China targets AI-powered price manipulation in new antitrust guidelines

Market regulator SAMR outlines risks of collusion, unfair subsidies and walled gardens in online platforms

Reading Time:2 minutes

The guidelines zero in on the sophisticated, often opaque ways in which online platforms with significant market power can exploit technological advantages to exclude rivals and harm consumer interests. Photo: Shutterstock

Published: 9:00pm, 17 Nov 2025Updated: 2:37am, 18 Nov 2025

China’s market regulator has moved to tackle the hidden risks of algorithm-driven price manipulation with newly proposed anti-monopoly guidelines for online shopping, food delivery and travel platforms.

A draft of the “Anti-Monopoly Compliance Guidelines for Internet Platforms”, published over the weekend by the State Administration for Market Regulation (SAMR), focused heavily on the sophisticated, often opaque ways in which online platforms with significant market power can exploit technological advantages to exclude rivals and harm consumer interests.

The document identified eight “new risks”, including collusion among platforms through coordinated algorithms to fix pricing and commission fees, dominant players forcing merchants to sign exclusive contracts, and practices like excessive subsidies or unjustified refusals to allow merchants to operate stores.

It explicitly prohibited platforms with a dominant market position from “imposing application-layer or network-layer blockade or exclusion measures against transaction counterparties”.

Chinese internet giants, including Alibaba Group Holding, Tencent Holdings and ByteDance, previously built “walled gardens” – like Alibaba’s Taobao not offering Tencent’s WeChat Pay, or WeChat not redirecting links of ByteDance’s Douyin short video app – but the restrictions started to ease last year. Alibaba owns the South China Morning Post.

China’s market regulator is soliciting public opinion on the draft regulation up to November 29. Photo: Getty

SAMR demanded that platform operators “conduct targeted screening and dynamic monitoring of core algorithmic models, including pricing algorithms, recommendation systems, ranking and advertising placement strategies”.

Select Voice

Choose your listening speed

Get through articles 2x faster

1.25x

250 WPM

Slow

Average

Fast

00:0000:00

1.25x