China’s AI chip industry has gained another financial marker as MetaX Integrated Circuits, known in some English reporting as Muxi, reported its first profitable half-year. The Shanghai company posted 1.324 billion yuan in revenue for the first six months of 2026, up 44.67% from a year earlier, and recorded a net profit of 612 million yuan after a 186 million yuan loss in the first half of 2025. Sina Tech, republishing an IT Home report on the company’s half-year disclosure, published the figures on August 30.
The results matter because MetaX is one of a group of Chinese chip companies trying to turn strategic demand for domestic AI computing into a sustainable hardware business. The company also said its Xiyun C600 graphics-processing unit reached mass production in May. Together, the financial shift and production update offer a more concrete signal than a product announcement alone: a domestic accelerator maker says it is now producing hardware at scale while reporting a move into profitability.
The secondary report includes an AI-generated-content notice and advises readers to refer to the official securities-disclosure system. The financial figures should therefore be treated as MetaX’s reported filing results, not independently audited market estimates. Even with that qualification, the change from loss to profit is notable in an industry where research costs, manufacturing access and software development can weigh heavily on young chip designers.
MetaX was mentioned as a related company in EastFrontier’s recent article about Mucang’s Series B and AI-cluster interconnects. Today’s news is different. It is a direct financial and production update from a company attempting to supply GPUs into China’s expanding AI-computing market.
MetaX’s First-Half Results Mark a Turn From Loss to Profit
MetaX reported H1 2026 revenue of 1.324 billion yuan and a net profit of 612 million yuan. In the same period one year earlier, it reported a 186 million yuan loss. These numbers come from the company’s half-year report as summarized by Sina Tech and IT Home, and they should be understood within that disclosure context.
The company also reported 525 million yuan of research-and-development spending, equal to 39.65% of its revenue for the period. That ratio illustrates the capital intensity of an AI chip business. A domestic GPU maker must invest not only in the chip design but also in software tools, driver support, compatibility work and the engineering required to operate in data-center environments.
Profitability does not mean those costs disappear. It means MetaX’s reported revenue and margins were sufficient to cover them in the first half. The durability of that improvement will depend on future orders, manufacturing yields, customer adoption and the expense of continuing product development. The current filing does not provide a guarantee about future periods.
Still, the numbers offer a useful contrast with the earlier stage of many Chinese AI chip startups, where products can spend years in development before reaching paying customers. MetaX’s reported shift suggests that demand for domestic computing products is becoming material enough to affect company financials, not only policy discussion.
The Xiyun C600 Enters Production as Compute Demand Rises
MetaX said its Xiyun C600 GPU reached mass production in May 2026. The report does not provide a full technical specification or a customer list, so those details should not be inferred. What is confirmed is the production milestone, which matters because chip companies face a major divide between a prototype and a product that can be manufactured and supplied in volume.
A GPU is only part of the system a customer needs. AI training and inference workloads also depend on memory, networking, storage, compilers, frameworks and cluster-management software. That is why MetaX’s prospects will be judged by practical deployment as well as unit production. Customers need to know whether a chip works with their model stack and whether it can be supported at the scale their workloads require.
China’s current demand environment gives domestic suppliers an opening because local AI developers, cloud providers and research groups are seeking more compute while high-end cards and large clusters remain scarce, as Caixin reports. EastFrontier’s coverage of China’s 2.45 million PFLOPS of AI capacity shows why the market is increasingly focused on clusters that can train and refine models, not just on individual chips that can run inference.
MetaX’s C600 milestone will therefore be viewed in relation to that wider need. Mass production can improve availability, but it does not automatically establish performance or software maturity for every workload. The commercial test is whether customers can integrate the GPU into useful systems without prohibitive porting, maintenance or reliability costs.
China’s Domestic GPU Market Is Moving Toward Financial Proof
MetaX’s half-year results arrive as China tries to build a more complete AI-computing stack. National policy and local investment have supported chip design, data centers and domestic model development. The next phase requires suppliers to show that their products can generate recurring commercial demand.
The company’s reported revenue growth and move into profit are therefore relevant beyond one balance sheet. They offer an indication that at least some domestic suppliers may be moving from a research-and-fundraising phase toward a period in which manufacturing and customer deployment determine success. That transition is crucial because AI chips require continuing investment even after an initial product reaches the market.
The broader market remains competitive. Domestic companies must contend with established global chip firms, other Chinese accelerator designers and customers that have already built software around particular hardware ecosystems. A shortage of high-end foreign hardware can create opportunity, but it can also create pressure to deliver products quickly while maintaining reliability.
MetaX’s R&D spending underscores that challenge. Spending 525 million yuan in six months is a large commitment for a company reporting 1.324 billion yuan of revenue. It suggests that the firm is trying to keep advancing its product line while serving current demand. The key question is whether that investment produces software and hardware advantages that remain valuable when supply conditions change.
For China’s AI sector, profitable domestic chip companies would carry strategic significance. They could provide more options for cloud services, model developers, and enterprises that are reluctant to depend entirely on overseas technology. But financial success alone will not establish technological independence. The ecosystem also needs production capacity, engineering talent, interoperable software and large-scale customers willing to run critical workloads on local platforms.
MetaX’s filing does not resolve those questions, but it supplies a concrete benchmark. The company says it produced the C600 at scale in May, increased first-half revenue by 44.67%, and moved from a 186 million yuan loss to 612 million yuan in profit. In a market often dominated by promises about future chips, those reported results give observers something more tangible to assess: whether a Chinese AI hardware maker can convert compute scarcity into a working commercial business.