Does Amazon Redshift Work in China? Data Residency, ICP & Cross-Border Rules
Amazon Redshift is AWS's managed data warehouse — usually the largest pool of analytical personal data a company holds at rest. AWS's global partition has no mainland-China region, though Redshift runs in the separate AWS China partition. A compliance-first look at residency, PIPL cross-border transfer and ICP.
Does Amazon Redshift work in China?
Whether you can use Amazon Redshift in China is a data-residency question, not a speed one. Redshift is a data warehouse — the single place your analytical data comes to rest, and usually the largest concentration of personal information on your whole stack.
In AWS's global partition there is no mainland-China region, so a Redshift cluster there keeps that data offshore — a cross-border transfer of personal information under PIPL (Articles 38–40). A genuine in-country option does exist: AWS's own China regional services list shows Amazon Redshift available in both AWS China Regions — Beijing (operated by Sinnet) and Ningxia (operated by NWCD) — but that is a separate partition, account and operator, and it is not automatic compliance. It still owes ICP filing for any China-facing surface, MLPS, and the data-localization duty (CSL Article 39, formerly Article 37) for a CIIO or high-volume handler.
This is a risk picture, not a verdict — your duties turn on your entity, your data volumes and your role. Our China team can map your Redshift residency exposure →
What Amazon Redshift's own documentation says about China
| Fact | Primary source |
|---|---|
| In AWS's own China regional services list: "Amazon Redshift" is shown as available in both AWS China Regions — the Beijing Region (operated by Beijing Sinnet Technology Co., Ltd.) and the Ningxia Region (operated by Ningxia Western Cloud Data Technology Co., Ltd.). That partition is reached through a distinct Amazon Web Services (China) account with its own ICP filing — an in-country option, but a separate operator and contract, not the service you run from a global account. | Amazon Web Services China — Amazon Services by Region (regional product services), retrieved 2026-10-14 |
| In AWS's global partition, the Amazon Redshift service endpoints span the Americas, Europe and Asia Pacific — the nearest being Hong Kong — with no mainland-China region among them. A cluster launched from a standard (global) AWS account therefore stores its data outside mainland China. | AWS General Reference — Amazon Redshift endpoints and quotas, retrieved 2026-10-14 |
| Holding the personal information of people in China in an offshore Redshift cluster is a cross-border transfer under PIPL Articles 38–40: the handler must give notice, obtain separate consent, and satisfy one transfer mechanism — a CAC security assessment, the CAC standard contract, or certification. | Personal Information Protection Law (PIPL), Articles 38–40 — cross-border provision of personal information |
| A critical-information-infrastructure operator or high-volume handler must store China personal information on the mainland — PIPL Article 40 and the data-localization rule in Cybersecurity Law Article 39 (formerly Article 37). A warehouse's large, long-horizon analytical volumes make that high-volume threshold easy to cross, and an offshore cluster cannot meet an in-country storage duty. | China Cybersecurity Law, Article 39 (formerly Article 37); PIPL Article 40 — data localization |
Sources verified by the 21YunBox compliance team on 2026-10-14.
For a mainland-China audience, the question worth asking about Amazon Redshift is not whether a cluster spins up or whether a BI tool can query it — both are routine. It is where the data that warehouse holds is allowed to come to rest. Amazon Redshift is AWS’s managed, petabyte-scale data warehouse: the place an organization consolidates its analytical data — customer tables, orders, events, product and behavioral logs — into one queryable copy. That makes it the most concentrated data-residency question on the whole stack, because a warehouse is usually the single largest pool of personal information a company keeps. AWS’s global regions include no mainland-China region to select, so a cluster provisioned in a global region keeps that consolidated data offshore. A genuine in-country option does exist — Redshift is offered in the separate AWS China partition — but that is a different operator, account and compliance surface, not the same service you reach from a global account.
Amazon Redshift in China at a glance
| What decides it | In Amazon Redshift's own terms — and China's law |
|---|---|
| Where the data physically rests (the region reality) | AWS's global partition has no mainland-China region. A cluster in a global region (Tokyo, Singapore, Hong Kong) keeps the warehouse offshore. Mainland capacity exists only in the separate AWS China partition — Beijing (operated by Sinnet) and Ningxia (operated by NWCD) — on a distinct Amazon Web Services (China) account, where Redshift is offered. |
| What a warehouse holds, and why it is personal information | Redshift is the consolidated analytical copy of the business: customer tables, orders, events and behavioral logs joined together. That is usually the single largest concentration of personal information you hold, and some of it — contact, financial and identity fields — is sensitive. Where it rests is a legal question, not a performance one. |
| Your China users' records when the cluster is offshore | Hold the personal information of people in China in an offshore Redshift cluster and you are making a cross-border transfer under PIPL (Articles 38–40): notice, separate consent, and one transfer mechanism — a CAC security assessment, the standard contract, or certification. |
| The in-country storage duty (CIIO / high-volume handler) | A critical-information-infrastructure operator or high-volume handler must keep China personal information on the mainland (PIPL Article 40; Cybersecurity Law Article 39 (formerly Article 37) — the 2025 amendment, in force January 1, 2026, renumbered the data-localization article from 37 to 39, substance unchanged). A warehouse's large analytical volumes make that threshold easy to cross, and an offshore cluster structurally cannot meet it. |
| An AWS China region is not automatic compliance | The in-country partition is a separate operator, account and contract, with its own service and feature gaps. Running Redshift there is a lawful option, but it still owes ICP filing for any China-facing surface, MLPS, separate consent and the data-localization analysis. In-country infrastructure is the start of compliance, not the whole of it. |
| Reachability is not the axis (and the China-facing surface) | Whether the warehouse is reachable from China is a delivery matter, not the legal test; data-at-rest residency is. The lawful levers are to run Redshift in-country — the licensed sovereign AWS China region, or another in-country deployment — and to put the dashboards, APIs and admin portals that read from it on ICP-filed delivery. |
Where the data actually rests
Choose a Redshift region from a standard AWS account and the list you pick from is the global partition — and it contains no mainland-China region. The nearest regions, in Tokyo, Singapore or Hong Kong, are all outside the mainland, so the warehouse and every row it stores sit offshore. (Hong Kong is Chinese territory, but it is treated separately from the mainland for data-localization purposes.) AWS’s own global service-endpoints reference lists Redshift across the Americas, Europe and Asia Pacific with no mainland-China endpoint among them.
Mainland capacity does exist, but through a different door. AWS China is a separate partition: the Beijing Region is operated by Sinnet (Beijing Sinnet Technology Co., Ltd.) and the Ningxia Region by NWCD (Ningxia Western Cloud Data Technology Co., Ltd.), and AWS’s China regional services list shows Amazon Redshift available in both. Using them means a distinct Amazon Web Services (China) account and an ICP filing — so an in-country, sovereign-cloud residency option for the warehouse genuinely exists.
What it is not is the same service you already run globally. It is a separate operator, a separate contract and account, its own service and general-availability gaps, and its own compliance surface. An in-country region is a lawful option; it is not automatic compliance.
What the warehouse holds is personal information
A data warehouse is not a cache or a CDN edge — it is the analytical system of record. Redshift is where an organization brings together customer tables, order and payment history, support interactions and the event and behavioral logs that describe what people did. Joined into one place, that is typically the single largest concentration of personal information a company holds, and much of it — contact details, financial data, identifiers — is sensitive personal information that draws heightened duties.
That is why residency, not reachability, is the test. When the personal information of people in China rests in an offshore Redshift cluster, you are making a cross-border transfer, and PIPL Articles 38–40 govern it: notice, separate consent, and one lawful transfer mechanism — a CAC security assessment, the CAC standard contract, or certification. The security-assessment measures set when that assessment is mandatory rather than optional, and the exposure scales with volume — which a warehouse, by design, accumulates.
On top of the transfer rules sits a harder duty. A critical-information-infrastructure operator or a high-volume handler must store China personal information inside the mainland — PIPL Article 40 and the data-localization rule in Cybersecurity Law Article 39 (formerly Article 37). Because a warehouse concentrates large, long-horizon analytical data, it is among the easiest systems to push a handler over the high-volume threshold — and a cluster in an offshore global region cannot satisfy an in-country storage duty, because the data is, by definition, not in the country.
Why a reachability check misses this
Point a probe at a Redshift endpoint from Shanghai and you can measure a connection — and learn nothing about the question that matters. The warehouse answers queries just as well from an offshore region; that is precisely the trap. The exposure is not that the cluster is slow or unreachable, it is that the consolidated personal information of people in China is sitting outside the country without a lawful basis, and that a CIIO or high-volume handler owes an in-country storage duty an offshore cluster cannot meet. A speed or uptime test cannot see where data rests or on what legal footing it crossed a border — so it reports “fine” on exactly the cases that carry the most regulatory risk.
The lawful path — map, localize, deliver
21YunBox is a compliant overlay, not a migration — and a partner to AWS and to your data team, not a competitor to them.
- Map. We read the PIPL cross-border, data-residency, data-localization (CII) and ICP obligations against your entity, your data volumes and whose personal information actually sits in the warehouse — so you know which duties bite before you move anything.
- Localize. We run the warehouse in-country so the personal information stays on mainland soil — on a licensed in-country / sovereign-cloud option such as Amazon Redshift in the AWS China partition, or another in-country deployment — and keep consented, in-country storage for what must stay. Localize means an in-country endpoint, not a route back to an offshore one.
- Deliver. Any China-facing surface in front of the warehouse — the dashboards, the analytics API, the admin and reporting portals your mainland users reach — runs over ICP-filed, in-country delivery (the 21YunBox Optimizer), in front of the stack you already run. No rebuild, no second codebase.
Running Redshift in the AWS China region keeps the analytical data in-country, but it is the beginning of compliance, not the end: the ICP filing, MLPS, separate consent and the localization analysis still have to be done. This page maps exposure; it is not a legal ruling. Your actual duties turn on your entity, how you classify the data, your transfer volumes and whether you are a critical-information-infrastructure operator — so settle those specifics with qualified China counsel before you rely on any single path. 21YunBox never uses or suggests circumvention of any kind.
Related reading
- Cross-border data transfers under PIPL
- China’s Cybersecurity Law and data localization
- China’s data-export security assessment measures
- How to get an ICP license for a website in China
