Cyanobacteria have promising potential as sustainable cell factories. However, one challenge that is still largely unreported in scaling-up cyanobacteria bioproduction is phenotypic instability, where the emergence and selection of nonproducing cells leading to loss in production has longer evolutionary timescales to take place in industrial-scale bioreactors. Quantifying phenotypic instability early on in strain development allows researchers to make informed decisions on whether to proceed with scalable designs, or if present, devise countermeasures to reduce instability. One particularly effective strategy to mitigate instability is the use of genome-scale metabolic models to design growth-coupled production strains. In silico studies have predicted that creating certain cofactor imbalances or removing recycling reactions in cyanobacteria can be exploited to stably produce a wide variety of metabolites.