GenAI can become a game changer for the automotive industry. One of the most promising application areas is the generation of code. Stunning results have been observed, and various teams are striving to apply GenAI in development processes. The potential for efficiency improvement is huge. How does this fit into Automotive development where software is developed based on SW requirements, SW architecture and SW detailed design? How can requirements and architectural design be systematically brought into the code generation process using prompts? After the evaluation of several options, including requirements and architecture via ‘In Context Learning’ seems to be the most promising approach. It requires advancements in the transformation of architectural specifications and requirements into a pure text format which can then be processed by Large Language Models. However, the transformation into pure text still must be facilitated and we can expect realistic projects to reach a size of more than 100,000 words. This is at the limits of even the most powerful Large Language Models today. The paper concludes with a discussion of the effects on positioning of the Automotive industry in the value creation chain.