China AI, decoded in English
The China AI stories English coverage misses.
I scan 100+ Chinese-language sources every day, the WeChat accounts, the Bilibili channels, the finance wires and the policy feeds, and write up what matters, in plain English, before it reaches the Western press.
120+ readers · Free · Every weekday morning
What you get
Every weekday
One email each morning. A lead story, a briefing, the signals, and the bigger picture.
100+ Chinese sources
WeChat accounts, Bilibili channels, finance wires and policy feeds, read so you do not have to.
Read the asterisk
Every story carries the honest caveat. Where the picture is disputed, we say so.
Latest issue20 Aug 2026The Internship
Start here
Five issues that show what this is
New to the Dispatch? These are the pieces that best capture the beat, the export-control fight, the model race, the memory squeeze, the grid.
The Grid
The US pulled data centers off its grid to keep home AC running through a heat wave. The same week, China wired a desert of solar straight into a server hall. The scarce input is shifting from chips to power.
Read on the newsletter →Conceded
On Nvidia’s earnings call, Jensen Huang said the quiet part out loud: China revenue is essentially zero, and the company has "largely conceded" the market to Huawei.
Read on the newsletter →The Week the Race Happened
DeepSeek open-sourced V4 under MIT with Huawei Ascend as a primary deployment path: 1.6 trillion parameters, million-token context, coding a hair behind the best Western models. The first sovereign stack.
Read on the newsletter →The Windfall
A Chinese memory middleman’s half-year profit jumped 743x. AI’s appetite for HBM is being paid for one layer down, by the budget phone in your pocket.
Read on the newsletter →The Firm Flag
Alibaba’s Qwen now accounts for more than half of every open-source AI model download on earth and has passed a billion downloads on Hugging Face, displacing Meta’s Llama, per SCMP and Pandaily, though those share figures trace to a single report. The lead is firm enough that Alibaba plans to start charging heavy users of its next open model, and in the same week Meta and Nvidia both re-entered the open-weight market they had backed away from, Meta with Muse Glimmer and Nvidia with Nemotron 3.5 Lightning, explicitly to catch the Chinese labs now setting the open-source default.
Read on the newsletter →Trackers
Maintained references, updated from the daily
Some stories are not one morning’s news, they are a running contest. We keep sourced, dated trackers on the ones that matter most.
The differentiator
China AI Chip & Export-Control Tracker
A maintained, sourced timeline of US export controls on AI chips and China’s response: Huawei Ascend, SMIC, CXMT and domestic HBM, DeepSeek efficiency, and the compute build-out. Every entry dated and sourced, updated from the reporting.
More trackers
The memory super-cycle, the domestic-chip scoreboard and the power build-out are next. See what is live and what is in progress.
All trackers →Topics
The threads we follow
The daily moves fast. These are the throughlines, each with the best issues we have written on it.
Chips & Export Controls
Ascend, SMIC, HBM, and the US control regime
Frontier Labs
DeepSeek, Moonshot, Zhipu, and the model race
Open Models
Open weights as China’s distribution strategy
Compute & Power
Data centers, the grid, and the electricity edge
Memory & HBM
The DRAM shortage, CXMT, and the HBM gap
Embodied AI
Humanoids, the capital flood, and the data shortage
AI Policy & Regulation
Beijing’s rules, Washington’s controls
Capital & Markets
Raises, IPOs, and the state’s money
Subscribe
The stories English coverage misses, every morning
I scan 100+ Chinese-language sources every day and write up the China AI stories English coverage misses. It goes out every weekday. Free.
Free. One email each weekday morning. Unsubscribe anytime.