Societal Impact Statement Large quantities of diverse native seeds are required to scale up global restoration efforts. However, it remains unclear for many ecosystems how the diversity of available seed in commercial stocks reflects the composition of the ecosystems where vegetation is being remade. This study highlights existing shortfalls in the diversity of seed presently available for use in restoration and identifies gaps in the seed supply chain while providing a new method for optimising species selection given these constraints. This work underscores how improved collaboration between stakeholders is required to strengthen the seed supply chain and help remake functionally diverse vegetation. Summary Restoration using native seed is frequently implemented to restore degraded ecosystems. However, it remains unclear how constraints on the diversity of germplasm available for use in restoration may limit the recovery of both species and plant functional diversity. Using a dataset of seed availability for Australia's major vegetation types, we explore variation in the diversity and breadth of functional traits (leaf mass per area, seed mass, plant height) for species where seed is available on commercial markets relative to unavailable. Using these data, we simulate theoretical seed mixes derived from two species pools: (1) constrained by the current market of commercially available seed, and (2) assuming all native species can be planted; then we compare differences in functional diversity (dispersion) as an exercise to explore possible limitations within the current seed supply. Seed from only 12% of Australian plant species (2992 species) is presently available for immediate purchase. Seed was more frequently available for trees and shrubs than for understorey species. Available species were on average taller, with thicker, longer‐lived leaves than unavailable species. Overall, functional dispersion was lower for seed mixes generated using available seed compared with those drawn from all species. Solutions are required to address seed shortfalls so that plantings are not only species rich but also functionally diverse. We develop two options: (1) quantifying and addressing gaps in the seed supply chain that currently limit the capacity for practitioners to restore diverse vegetation; and (2) applying a trait‐informed species‐selection method to plantings that maximises functional diversity using available seed.