Key details

  1. Effective October 1, 2026: 20–45% reductions on existing Key Value plans with 1GB or more RAM.
  2. Existing eligible Key Value instances receive the new prices automatically.
  3. New 3GB tier: up to 3,000 simultaneous connections.
  4. New 50GB tier: up to 50,000 simultaneous connections.
  5. New Key Value instances run Valkey 8; legacy instances created before February 12, 2025 run Redis 6.
  6. Render currently permits Key Value upgrades, not downgrades; upgrades briefly interrupt service.
  7. Historical dollar comparisons from independent pricing monitoring are not a substitute for current checkout pricing.

What builders should take away

  1. Check the October bill for the automatic rate change before changing a working Key Value deployment.
  2. Compare current memory use and connection count against the new 3GB tier, but remember that a Key Value plan cannot presently be downgraded.
  3. Treat cache and job-queue persistence differently; choose an eviction policy appropriate to the workload.
  4. Check the current Render pricing table for the exact region and tier rather than relying on older plan names or third-party price snapshots.
  5. Do not confuse this Key Value reduction with Render's August general web-service memory-plan expansion or September Workflows billing changes.

What changed

On October 1, 2026, Render cut compute prices by 20–45% on its paid Key Value plans with 1GB or more RAM. It also introduced a 3GB tier supporting up to 3,000 connections and a 50GB tier supporting up to 50,000. Render says existing eligible services receive the lower price automatically, without a migration. This applies specifically to its Redis-compatible Key Value offering, which uses Valkey 8 for newly created instances. It does not mean ordinary web-service compute prices have been reduced.

Why it matters

For small SaaS teams, a managed cache or background-job queue can become a persistent infrastructure bill even when traffic is modest. Cutting that bill without a migration changes the economics of retaining managed state rather than running Valkey yourself. The new 3GB option also fills a meaningful gap between 1GB and 5GB; teams can buy a closer fit instead of overprovisioning. But the operational decision still depends on connection limits, persistence and the fact that Render does not currently allow Key Value plan downgrades.

The lower rates apply to existing Key Value customers

Render's October 1 announcement says all Key Value plans with at least 1GB RAM have lower prices, with reductions of 20–45%. Existing paid instances on eligible tiers are automatically billed at the new rates from October 1; no manual switch is required. An independent pricing-history snapshot compared September 28 and October 5 rate cards and reported the 1GB tier falling from $32 to $20 monthly, 5GB from $135 to $100, 10GB from $250 to $200, 20GB from $550 to $350, and 40GB from $1,100 to $600. Those historical figures are third-party observations; Render's current pricing page is the authority for a new purchase.

New 3GB and 50GB sizes fill the gaps

Render's changelog confirms two additional tiers: 3GB with up to 3,000 simultaneous connections, and 50GB with up to 50,000. The documented Key Value plan IDs are `3g` and `50g`; the existing tiers remain available. The 3GB step is useful when 1GB is too tight but 5GB would be wasteful. The 50GB option raises the single-instance ceiling beyond the previous 40GB tier.

It is a Valkey-compatible service, not an ordinary database upgrade

Render Key Value uses Valkey 8 for newly created instances, with Redis-compatible clients and private-network connectivity. Older instances created before February 12, 2025 can still run Redis 6 and do not receive version updates. Paid instances support disk persistence; free ones do not. The price announcement is not a promise of Redis 6 upgrades, cross-region replication, or any change to Render Postgres.

Plan changes still require operational care

Render's current documentation says Key Value plans can be upgraded but not downgraded. Upgrades cause a minute or two of unavailability and can lose data depending on persistence mode; free instances always lose their data on upgrade. Teams running queues should use a suitable persistence and eviction policy, verify client connection limits, and plan any upgrade window rather than assuming a cheaper plan means an automatic right-size migration.

What to watch next

  • Whether Render enables safe Key Value downgrades or offers in-place resizing without downtime.
  • Changes to connection limits, storage/persistence charges or regional availability.
  • Whether price reductions lead to comparable changes from competing managed Redis/Valkey services.

Still unclear

  • Render's announcement gives a percentage range; the historical before/after dollar examples come from independent pricing monitoring and may not capture every SKU or discount.
  • The impact on a particular bill depends on memory tier and any other account charges.
  • The announcement does not promise new durability or high-availability guarantees.

Sources

Direct reading behind this dossier.

4 sources
Render Key Value documentation
Render primary_documentation

Valkey/Redis compatibility, persistence, connectivity, plan changes and downtime caveats.

Discussion

Discussion is reader-contributed. Comments are not part of the BTN dossier or its editorial evidence.

0 visible comments

Join the discussion

Keep comments useful and relevant. Reader contributions may be moderated and are not BTN editorial evidence.

Sign in to comment