Emerging PV technologies cells have a continuous strong track record in performance during the last years. With these performance values, solution processed emerging photovoltaic technologies are reaching out to applications that are going beyond the typical niche markets. The first generation of commercially available printed PV modules showed a lifespan in the order of beyond 5 years and more under outdoor conditions (OPV) while long-time outdoor data for perovskite modules are still missing. Interestingly, several experiments are strongly suggesting that solution processed semiconductors like organics or perovskites can be stable under light and, to some extent, under oxygen as well. Despite these impressive numbers, one should not forget that these are “best you can do” lifetime values. On the other hand, the community did not progress significantly in overcoming the fundamental limitations of printed PV. This is more expressed for organics than for perovskites: The energy gap law for excitonic materials, the precise microstructure control of binary or ternary composites, the design principles for environmentally stable materials or the Kirchhoff law for multi-junction cells continue to be major barriers for this technology. We briefly introduce into these long-time challenges and then discuss concepts and strategies how to resolve them for excitonic absorbers. Among them, the development of a digital twin which has inverse predictive power is a most promising concept. “Solar FAU”, an alliance of research partners in the Erlangen-Nürnberg region that is headed by Friedrich Alexander University, is exploring the basic concepts and methodologies how to build a digital twin for emerging-PV technologies. The central and most desired element of the digital twin is the power of inverse design, e.g., inventing molecules, device architectures and processes with tailored properties. Insight from first pieces (agents) of the digital twin strongly supports the assumption that inverse design capability is possible, even in the case of considerable experimental uncertainty. Coupling the digital twin to Material Acceleration Platforms (MAP) reduces experimental uncertainty and allows to learn predictions which otherwise would be impossible. We have recently demonstrated the power of such coupled systems and demonstrated correlations which were previously unthinkable, like the prediction of performance and lifetime of OPV cells from simple absorption data or the identification of he best process for perovskites under environmental conditions merely at the hand of photoluminescence data.