HGTech generated RMB 7.8218 billion in revenue during the first six months of 2026, an increase of 2.53% from the same period a year earlier. Its intelligent-manufacturing product group grew much faster, rising 15.09% to RMB 1.9293 billion. HG Laser's metal additive manufacturing activity sits inside that group and is now being presented not merely as a machine business but as a route to copper thermal-management hardware for consumer electronics and artificial-intelligence systems.

Those figures must not be treated as though they describe one business. The RMB 1.9293 billion line covers laser-processing equipment and intelligent production lines as a whole. HGTech does not disclose separate revenue, orders, margin or profit for metal printers, additive services or automated 3DP cells. The filing therefore establishes that additive manufacturing is embedded in a growing industrial segment, but it does not prove that AM was responsible for the group's financial performance.

What the first-half filing actually reports

HGTech, listed in Shenzhen under code 000988, organises itself around optical connections, intelligent sensing and intelligent manufacturing. Group revenue increased from RMB 7.6290 billion to RMB 7.8218 billion. Net profit attributable to shareholders reached RMB 1.1860 billion, 30.17% higher year on year, while basic and diluted earnings per share increased to RMB 1.18.

The headline profit growth requires a second reading. Net profit attributable to shareholders after non-recurring items was RMB 746.98 million, up only 2.42%. The reconciliation includes RMB 471.04 million from changes in the fair value and disposal of financial assets and RMB 100.21 million in government grants recognised during the period. A 30% rise in reported profit therefore does not mean that underlying operations improved by the same percentage.

Cash generation also moved in the opposite direction. Operating cash flow was negative RMB 1.198 billion, compared with negative RMB 481.44 million a year earlier. HGTech attributes this largely to inventory built in response to higher orders and possible raw-material supply disruption. Inventory rose to RMB 4.508 billion at the end of June from RMB 3.581 billion at the end of 2025. That can protect delivery, but it ties up cash and adds demand, price and obsolescence risk.

Research and development spending increased 29.89% to RMB 599.02 million. Overseas revenue more than doubled to RMB 1.648 billion, representing 21.07% of group sales and a 115.67% year-on-year increase. The filing does not identify how much of that international expansion came from additive equipment, so it cannot be used as a proxy for HG Laser's metal-AM exports.

A strong intelligent-manufacturing segment is not a pure AM result

Laser-processing equipment and intelligent production lines contributed 24.66% of group revenue. Sales rose 15.09%, while the associated cost of sales increased slightly faster at 15.50%. Gross margin was 31.86%, down 0.24 percentage points. The numbers describe a growing and profitable activity, but not one in which revenue growth translated into wider product-level gross margin during the period.

This category is much broader than 3D printing. It includes laser cutting, welding, precision processing, automated lines and smart-factory systems. The company discusses new-energy vehicles, shipbuilding, electronics, semiconductors and AI connectors across the segment. Additive manufacturing should consequently be understood as one component of a much larger laser-industrial platform rather than a separately reported division.

The filing nevertheless names tangible AM developments. The 3DP activity is following an 'equipment plus service' model. On the hardware side, HGTech identifies an eight-laser machine, a green-laser additive system and an automated 3DP-Cell line. On the services side, it says batch orders have been won and delivered to industrial customers. It does not provide customer names, volumes, prices, qualification status or the revenue attached to those deliveries.

Why AI hardware and copper cold plates?

HG Laser positions its new green-laser PBF-LB system for two-phase cold plates, consumer electronics and AI hardware. A cold plate collects heat close to a processor or other high-power component and transfers it into a coolant. In a two-phase system, part of that fluid evaporates, allowing latent heat to carry a large thermal load. Channel geometry, flow distribution, pressure containment and leak resistance are therefore critical engineering properties.

The additive value proposition is not that every flat plate becomes cheaper to print. Powder-bed fusion can form branching internal microchannels, thin walls and topology-optimised heat-transfer surfaces in one component. Monolithic production may remove brazed or welded interfaces that add thermal resistance and potential leak paths. Those are credible design advantages, but each cold plate still requires flow, thermal, pressure, fatigue and life-cycle validation before it can be trusted in expensive computing hardware.

The product page names pure copper and copper alloys. Their high thermal and electrical conductivity makes them useful for cooling and power electronics, but those same properties complicate laser processing. Copper reflects much of the near-infrared light used by conventional PBF systems and rapidly removes heat from the melt pool. The consequences can include lack of fusion, porosity, unstable melting, optical back-reflection and a narrow process window.

A shorter green wavelength is therefore a technically grounded approach rather than a label invented for the product launch. Earlier Fraunhofer ILT work used green light to produce high-purity copper at more than 99.8% relative density and 58 megasiemens per metre of specific electrical conductivity. That independently supports the underlying process principle. It does not validate HG Laser's throughput, yield, repeatability or part life.

The manufacturer's performance claims

HG Laser says its green-laser system achieves material absorptivity above 40%. It advertises monolithic microchannels as small as 0.05 mm, water cooling and optical protection intended to limit thermal drift, and AI vision that observes powder recoating. Multiple process-state inputs are claimed to improve equipment utilisation by 10% while supporting stable production of complex copper-alloy parts.

These figures remain manufacturer claims. The public page does not provide the test geometry, measurement method, powder grade, layer thickness, number of builds or independent validation. It is not clear whether 0.05 mm describes a nominal designed feature, a locally measured opening or a channel that remains clear and dimensionally controlled over a useful length. The baseline behind the 10% utilisation improvement is likewise not published.

HG Laser's broader platform also lists stainless steel, titanium and aluminium alloys. Its LSP260E is described as a four-laser system with coordinated scanning, sealed powder handling and central powder delivery for series production. Public material states a 260 mm forming size and a 1,700 mm machine size without fully defining every axis or configuration. Those limited data are insufficient for a complete comparison with competing production systems.

Laser count is not the same as production capacity

Four- and eight-laser architectures can increase nominal build rate, but they also make calibration of overlap zones, optical power balance and thermal conditions across the bed more demanding. In a production case, the useful metrics are accepted-part cycle time, first-pass yield, reusable-powder rate, post-processing labour and the stability of critical properties across positions and builds—not the number of lasers printed on a specification sheet.

Cold plates create an especially demanding acceptance problem. Residual powder must be removed from internal passages; channel openness, wall thickness and roughness must be verified; and the finished component needs pressure, leak, thermal-cycle and corrosion testing. Computed tomography may be necessary for concealed features, but it is expensive at scale. A geometry that can be printed once is not automatically a geometry that can be qualified and produced economically.

Powder safety is also part of the manufacturing system. HG Laser highlights sealed handling and central powder delivery, which may reduce operator exposure and dust-explosion risk. They do not remove the need for material-specific risk assessment, inert-gas controls, extraction, protective equipment, housekeeping and waste procedures. Safety cannot be reduced to one equipment feature.

Strategic advantage: HG Laser is not a stand-alone printer startup

HGTech combines optical-connectivity, sensing, laser-source, automation and intelligent-manufacturing capabilities. Its filing describes internally developed CAM, machine vision, motion control and laser-plus-AI coordination software. A cold-plate production line can require printing, welding, marking, leak testing, traceability and factory data integration. A supplier that can combine several of those stages may possess an integration advantage beyond the printer itself.

That breadth also creates a reporting hazard. Strong results from a diversified group can lend more confidence to a young product line than its evidence warrants. RMB 7.82 billion in group revenue, a 31.86% intelligent-manufacturing gross margin and rapidly rising overseas sales do not demonstrate market acceptance of one green-laser machine. Installed-base data, qualified applications, identified customers and repeatable production results would be needed for that conclusion.

What an industrial buyer should request

A potential customer should commission trials with its own copper grade and a representative cold-plate design. Each run should document powder lot, particle-size distribution, oxygen content, layer thickness, laser parameters, build time, depowdering and post-processing. Density, conductivity, channel dimensions, wall thickness, roughness, flow, pressure drop and thermal resistance should then be measured across multiple builds, not inferred from one demonstration part.

The procurement audit must cover the process as well as the machine. Buyers need the terms for materials and parameter licences, the calibration procedure, validation across multi-laser boundaries, maintenance of powder and filtration systems, regional availability of critical parts and export of layer-by-layer monitoring data. AI inspection becomes valuable only when false alarms, missed defects and the path from signal to manufacturing decision are understood.

The economic model must include more than hardware price and raw powder. It should cover unrecovered material, inert gas, electricity, filters, build plates, stress relief, separation, finishing, CT or other non-destructive inspection, leak testing, rejects and engineering qualification. A monolithic cooling channel creates business value only where better thermal performance, lower mass or shorter development time outweighs the cost of the complete production route.

Verified facts, claims and editorial assessment

It is verified that HGTech generated RMB 7.8218 billion in first-half revenue, that the intelligent-manufacturing product group grew 15.09% to RMB 1.9293 billion and that R&D spending reached RMB 599.02 million. The company filing also identifies eight-laser, green-laser and automated-cell developments and says its additive service activity delivered batch orders.

The 0.05 mm microchannel, above-40% absorption, 10% utilisation improvement and stable mass-production positioning are manufacturer claims. Independent research supports the technical rationale for using green wavelengths in copper PBF-LB, but no public independent benchmark of this HG Laser system was found. HGTech's half-year report is also unaudited, as the document itself states.

LAYERTIMES' assessment is that the important shift is not the group's roughly billion-dollar scale in US-currency terms. It is the narrowing of metal AM toward a specific high-value thermal problem, with machine, material process and automation being offered together. That is a more mature proposition than a generic promise to print any geometry. Commercial proof will still require additive-specific financial disclosure, named production customers and reproducible qualification results.