Report: MetaX Sales Jump More Than Fourfold as Nvidia Faces Curbs

Three Chinese AI-chip makers moved into profit in the third quarter, according to the South China Morning Post, as restrictions on Nvidia processors opened a market for domestic alternatives. ^[1]

The November 6 report named Moore Threads, MetaX and Iluvatar CoreX among companies making the turn. They are competing for Chinese enterprise customers and major cloud-service providers seeking computing power that US export controls have made harder to buy from Nvidia. ^[1]

The commercial stakes go beyond winning orders. China's domestic chip developers need to show that they can supply customers at prices that cover their costs, then keep doing so as production grows. The Post said soaring memory costs and other supply-chain hurdles threaten long-term growth, and that major investment banks have cut their margin forecasts. ^[1]

The three-company profit claim remains unconfirmed in detail. Separate third-quarter figures and comparable accounting measures for Moore Threads and Iluvatar CoreX have not been established, nor has the contribution of core product sales to each company's profit. The size of the memory-cost increases and the banks' forecast cuts is also not known. ^[1]

MetaX offers the clearest reported financial example. According to BigGo Finance, its third-quarter revenue reached 1.5 billion yuan, up 367.12% from a year earlier, while adjusted net profit attributable to shareholders was 308 million yuan. The outlet described it as MetaX's second consecutive quarter of adjusted profitability. ^[2]

That is a substantial sales increase: quarterly revenue was more than four times the year-earlier level. It gives the domestic-chip push a commercial result to examine, rather than just a promise of future demand. ^[2]

MetaX puts numbers behind the shift

BigGo Finance reported MetaX's revenue for the first nine months at 2.824 billion yuan. Net profit attributable to shareholders was 639 million yuan, while adjusted net profit was 259 million yuan. Both profit measures were positive after losses a year earlier, according to the outlet. ^[2]

The third quarter therefore accounted for just over half of the company's reported nine-month revenue. That concentration shows how heavily the latest quarter contributed to the year's sales. ^[2]

The gap between the two nine-month profit measures deserves attention. Adjusted earnings remove selected items from the headline calculation, so they answer a different accounting question. The reported 639 million yuan in net profit and 259 million yuan in adjusted profit should not be treated as interchangeable. ^[2]

For the third quarter, the reported adjusted profit was equivalent to about 20.5% of revenue, calculated from BigGo Finance's figures. That gives a sense of the earnings relative to sales, but it is an adjusted measure, not a substitute for a comparable unadjusted result across all three companies. ^[2]

MetaX's move into profit also predates the latest quarter. In August, the Post reported that the company had earned 612 million yuan in net profit during the first half of 2026, reversing a loss of 186 million yuan a year earlier. The same account described sharply different financial results across Chinese chipmakers. ^[3]

The emerging picture is of individual businesses finding buyers at different speeds. It does not support treating domestic AI chips as a single industry moving together into profit.

Orders still have to become working systems

The products involved are GPUs, processors that carry out many calculations at once. They underpin much of the work involved in training AI models and running them after training. For enterprise and cloud customers, the purchase is ultimately about usable computing capacity.

BigGo Finance reported that MetaX's flagship Xiyun C600 entered mass production in May and that a computing cluster using 10,000 cards began operating in September. The cards are boards carrying the processors, linked together in a cluster to work as one system. Those reported milestones put hardware production and deployment alongside the financial gains. ^[2]

A cluster of that size makes the business challenge concrete. Selling processors is one step. Supplying and supporting systems that customers can keep using is how a chipmaker earns repeat business.

US restrictions give domestic suppliers an opening with customers who cannot readily obtain Nvidia's high-end processors. ^[1] The opportunity is real even without a claim that these companies have matched Nvidia across performance, software or cost. Customers with fewer purchasing options still need machines that work.

Memory puts the profits under pressure

Memory is a key component in systems built around AI processors. When its price rises, more of the money from a sale goes towards building the system, leaving less as profit unless the supplier can charge more or reduce other costs.

That is the pressure identified by the Post, alongside broader supply-chain difficulties and lower bank forecasts for margins. ^[1] Strong demand can help fill an order book. It cannot guarantee that each order will remain profitable when components become more expensive.

For China's emerging Nvidia alternatives, winning a customer is the opening transaction. Keeping a margin is the business.