Moore Threads Opens a Hong Kong Listing Route, but Not Yet a Deal

Revenue growth has not erased the operating burden

For the six months ended June 30, Moore Threads reported revenue of RMB 1.736 billion, against RMB 701.8 million a year earlier. Net loss attributable to shareholders narrowed to RMB 11.6 million from RMB 270.9 million.

The company attributed the increase to AI demand, demand for full-function GPUs, the commercialisation of its KUAE computing clusters and stable product supply. Those are the company’s explanations for its sales change, rather than an independently measured account of market share or product performance.

Its cloud-products revenue was RMB 1.692 billion in the half, compared with RMB 39.7 million for edge and terminal products and RMB 3.1 million of other revenue. The result makes the growth story chiefly a cloud and data-centre business story, not a consumer-device one.

The reported loss also needs context. Excluding non-recurring gains and losses, the attributable loss was RMB 150.8 million. Research and development expense rose 38.16% to RMB 769.1 million. The company’s R&D spend was 44.30% of first-half revenue, down from 79.33% a year earlier because revenue grew faster.

Cash moved in the opposite direction. Net cash used in operating activities was RMB 2.169 billion, compared with RMB 1.164 billion in the prior-year half. Moore Threads said the increase reflected operating purchases needed to expand product production. As a fabless company, it designs and sells chips and related systems while outsourcing wafer fabrication, packaging and testing; that avoids owning fabs but does not remove the need to pay suppliers and support product development.

Moore Threads’ company-reported MTT KUAE large-model pretraining platform. Source: Shanghai Stock Exchange.

A domestic GPU claim meets entrenched alternatives

Moore Threads describes its MUSA platform and products as serving AI acceleration, graphics rendering, physical simulation and scientific computing across cloud, edge and terminal systems. It says KUAE can support clusters of 10,000 cards or more. The interim report says the company has released five GPU architectures since its founding, including its Huagang architecture.

Those are product and scale claims from the company, not a like-for-like benchmark against rival systems. Its own risk disclosure makes clear why the comparison should remain cautious: it identifies international leaders, including Nvidia, as having advantages in product iteration, process technology, software ecosystems and supply-chain integration, including cost control. It also flags intensifying domestic competition as policy support draws more capital and talent into China’s GPU field.

The commercial question is therefore broader than whether Moore Threads can sell more cards and clusters. Its own risk disclosure identifies software, supply-chain integration and cost control as competitive constraints. A second listing could broaden the company’s financing options, but the company has not said that is the purpose of this proposal.

What the next disclosure must answer

The immediate next decisions lie with shareholders, regulators and the company’s choice of an issuance window. A later offering document would need to establish the proposed share count, price range, use of proceeds and dilution before the Hong Kong plan can be judged as a capital-raising event rather than a corporate option.

The operating evidence investors need is similarly concrete: whether cloud-products revenue can continue to grow, whether recurring losses shrink, and whether the cash required for production, R&D and supplier commitments moderates. The first-half report shows momentum and a narrower headline loss. It does not yet resolve the economics of sustaining that momentum against established GPU platforms.