The messenger RNA lipid nanoparticle (mRNA-LNP) platform reached the clinic at extraordinary speed, yet it carried an inconvenient inheritance: a liquid drug product that survives only at deep sub-zero temperatures. The frozen chain that this demands, running from roughly -80 to -20 degrees Celsius, is precisely the infrastructure that low- and middle-income countries (LMICs) most often lack, and it converts a scientific triumph into a distribution failure at the last mile. This review argues that the instability is a formulation problem with a formulation solution, and that lyophilisation is the most mature route to a refrigerator-stable or ambient-stable mRNA vaccine. It examines why the aqueous particle degrades, working through the chemistry of RNA backbone hydrolysis and lipid ester cleavage that the resident water in the particle core sustains. It then treats the two levers that a formulator actually controls. The first is excipient selection, where disaccharides such as sucrose and trehalose vitrify the matrix and replace hydrogen bonds at the lipid headgroups, while bulking agents and a matched buffer set the collapse behaviour of the cake. The second is process design, where the freezing, primary drying and secondary drying stages must each be held below the critical temperature of the freeze concentrate to avoid collapse, aggregation and residual moisture that would defeat the exercise. Evidence that an optimised cycle can compress a multi-day process to under a day, hold particle size and encapsulation within narrow bounds, and confer refrigerated shelf lives measured in months is now consistent across several independent groups. The review closes on the distribution question, quantifying what a warm-stable presentation would mean for LMIC immunisation and setting out the manufacturing and regulatory work that still stands between the bench result and a licensed product.