III-N semiconductor materials possess superior physical properties, which make them very attractive for various optoelectronic and microelectronic applications. Despite their great features, such as the large band gap energy, the high critical electric field and the superior thermal and chemical stability, III-N materials have not been well-established at the semiconductor market, yet. A major challenge for the industrial growth of III-N structures, is the lack of economically reasonable substrate materials, which fulfil all significant requirements, such as electric and thermal conductivity, compatible crystal structure and negligible thermal and lattice mismatch. In the semiconductor industry, the mainly used substrate materials are SiC, Al2O3, and Si. Due to its attractive economical features combined with good electrical properties, Si substrates are very promising for the application in the microelectronic sector. The major drawbacks of Si substrates are the large lattice and thermal mismatches with respect to III N semiconductors. In order to compensate these mismatches and to fulfil all electronic requirements, III-N structures are usually grown as multilayer stacks. Although the overall stress is negligible, there are locally high sublayer stress concentrations, which have a significant influence on the reliability of the actual device. The stress profile of various high electron mobility transistors (HEMTs) deposited on Si have been characterized. Typically, a HEMT is based on AlxGa1-xN alloys and consists of a nucleation-, a transition-, a buffer- as well as a thin barrier-layer on top. For a comparative study, there have been applied several techniques based on (i) focused ion beam milling combined with digital image correlation (FIB-DIC) (ii) transmission electron microscopy (TEM), (iii) X-Ray diffraction, (iv) wafer curvature and (v) Raman spectroscopy. All analyses have indicated a high tensile stress in the nucleation layer, a compressive to tensile stress transition in the transition layer and tensile stress within the buffer layer. Subsequently, the stress profiles have been correlated with the sublayer microstructures using TEM. The comparison of different heterostructures has shown, that the transition layer design has a significant impact on the microstructure and the stress gradient across the multilayer stack. In a final step, the advantages and disadvantages of the individual techniques have been discussed.