Alibaba is developing a new chip for AI inferencing, according to sources cited by CNBC.

The chip will be used to run AI applications, not for training large models.

Alibaba has a semiconductor design arm called T-head and previously released the Hanguang 800 inference chip in 2019.

The new chip will not be sold, but customers can rent computing power from Alibaba.

Chinese tech firms have been focusing on developing domestic chips amid ongoing US-China tensions and restrictions on access to American technology.

.source-ref{font-size:0.85em;color:#666;display:block;margin-top:1em;}a.ask-tia-citation-link:hover{color:#11628d !important;background:#e9f6f5 !important;border-color:#11628d !important;text-decoration:none !important;}@media only screen and (min-width:768px){a.ask-tia-citation-link{font-size:11px !important;}}

🔗 Source: CNBC

🧠 Food for thought

1️⃣ Chinese tech giants are pivoting to integrated chip-as-a-service models rather than direct sales

Alibaba’s approach of keeping its AI chips in-house while renting computing power represents a strategic departure from traditional semiconductor business models.

Unlike companies that sell chips directly, Alibaba will use its new inference chips internally and offer cloud computing services to customers1. This mirrors how the company previously handled its Hanguang 800 chip released in 20191.

This integrated approach allows Chinese firms to maintain control over their technology while generating revenue through services rather than hardware sales.

The strategy appears to be working financially, with Alibaba’s cloud division reporting 26% year-over-year revenue growth and AI-related product revenue maintaining triple-digit growth for eight consecutive quarters1.

2️⃣ Geopolitical tensions are accelerating China’s domestic semiconductor capabilities with measurable results

The recent surge in Chinese AI chip development is producing tangible performance improvements and significant revenue growth across multiple companies.

Baidu’s newly launched Baige 5.0 platform, powered by domestic Kunlunxin chips, delivers a 50% efficiency improvement for AI inference tasks and has already secured over $139 million in chip orders from China Mobile2.

Meanwhile, Cambricon, another Chinese AI chip designer, reported significant revenue growth and record profits in the first half of the year1.

These developments coincide with ongoing US export restrictions, including Nvidia’s blocked H20 chip sales to China, which has forced Chinese companies to accelerate their domestic alternatives1.

The trend demonstrates how geopolitical pressure is driving not just investment but actual commercial success in China’s homegrown semiconductor sector.

Recent Alibaba developments

……