/
Navigation
Chronicles
Browse all articles
Explore
Semantic exploration
Research
Entity momentum
Nexus
Correlations & relationships
Story Arc
Topic evolution
Drift Map
Semantic trajectory animation
Posts
Analysis & commentary
Pulse API
Tech news intelligence API
Browse
Entities
Companies, people, products, technologies
Domains
Browse by publication source
Handles
Browse by social media handle
Detection
Concept Search
Semantic similarity search
High Impact Stories
Top coverage by position
Sentiment Analysis
Positive/negative coverage
Anomaly Detection
Unusual coverage patterns
Analysis
Rivalry Report
Compare two entities head-to-head
Semantic Pivots
Narrative discontinuities
Crisis Response
Event recovery patterns
Connected
Search: /
Command: ⌘K
Embeddings: large
TEXXR

Chronicles

The story behind the story

days · browse · Enter similar · o open

Alibaba is developing its own neural network chip, the Ali-NPU, and has acquired Chinese IoT chip designer C-SKY Microsystems

Masha Borak / TechNode :

TechNode Masha Borak

Context & Ripple Effects

This 2018 report is the origin point of Alibaba's chip program. Two years after AliCloud partnered with Nvidia on a GPU-based cloud platform in China, Alibaba flipped from buying compute to designing it: an in-house neural network processor, the Ali-NPU, plus the acquisition of IoT chip designer C-SKY Microsystems for embedded silicon know-how.

The arc since then validates the move. The chip subsidiary announced months later became T-Head, which shipped the Hanguang 800 for internal search and recommendation workloads and, by 2026, the Zhenwu M890 with planned annual upgrades — while a 2025 inference chip competing with Nvidia's H20 moved manufacturing to a Chinese foundry.

First-order effects

  • Alibaba gains in-house design capability for both datacenter AI (Ali-NPU) and connected devices (C-SKY's IoT portfolio), reducing its dependence on merchant silicon for its cloud and commerce workloads.
  • C-SKY's team and IP are absorbed into Alibaba rather than remaining an independent supplier to other Chinese device makers.

Second-order effects

  • Nvidia's relationship with Alibaba shifts from vendor-customer toward rivalry: the 2016 GPU cloud partnership now coexists with Alibaba building inference chips that compete directly with Nvidia parts like the H20.
  • Alibaba's chip roadmap pulls toward domestic manufacturing — the 2025 H20-class inference chip was made by a Chinese company, unlike its earlier TSMC-fabricated AI chip — tightening ties between its design unit and the local foundry ecosystem.

Third-order effects

  • If the pattern holds, Alibaba converges on a full proprietary-plus-open stack — custom silicon, Qwen models, and open-sourced chip software following Huawei and Moore Threads — aimed at breaking Nvidia's CUDA lock-in for Chinese developers.
  • China's cloud giants structurally become chipmakers, with T-Head's annual-upgrade cadence signaling that AI accelerators are now a core product line rather than a side project.

The trend: Chinese cloud platforms are vertically integrating from AI buyers into AI silicon suppliers, with Alibaba's 2018 NPU-and-acquisition start maturing into a full-stack challenge to Nvidia's ecosystem.