Home / Models & DBs / Supabase Pricing Explained

BaaS Cloud Economics • 2026 Edition

Supabase Pricing Explained: Pro Tier, Compute Add-ons & Bill Optimization (2026)

Why production apps on Supabase rarely cost $25/mo. Understand Compute Add-ons ($10-$960/mo), the permanent disk auto-expansion ratchet, and $0.09/GB egress billing.

⚡ 60-Second Supabase Verdict ✓ Verified October 2026

All-In-One BaaS Platform • PostgreSQL 16 + pgvector

Supabase Production Cost Verdict: Supabase Pro costs $25/month, bundling managed Postgres, 50k MAU free authentication, 100GB storage, and 250GB unified bandwidth egress. High-traffic apps scale smoothly via dedicated compute add-ons (Micro $10/mo to 16XL $1,010/mo), eliminating the complex multi-service invoice sprawl of AWS RDS, S3, and Cognito. Compare database architectures in Supabase vs Neon and check Supabase Overage Guide.

Pro Plan Base $25.00 / month

Free Auth MAUs 50,000 MAUs Included

Included Bandwidth Egress 250 GB / month

AI Vector Search Native pgvector Extension

Direct Answer • Supabase Pro Realities

Supabase Pro is $25/month base (includes Nano compute, 8 GB disk, 250 GB egress, and 100k MAUs). Real production invoices typically reach $65 to $250+/month due to: 1) Compute Add-ons (Small is +$20/mo, Medium is +$40/mo for RAM/CPU headroom), 2) Disk Auto-scaling ($0.125/GB) that expands automatically and never shrinks, and 3) Egress ($0.09/GB) across REST endpoints and direct Postgres connections.

True Supabase Monthly Invoice Calculator

Adjust compute instance size, database disk usage, and monthly egress to calculate your realistic invoice.

Dedicated Compute Add-on Small (+$20/mo • 2 vCPU, 2GB)

Database Disk Allocation (8 GB Included) 40 GB

Monthly Data Egress (250 GB Included) 600 GB

Total Projected Monthly Bill

$80.50

$25 Base + $20 Compute + $4.00 Disk + $31.50 Egress.

Base Pro Subscription: $25.00

Disk Overage ($0.125/GB over 8GB): $4.00

Egress Overage ($0.09/GB over 250GB): $31.50

1. Supabase Compute Add-on Matrix

In real production, the shared Nano instance included in Pro frequently runs out of memory (OOM) during migrations or indexing. Compute upgrades are billed hourly:

Instance Size	vCPU Specs	Dedicated RAM	Monthly Add-on Fee	Recommended Workload
Nano (Default)	2-core ARM (Shared)	1 GB RAM	$0 (Included)	MVPs, low traffic, development
Micro	2-core ARM (Dedicated)	1 GB RAM	+$10 / mo	Eliminates noisy-neighbor CPU throttling
Small	2-core ARM (Dedicated)	2 GB RAM	+$20 / mo	Standard SaaS launch tier
Medium	2-core ARM (Dedicated)	4 GB RAM	+$40 / mo	Sustained API traffic, pgvector search
Large	2-core ARM (Dedicated)	8 GB RAM	+$80 / mo	High connection volume, heavy background jobs

2. The Disk Auto-Expansion Ratchet Trap

Supabase implements automatic disk expansion to protect databases from halting when storage fills up. While beneficial for reliability, it introduces an irreversible cost ratchet:

When your database crosses 90% utilization, Supabase automatically increases disk space by 50% or 10 GB (whichever is larger).
AWS EBS volumes can only increase in size; they cannot be shrunk online.
If a temporary migration script or bulk analytics CSV temporarily fills 100 GB, your disk expands to 150 GB. You will pay for 150 GB ($17.75/mo overage) indefinitely, even if your actual data size shrinks back to 5 GB.

“We ran an unoptimized pg_dump export on our Supabase database that duplicated logs into a scratch table. Disk ballooned from 15GB to 80GB in an hour. Even after dropping the table and running VACUUM FULL, our bill permanently climbed by $9/mo because AWS EBS doesn't support downscaling without a full database recreation.”

SaaS CTO • Hacker News Supabase Discussion

Frequently Asked Questions

Why do Supabase production bills exceed the $25/mo Pro tier?

The $25/mo Pro plan includes a shared Nano compute instance (2-core ARM, 1GB RAM) with 8GB disk storage and 250GB egress. Production workloads routinely require Compute Add-ons (Small $20/mo, Medium $40/mo, Large $80/mo), additional disk storage ($0.125/GB), and extra egress ($0.09/GB).

What is the Supabase disk auto-scaling ratchet trap?

When your Supabase database reaches 90% disk capacity, Supabase automatically resizes your disk by 50% or 10GB to prevent read-only lockouts. However, disk size cannot automatically downscale. Even if you delete half your tables, you continue paying for the enlarged disk allocation forever unless you dump and restore into a new project.

How does Supabase charge for database egress?

Pro plans include 250GB/month of pooled egress. Bandwidth exceeding 250GB is billed at $0.09 per GB. Egress includes REST API responses, Auth callbacks, Storage asset downloads, and direct Postgres port 5432 queries.

Compute Add-On Architecture, pgvector RAM Contention & Egress Bandwidth Pricing

Supabase provides an open-source Firebase alternative built around dedicated AWS PostgreSQL instances. Supabase bundles relational persistence, Row Level Security (RLS), real-time WebSockets, automated auth, and pgvector vector search into a unified platform. Supabase's Pro Plan ($25/month) includes a shared Micro compute instance (1 GB RAM, 2 shared vCPUs). However, scaling AI retrieval workloads with millions of vectors requires upgrading to dedicated Compute Add-Ons (ranging from Small at $10/mo to 16XL at $1,970/mo) to prevent PostgreSQL from killing processes due to Out of Memory (OOM) errors.

Supabase Compute Add-Ons & RAM Allocation Matrix

Compute Add-On	vCPU / RAM	Monthly Cost	Max pgvector Vectors	Target Production Workload
Micro (Default Pro)	2 shared / 1 GB RAM	$25.00 (Base)	~100,000 (1536-dim)	MVP, staging, small apps
Small Add-On	2 dedicated / 2 GB RAM	$35.00 / mo	~350,000 (1536-dim)	Growing SaaS with search
Medium Add-On	2 dedicated / 4 GB RAM	$85.00 / mo	~1,000,000 (1536-dim)	Production AI apps, RAG
XL Add-On	4 dedicated / 16 GB RAM	$235.00 / mo	~5,000,000 (1536-dim)	High-concurrency enterprise

Egress Bandwidth Shock Prevention & Cloudflare Proxying

Supabase bills database egress at $0.09 per GB beyond the included 250 GB on the Pro tier. If client applications query large JSON arrays or stream media files directly from Supabase Storage without caching, egress bills can skyrocket. Placing Cloudflare in front of Supabase Storage and using cached pre-signed URLs eliminates storage egress fees and offloads read bandwidth to Cloudflare's free global CDN edge.

Point-in-Time Recovery (PITR) & Storage Autoscale Guardrails

Enable Point-in-Time Recovery (PITR at $100/mo) for mission-critical financial databases to roll back state to specific seconds in the event of accidental data corruption. Furthermore, configure database disk auto-scaling with caution: disk volume automatically expands when capacity hits 90%, but PostgreSQL disk sizes cannot be shrunk without dumping and rebuilding databases.

Supabase Disk Autoscaling & Backup Retention Policies

Supabase projects automatically expand database disk capacity when storage utilization reaches 90%. Because cloud block storage volumes cannot be shrunk after expansion, staging and test environments should establish automated data cleanup crons to purge stale vector embeddings and temporary upload blobs, avoiding permanent storage tier creep.