# Vercel Sandbox routing is now 18x faster globally

Source: [Vercel](https://vercel.com/changelog/vercel-sandbox-routing-is-now-18x-faster-globally)  
Feed7 permalink: https://feed7.dev/p/vercel-sandbox-routing-is-now-18x-faster-globally-1xr4uw7  
Published: 2026-09-08T20:00:00.000Z  
Trust: Official Source (official_source)

## Why Included

Vercel cut median Sandbox domain lookup latency from 62 ms to 3.4 ms by resolving domains from regional replicas. Agent sandboxes get the change automatically with no pricing change.

## Source Summary

Vercel now resolves Sandbox public domains from the nearest regional replica rather than one centralized store. **Median lookup latency fell from 62 ms to 3.4 ms**, an **18× reduction**.

## Practical Implication

Builders running browser previews, callbacks, or agent-generated services on Sandbox domains get lower routing overhead automatically. No SDK migration or configuration change is described.

## Agent-Ready Context

Vercel now resolves Sandbox public domains from the nearest regional replica rather than one centralized store. **Median lookup latency fell from 62 ms to 3.4 ms**, an **18× reduction**.

Builders running browser previews, callbacks, or agent-generated services on Sandbox domains get lower routing overhead automatically. No SDK migration or configuration change is described.

The largest reported gains are regional p99 figures—up to **112× in Sydney** and **146× in Cape Town**—so they should not be read as typical global latency. Process startup and application latency are outside this measurement.

## Connected Context

Feed7 judgment across 713 accumulated Signals:

This removes a specific infrastructure tax for public Sandbox endpoints by distributing domain resolution, benefiting previews, callbacks, and agent-hosted services without a described migration. It complements sandbox capability and isolation improvements, but the measurement is only routing lookup latency; it neither demonstrates faster process startup nor establishes lower end-to-end application latency.

- [Vercel AI Gateway and Vercel Sandbox now available on Hermes Agent](https://feed7.dev/p/vercel-ai-gateway-and-vercel-sandbox-now-available-on-hermes-agent-04gyfsv) — Hermes can use Sandbox as an opt-in execution backend; the routing change reduces lookup overhead for public endpoints created by that deployment choice without changing its runtime, linking, or OIDC prerequisites.
- [How Sandstone grew 40x in 147 days on Vercel](https://feed7.dev/p/how-sandstone-grew-40x-in-147-days-on-vercel-1bhjz28) — Sandstone uses preview deployments as part of an agent workflow, making lower public-domain routing overhead relevant to one concrete deployment pattern, while still leaving its missing end-to-end latency and reliability evidence unresolved.
- [TencentCloud/CubeSandbox](https://feed7.dev/p/cubesandbox-1d2cg1q) — CubeSandbox’s cold-start claim measures sandbox creation, while Vercel measures domain lookup; the contrast prevents treating either number as a complete measure of time to a usable agent service.
- [Scaling Managed Agents: Decoupling the brain from the hands](https://feed7.dev/p/managed-agents-00zpz67) — Managed Agents reports improvement in time-to-first-token through harness–sandbox decoupling, whereas this change optimizes endpoint routing; they address distinct latency stages in hosted agent infrastructure.

## Context Map

- Layer: infra
- Domains: coding
- Topics: sandboxing

## Uncertainty

- The largest reported gains are regional p99 figures—up to **112× in Sydney** and **146× in Cape Town**—so they should not be read as typical global latency. Process startup and application latency are outside this measurement.

## Agent Instruction

Use this item as source-backed context. Do not invent claims beyond the linked source. If this item conflicts with another source, call out the conflict.