Ok, but Does It Scale?

Here's the thing about scaling claims in open source: everybody's got one, and almost nobody can prove it. SpacetimeDB's pitch is that by fusing the database with the application server into a single "relational database with WASM modules" running in-memory, you sidestep the network latency problem that kills most real-time architectures. That's a legitimately interesting architectural bet—treating your database as a stateful compute engine rather than a dumb storage layer. But the history of databases that claim to "just scale" is littered with caveats, weird benchmark conditions, and the quiet admission that "it depends on your workload." What makes SpacetimeDB worth watching isn't the scaling math. It's the philosophical stance: that the database should own your application logic, not just your data. That's a radical inversion of the modern microservices gospel, where every service gets its own Postgres instance and you duct-tape them together with Kafka. If SpacetimeDB can genuinely deliver horizontal scale with strong consistency and server-side logic, it would eat a massive chunk of the real-time backend market—think multiplayer games, live dashboards, collaborative tools. That's a big "if," and the blog post, for all its technical posturing, needs to answer the hard questions: What happens at the tail? What's the failure mode? Where's the load test with real traffic patterns? The broader pattern here is the eternal tension between engineering and marketing in the open source world. A scaling blog post is a rite of passage—every project writes one, usually right before the community discovers the pathological case that makes it fall over. The honest projects publish their benchmarks, their y-axis labels, and their hardware specs. The less honest ones publish vibes. SpacetimeDB, to its credit, has generally been a serious engineering effort with a real community around it, and the WASM module approach is genuinely novel. But the burden of proof is on them, and the HN thread sitting at zero comments suggests the crowd is waiting for receipts, not rhetoric. Source: https://spacetimedb.com/blog/how-does-spacetime-scale