximate the distribution of the largest inclusions. The precision of this method is still very uncertain. Different methods of this kind of analysis do exist. The method applied in this work of degree is extreme values analysis by statistics of extreme values (SEV). To ensure the pre-conditions of the extreme values analysis, a large quantity of powder steel was analyzed. This was done by light-optic microscopy (LOM) and exclusively performed by Uddeholms AB at their R&D-facility (Research and development-facility) in Hagfors. At Karlstad Universityś test-facility specimens were subjected to ultrasonic fatigue testing at 20 kHz. The specimens derived from the same material as the ones examined in LOM. The result from the ultrasonic fatigue testing were that the largest inclusions in the material were found. This was done by applying very high cycle fatigue (VHCF). To calculate the maximum theoretical inclusion size, SEV was used. The extreme values analysis was performed on gathered data from both LOM and VHCF. All fracture surfaces from the ultrasonic fatigue testing were then examined in scanning electron microscope (SEM) at Karlstad University. In SEM, the largest inclusion in every fractured surface was identified and measured. The maximum real inclusion size from the fractured surfaces was then compared to the maximum theoretically calculated inclusion size from the extreme values analysis. As it turned out the real inclusion size proved to be slightly larger than the theoretical. The difference between them was found to be 3,25 µm. Conclusions drawn were that Uddeholms powder steel exhibits very high purity and that extreme values analysis as an analytical method is recommended. However, the analysis should be repeated to underline the chosen solution methodology.