In a multi-tenant SaaS product, many customers ("tenants") share the same underlying application and infrastructure, with their data logically separated. In a single-tenant setup, each customer gets their own dedicated instance of the software. Both are legitimate architectures, and the right choice depends on what you're buying — or building.
Multi-tenant systems are typically cheaper to run at scale, since infrastructure costs are shared across many customers, and updates roll out to everyone at once. This is why most subscription software you use day-to-day — accounting tools, CRMs, project management platforms — is multi-tenant. The trade-off is less customization per customer, since changes generally need to work for everyone on the platform.
Single-tenant setups offer more control: a dedicated database, sometimes dedicated infrastructure, and room for customer-specific customization that wouldn't make sense to build into a shared product. This usually comes at a higher cost and slower rollout of updates, since each instance may need to be upgraded individually.
For businesses building their own SaaS product rather than buying one, this decision shapes the entire technical architecture from day one — it is far more disruptive to change later than to decide correctly at the start. It's a conversation worth having explicitly with a development partner rather than leaving as an unstated assumption.