Following its entry into the discrete GPU sector, Moore Threads is now making waves in the APU market with the introduction of a high-performance system-on-chip (SoC) designed for laptops.
Moore Threads’ Latest Laptop Chip Delivers Exceptional Edge AI Performance, Competing with Current Trends
Moore Threads has become a recognizables name in the tech industry, frequently gaining attention for its innovative solutions. For instance, the Chinese GPU maker was among the pioneers to unveil a PCIe 5.0 GPU, although it later proved less powerful than NVIDIA’s GTX 1050 Ti. Now, the company is pivoting towards the laptop SoC arena with its proprietary ‘Yangtze’ architecture, which integrates native GPU capabilities, sophisticated software, and advanced CPU microarchitectures.

Examining its technical specifications, the Yangtze chip boasts a robust 12-core configuration built on the ‘ARMv8’ architecture, complemented by an integrated GPU leveraging the company’s MUSA architecture. The chip operates at a base frequency of 2.65 GHz and is embedded in the MTT AIBOOK, an OEM product released by Moore Threads. This laptop is also equipped with 32 GB of DDR5-7500 memory and a substantial 1 TB SSD, all backed by an OLED display offering a refresh rate of 120 Hz.

This laptop is positioned as a strong contender for mid-to-high-end workloads, yet the standout feature is undoubtedly the custom-built chip. While the previous iteration of the chip featured an 8-core setup, this new version has seen a significant upgrade. Notably, it includes a neural processing unit (NPU) capable of achieving 50 TOPS (INT8) performance, which is optimized for tasks like speech and image recognition. Here are some highlights of the chip:
- 12-core CPU
- Boost Clock of 2.65 GHz
- High-Performance Integrated GPU
- 50 TOPS Neural Processing Unit
- Support for H.265/H.264/AV1 Video Codecs
- Targeted Towards the AI PC Market
The AIBOOK aims to leverage the growing AI PC trend in China, where utilizing models on local hardware has gained massive popularity, influenced by the country’s commitment to developing open-source large language models (LLMs).While the specifications look promising on paper, the real test will be performance benchmarks in practical applications, as theoretical capabilities do not always translate into real-world efficiency.
Regarding pricing, the MTT AIBOOK is currently listed at approximately $1, 440 on JD.com in China. While this price tag may seem steep, it reflects Moore Threads’ first foray into the APU segment.
For further information, consult the news source: Videocardz.
Leave a Reply