Spacetime details horizontal scaling strategy

Spacetime’s scaling story is credible because it treats coordination as a real cost—but it also makes workload topology part of the developer contract.

• –Single-threaded execution is a deliberate optimization for contention, while replication primarily adds availability rather than write throughput.
• –Current per-database state remains memory-bound; planned disk and object-storage tiers address capacity and cost.
• –Actor-style sharding can scale regions or tenants concurrently while keeping local transactions fast.
• –Cross-shard workflows require asynchronous messaging or planned two-phase commits, adding operational and consistency complexity.
• –Standalone SpacetimeDB and proprietary Spacetime Cloud offer different scaling capabilities, making portability and licensing important considerations.