# Amazon — How Salesforce is advancing global-scale model serving with cross-Region inference routing in Amazon SageMaker AI

- Company: Amazon (amazon.com)
- Announced: 2026-11-18T20:13:18+00:00
- Category: not stated
- Coverage: not counted
- Announcement: no
- Group: routine
- Source: https://aws.amazon.com/blogs/machine-learning/how-salesforce-is-advancing-global-scale-model-serving-with-cross-region-inference-routing-in-amazon-sagemaker-ai/
- Record: https://forck.live/items/38211-how-salesforce-is-advancing-global-scale-model-serving-with-cross-region
- Subject: Bedrock / Nova

Salesforce's Agentforce Foundation team is using cross-Region inference routing in Amazon SageMaker AI to serve models across 30+ global Regions without splitting capacity or breaking data residency commitments. Hosting models separately in each Region fragmented capacity, leaving some Regions with idle instances while others ran out, and the team's own gateway had no capacity signals from local or remote endpoints. The router is a Regional SageMaker AI resource that applications invoke instead of an endpoint, passing a routing target name that lists the endpoints and inference components allowed to serve one logical model by their ARNs. Destinations can sit in another Region or AWS account, with the owning account granting access through a resource-based endpoint policy that allows one named router to invoke it. The router prefers its own Region and shifts a growing share of new requests to healthy destinations within about a minute when platform faults, 5xx errors, throttles or timeouts persist, returning traffic more slowly after recovery. Client errors and model latency do not move traffic, and the router does not deploy models or add instances, relying on each endpoint's own autoscaling. In Salesforce's Internal Gateway the router call sits in the per-request egress path after the data residency check, and requests tagged with a residency requirement are pinned to an in-Region routing target that can never fail over across a boundary.

## Evidence

Verbatim from https://aws.amazon.com/blogs/machine-learning/how-salesforce-is-advancing-global-scale-model-serving-with-cross-region-inference-routing-in-amazon-sagemaker-ai/:

> Thanks to Amit Gupta, Senior Director- AI/ML and Mark Wang, Software Engineering PMTS from Salesforce for your contributions in the vision and partnership of cross-Region inference routing on SageMaker AI.
> 
> Salesforce’s Agentforce Foundation team serves a wide variety of models for products across Salesforce, including Agentforce and the Salesforce Trust Layer, in 30+ global Regions. These include both models trained in house and open source models. Hosting models separately in each Region splits capacity and makes traffic spikes harder to absorb, which turns a capacity shortfall into an availability problem.
> 
> This post shows how Salesforce set up the cross-Region inference routing in SageMaker AI and how it lets the team add capacity securely while keeping its data residency commitments. It also walks through how Salesforce sees where each request was served and why. The cross-Region inference routing gives an application one invocation path across SageMaker AI endpoints in several Regions and accounts, and sends each request to an endpoint that can serve it. Those endpoints are Salesforce’s own, already deployed.

---

Record: https://forck.live/items/38211-how-salesforce-is-advancing-global-scale-model-serving-with-cross-region
Catalogue: https://forck.live/llms.txt
Current issue: https://forck.live/feed.md
