Back to Tech Radar
infrastructureNov 2, 2026

A Cache Migration Can Change Costs Without Changing Code

Vercel is moving Next.js Data Cache to Runtime Cache without requiring code changes. The migration also changes how previously unbilled reads and writes are priced, making usage and upstream load important to review.

Editorial abstract cover for A Cache Migration Can Change Costs Without Changing Code, showing a vercel and caching update as connected infrastructure signals

Summary

A migration can be invisible in the code and still change the cost of running an application. Vercel is moving Next.js Data Cache into Runtime Cache. Existing fetch caching and unstable_cache usage will switch without code changes, but previously unbilled Data Cache reads and writes will be subject to Runtime Cache pricing after migration.

The practical question for teams is not simply whether their application keeps working. It is what the same caching behavior costs under the new model—and what happens to upstream services if they cache less.

Key Updates

Vercel says teams created on or after October 30, 2026 use Runtime Cache by default. For existing teams, the schedule depends on the plan:

  • Hobby: Vercel says it will migrate teams on November 29, 2026.
  • Pro: Vercel says it will notify teams of their migration date and when billing changes take effect.
  • Enterprise: Account teams will provide migration details.

Existing Hobby and Pro teams can opt in early. Runtime Cache pricing applies as soon as they do. Hobby includes one million read units and 200,000 write units free each month across all projects. Pro usage draws from its monthly $20 infrastructure credit; usage beyond that is billed at regional rates. Enterprise pricing requires an account-team conversation. Vercel lists starting rates of $0.40 per million read units and $4.00 per million write units, varying by region. Units are billing measures, so check the pricing documentation rather than equating them automatically with individual application requests.

The migration also increases storage allocations. Vercel says each Pro project and environment receives an allocation with eight times the space of Data Cache’s previous team-wide pool; Enterprise projects get ten times their previous Data Cache storage. These storage comparisons do not predict an individual team’s bill.

Why It Matters

Caching trades repeated work at the data source for storage and cache operations. A team that removes caching to reduce cache charges may send more traffic to its database or API, increasing latency or costs elsewhere. Vercel says disabling fetch and unstable_cache caching in CDN settings does not stop the migration.

Runtime Cache is regional: on a miss, the application or framework fetches or computes the value and writes it in that region; the service does not look for a cached copy in another region. Vercel says tag invalidation updates the receiving region immediately and typically reaches other regions within a few hundred milliseconds. It also describes a circuit-breaker fallback during which recently invalidated data can briefly be treated as fresh. Applications with strict freshness requirements should evaluate that behavior against their own needs.

Builder Takeaway

  1. Find the affected cache usage. Review Data Cache usage in Vercel Observability before the team migrates. Use Runtime Cache observability after migration to investigate revalidation causes and write utilization.
  2. Estimate the whole tradeoff. Check your plan, region, allowances and migration notice, then compare cache usage with the database or API traffic a change would create. Do not assume every request equals one billing unit.
  3. Reduce unnecessary work deliberately. Review frequently written entries that are rarely read and validate cache lifetimes and invalidation against actual freshness needs.
  4. Test under realistic conditions. Measure cache misses, upstream load, latency and freshness across the regions you serve. If opting in early, remember that Runtime Cache pricing starts at opt-in.

The most useful outcome is a cache policy that fits both the application’s data requirements and its operating budget—not simply a smaller cache bill.

Sources

    How strong is this signal for builders?

    Signal feedback is stored anonymously and used to improve Tech Radar editorial quality.

    If this was useful, we can send you the next one.

    We follow changes in software, AI, and infrastructure—and explain what they mean for people building and running systems.

    Occasional Tech Radar emails from uniQubit. Unsubscribe anytime.

    Sent by uniQubit · 1407 - 177 Nonquon Rd., Oshawa, ON L1G 3S2, Canada · info@uniqubit.ca
    Privacy policy · Your confirmation email includes the controls to unsubscribe or change tracking.

    Want more operational technology signals?

    Follow uniQubit Tech Radar or contact uniQubit about a product, partnership, or operational software need.