China's leading GPU developer, Moore Threads, has delivered a highly encouraging set of preliminary half-year financial results.
According to a performance forecast released by the company on July 16 for the first half of 2026, it anticipates generating revenue between 1.65 billion and 1.75 billion yuan, representing a substantial year-on-year increase ranging from 135.12% to 149.37%. This robust growth, measured against a base of 702 million yuan in the same period last year, is notably exceptional within the domestic semiconductor industry. The forecast did not include profit figures.
The primary driver behind this impressive expansion is clear: the ongoing surge in artificial intelligence has fueled strong demand for fully-featured GPUs. Concurrently, the company's flagship product, the MTTS5000, has entered mass production, and the commercialization of its Kuae intelligent computing clusters is accelerating, enabling Moore Threads to rapidly translate its technological strengths into tangible commercial gains.
Revenue More Than Doubles, Scaling New Heights
Examining the figures, Moore Threads' projected revenue for H1 2026 falls between 1.65 billion and 1.75 billion yuan, which translates to growth of 135.12% to 149.37% year-on-year, or approximately 2.35 to 2.49 times the previous year's level.
This growth rate significantly surpasses general market expectations for domestic GPU manufacturers and signals the company's formal transition from a phase of "high-speed startup growth" to a new stage of "scaled commercial monetization."
Using the midpoint estimate, first-half revenue is around 1.7 billion yuan, compared to just 700 million yuan a year ago, representing an almost qualitative leap in the company's revenue scale within a single year.
Flagship Product Reaches Mass Production, Performance Rivals International Peers
The core engine driving this revenue surge is Moore Threads' flagship, fully-featured GPU for training and inference, the MTTS5000 intelligent computing card.
The announcement explicitly states that this product's performance has reached a market-leading level and that it has achieved mass production. This statement is significant, as within the domestic GPU landscape, very few players can claim both leading performance and mass production capabilities for a flagship product.
More crucially, the company is continuously delivering large-scale clusters based on the MTTS5000, capable of efficiently supporting the training of trillion-parameter large language models. Its computational efficiency is reportedly at an advanced level comparable to that of international, same-generation GPU clusters of equivalent scale.
This indicates that Moore Threads now possesses the confidence to compete directly with international mainstream products in the core AI computing arena of large model training, moving beyond merely being a "good enough" alternative.
Kuae Intelligent Computing Cluster Commercialization Accelerates, Creating a Complete Chip-to-Service Chain
Moore Threads' growth strategy extends beyond merely selling chips. The announcement specifically highlights the accelerated commercialization of its Kuae intelligent computing clusters as a key factor driving rapid revenue growth.
The Kuae cluster is a large-scale AI computing infrastructure product built on top of Moore Threads' GPU hardware. Its accelerated commercialization signifies that the company is successfully creating a complete industrial chain from chip R&D and system integration to computing service delivery. This integrated "hardware + cluster + service" business model not only helps increase the value of individual customer orders but also strengthens customer loyalty, providing a foundation for sustained future revenue.
AI Industry Boom Provides Major Tailwind, Domestic Computing Power Benefits from Policy and Demand
This performance surge by Moore Threads is underpinned by powerful macro tailwinds.
During the first half of 2026, the domestic demand for AI large model training and inference continued to explode. Coupled with external supply constraints on imported high-end GPUs, there has been a significant increase in procurement interest in domestic GPUs from Chinese tech firms, cloud platforms, and government computing projects. By consistently advancing product performance, ensuring stable supply, and building a growing reputation, Moore Threads has captured a considerable market share during this historic window of opportunity.
The announcement's statement that "product performance has been highly recognized by customers and stable supply has been achieved" precisely highlights the two core challenges in the current domestic GPU competition: credible performance and reliable delivery—both areas where Moore Threads is demonstrating continuous progress.
Focus on Fully-Featured GPU Strategy, Continued Enhancement of Cloud-Edge-End Portfolio
Strategically, Moore Threads adheres to a differentiated path focused on "fully-featured GPUs," continually advancing rapid chip architecture iteration and enriching its portfolio of cloud, edge, and end-point products.
A "fully-featured GPU" refers to a chip capable of both graphics rendering and general-purpose computing (GPGPU). Compared to pure inference chips or specialized AI accelerators, fully-featured GPUs offer greater flexibility and a broader range of applications. By steadfastly pursuing this strategy, Moore Threads can simultaneously address diverse scenarios such as cloud-based large model training, edge inference deployment, and end-point applications, securing a wider position within the complete AI infrastructure ecosystem.
Comments