←News detail

Jiwei.comUpdated Oct 28 15:58

Huaqin Technology's Q3 results hit a record high, with batch delivery of Super Nodes driving an explosive surge in computing power business

AI summary

Huaqin Technology's Q3 2026 single-quarter revenue grew 60.86% year-on-year and 37% quarter-on-quarter, with non-recurring net profit of 1.19 billion yuan; both revenue and non-recurring profit hit record highs for a single quarter. Super Node product revenue in Q3 was 6.191 billion yuan, with monthly shipments already exceeding one hundred racks, driving data business Q3 revenue to 22.456 billion yuan, up 228% year-on-year; full-year Super Node revenue is expected to exceed 10 billion yuan. In the first three quarters, computing and data business revenue was approximately 82.706 billion yuan, surpassing the mobile terminal business's 68.784 billion yuan for the first time.

Body · AI translation

On the evening of October 27, Huaqin Technology disclosed its 2026 third-quarter report and simultaneously held an earnings call. In the first three quarters, the company achieved operating revenue of 166.014 billion yuan, up 28.81% year on year; net profit attributable to shareholders of 4.323 billion yuan, up 39.53% year on year; and net cash flow from operating activities of 781 million yuan. Alongside steady growth in revenue and profit, operating cash flow continued the positive trend of turning positive in the first half, with its "cash generation" capability continuously strengthened; combined with the supply chain scale and cost control advantages supported by the world's largest smart hardware ODM platform, this has built a solid financial foundation for the scaled expansion of new computing power businesses.

On a quarterly basis, Q3 growth momentum became more pronounced: single-quarter revenue rose 60.86% year on year and 37% quarter on quarter, while net profit excluding non-recurring gains and losses reached 1.19 billion yuan, up 22% year on year and 39% quarter on quarter. Both revenue and non-recurring net profit hit record highs for a single quarter, fully confirming the company's operational resilience from multi-category coordinated development and continuous product mix upgrading.

At the earnings call held the same day, management further clarified the ramp-up pace of the super node business. The product achieved first delivery to a major customer in the second quarter and entered the mass delivery stage from the third quarter, with shipments rising rapidly quarter by quarter. Current monthly shipments have reached more than 100 racks, and super node revenue in the quarter reached 6.191 billion yuan, with full-year revenue expected to exceed 10 billion yuan. The clear delivery milestones and the third-quarter results corroborate each other, marking the smooth fulfillment of the first-phase strategic goals of Huaqin's transformation.

Leveraging the industry's scarce full-stack system capability of "computing + networking," Huaqin is leaping from a globally leading consumer electronics ODM leader to a rare domestic provider of computing infrastructure and AI hardware system solutions, seizing the early-mover high ground in the window of change in domestic computing power architecture.

Mass delivery of super nodes lands, building a unique system-level moat

Domestic AI computing power is undergoing a paradigm shift in underlying architecture. As large models evolve toward MoE sparse architectures, the shortcomings of traditional scale-out single-machine stacking solutions in communication latency, computing utilization, and cluster TCO continue to amplify. Computing power competition has moved beyond comparisons of single cards or single machines to system-level competition integrating computing, high-speed networking, high-power power supply, and liquid cooling. Scale-up super nodes with high interconnect density and low latency have become the preferred route for training and inference clusters of leading cloud vendors; the value of pure server assembly continues to dilute, and integrated in-house design capability of "computing + networking" has become the scarcest core barrier in the current computing power track.

Huaqin is one of the few players in the industry to have closed the in-house loop between computing nodes and switching nodes, and its ramp-up pace precisely hits the main line of this round of computing power upgrading. Data business revenue in the first half was approximately 20 billion yuan, under year-on-year pressure at one stage due to the switching of domestic GPU platforms, but it improved quarter by quarter, with resilience standing out; in the second quarter it completed first delivery of super nodes to a major customer, and in the third quarter it officially entered mass delivery. Q3 data business revenue reached 22.456 billion yuan, up 228% year on year, of which super node products contributed 6.191 billion yuan in revenue, with a steep ramp-up slope. Looking ahead to the full year, data business revenue is expected to exceed 60 billion yuan, up 50% year on year, with strong growth momentum; of this, super nodes will contribute more than 10 billion yuan in revenue in the second half, and in-house switches will generate about 6 billion yuan in full-year revenue, achieving double-digit growth, with clear ramp-up paths for the two core computing power products. In 2027, the two core products will still maintain rapid growth and continue to serve as the core engines of the company's performance growth, further opening up the medium- and long-term growth space of the computing power business.

Huaqin's scarcity lies not in single-point hardware manufacturing, but in the complete closed-loop system capability. The vast majority of domestic system integrators can only take on computing-side server orders, while switching nodes, as the segment with the highest barriers and strongest bargaining power in computing power clusters, have long been monopolized by a few vendors. As the lead unit of the ODCC ETH-X super node rack specification, the company leads the formulation of next-generation computing power architecture standards such as in-rack orthogonal backplanes and inter-rack optical interconnection, holding the discourse power over industry rules; at the same time, it has made forward-looking arrangements in NPO near-packaged optics technology, and super node products based on NPO solutions have entered the verification stage with leading customers, continuously positioning itself for next-generation high-speed interconnection. "In-house hardware development—standard definition—coordinated design of computing and networking" forms a complete closed loop, and on the supply side it has built a dual-backup platform for overseas and domestic switch chips, which can flexibly adapt to the differentiated technical routes of different cloud vendors and government and enterprise customers, with supply chain risk resistance significantly better than peers.

Third-party industry assessments also corroborate the value of Huaqin's early-mover positioning. Gartner's "Hype Cycle for Infrastructure Strategy in China, 2026" lists "China AI Super Node" as a disruptive technology, and Huaqin's computing power platform EVEX Yuantu, together with Alibaba Cloud and Huawei, was selected in the list of benchmark vendor examples. Currently this technology is still in the embryonic stage, with industry penetration below 1%, and is expected to enter a stage of large-scale adoption in 2–5 years. 2026 is precisely the critical window for the industry landscape to take shape, and Huaqin's early achievement of mass production is intended to lock in system solution premiums on the eve of the industry explosion in 2027–2028 and establish a first-mover advantage that is difficult to catch up with.

The stable delivery of super node solutions stems from the deep synergy of multiple layers of core capabilities. With advanced computing architecture and full-link liquid cooling design as the foundation, relying on a rich product matrix covering multiple cards and multiple U form factors, combined with L12-level complete machine system solution capability, and with delivery guaranteed by its own large-scale full-capacity production capacity, the company ultimately forms an integrated closed loop of "solution definition—verification—mass production delivery," building the core competitiveness for stable, high-quality delivery of the super node business.

In terms of profitability, the ROI within super nodes is significantly higher than that of traditional AI servers. As shipment scale continues to rise, there is room for gross margin and net margin to improve beyond peers. To address the industry-wide challenge of periodic tight supply of switch chips, Huaqin, on the one hand, has established priority supply guarantee mechanisms with leading chip manufacturers and core cloud customers; on the other hand, it has strategically stockpiled key materials such as switch chips and memory in advance, effectively hedging supply chain fluctuation risks. The customer structure continues to optimize, with the three leading CSPs all expected to contribute revenue on the scale of ten billion yuan for the full year, and breakthroughs in the government-enterprise and carrier markets are accelerating; the company aims to rank in the first tier of the super node supply chains of leading cloud vendors and grow into a core system supplier for domestic data centers.

Revenue structure makes a historic shift, and the data business enters a high-speed growth channel

In the first three quarters of 2026, Huaqin's computing and data business revenue was approximately 82.706 billion yuan, historically surpassing the mobile terminal business (68.784 billion yuan) that it had cultivated for 21 years. It is worth emphasizing that this structural shift is not a passive yielding under the contraction of the traditional main business—during the same period, the mobile terminal business still maintained steady growth of 15.6%, and the company's overall foundation remains solid; the essence of the surpassing is a landmark inflection point marking the company's proactive grasp of the computing power wave and the rise of its second growth curve.

The third-quarter report is the core window for verifying the ability to deliver on this computing power main line. Following the first delivery to a major super node customer in the second quarter, the data center business entered the fast lane of growth across the board in the third quarter, with single-quarter revenue reaching 22.456 billion yuan, a year-on-year increase of 228%, showing a steep ramp-up slope; combined with the computing business, single-quarter revenue from the computing and data business was approximately 39.4 billion yuan, accounting for more than 54% of revenue. A single-quarter revenue structure in which this segment accounts for more than half means that the computing power business has substantively taken over as the company's growth engine and is a core indicator for observing the implementation of full-year performance guidance and the improvement of computing power profitability.

The valuation logic is subsequently being reconstructed. In the past, the market applied a consumer electronics ODM valuation system, assigning a 10–15x PE; currently, the market is gradually switching to a sum-of-the-parts valuation framework, with the computing power business benchmarked against leading server and data communications equipment companies, handsets, PCs, and AIoT forming the stable cash flow foundation, and automotive electronics and robotics serving as long-term growth options. As the data business's full-year high growth rate of 50% is significantly higher than that of the group as a whole, the weight of the computing power segment in the 2027 revenue structure will continue to increase, driving the valuation center upward.

The terminal segment captures structural dividends under the AI wave, forming a high-quality combination of "upward computing power elasticity + terminal cash flow support." The consumer electronics market as a whole is under pressure, but AI PCs, AI phones, smart wearables, and AIoT are opening up incremental tracks, ODM industry concentration continues to shift toward leaders, and Huaqin continues to capture share by virtue of its advantages in scale, supply chain, and R&D delivery.

Mobile terminals are the ballast of the group's steady cash flow. ODM shipments in the three major categories of phones, tablets, and wearables all rank first globally, and continuous product structure upgrading drives steady revenue growth, with full-year revenue expected to rise from more than 80 billion yuan last year to more than 90 billion yuan. The PC business is a benchmark for the company's cross-track capability migration, with first-half revenue growing more than 50% year on year, a full-year shipment target of 20 million units, and share continuing to penetrate leading US and Taiwanese customers; relying on the integrated Nanchang base, it achieves efficient coordination of R&D, manufacturing, delivery, and supply chain, with outstanding cost and delivery advantages.

AIoT and innovation businesses both grow rapidly, and the second-tier growth poles are accelerating into shape. AIoT is a high-quality segment with high growth and high gross margin, achieving revenue of 9.196 billion yuan in the first three quarters, a year-on-year increase of 68.4%, with a clear trend of simultaneous volume and profit growth. Innovation businesses are being cultivated in an orderly manner and gradually scaling up, achieving revenue of 4.120 billion yuan in the first three quarters, more than doubling year on year with growth of 119.3%. Among them, automotive electronics has a full-year target of 2.5–3 billion yuan, plans to reach the ten-billion-yuan scale and achieve profitability by 2028, and is deepening its presence in the Japanese automotive market through a joint venture with Toyota Boshoku, opening a channel to overseas high-end customers; in the robotics business, shipments of data collection products exceed 400 units, and more than 4,000 units have been manufactured on an OEM basis for leading manufacturers. Its self-developed wheeled dual-arm robot has completed validation on internal production lines, opening up an upgrade path "from OEM to self-developed complete machines" and securing an early position in the humanoid robot industry wave.

More critical than the leap in revenue scale is the comprehensive upgrade of the cooperative value chain. The delivery vehicle has upgraded from motherboards and complete terminal devices to self-developed switches and super node full-rack systems; the cooperation model has upgraded from traditional price-comparison bidding to deep JDM joint development with leading cloud vendors, participating upfront in the definition of computing power architecture; the pricing model has moved away from simple cost-plus pricing to value-based pricing centered on project ROI, relying on unique complete-machine system engineering capabilities to obtain solution premiums, with gross profit per super node rack significantly higher than that of a single AI server.

High-intensity R&D continues to provide the foundation, and the A+H capital platform consolidates the confidence for long-term expansion

The computing power track has the dual attributes of being technology-intensive and capital-intensive. Undertaking continuously surging orders and rapidly iterating system solutions cannot be achieved without upfront R&D investment, smooth and diversified capital channels, and continuously recovering operating cash flow.

On the R&D side, the company maintains a long-termist approach and a high-investment pace. From 2023 to 2025, cumulative R&D investment reached 16.1 billion yuan; in the first three quarters of 2026, R&D investment was 5.66 billion yuan, up 23% year on year, of which R&D investment related to super nodes has maintained a high growth rate of 50% for three consecutive years. The company adheres to the iterative rhythm of "mass-producing one generation, developing one generation, and pre-researching one generation," and continues to deepen underlying core technologies such as high-speed interconnect, high-power power supply, and liquid cooling. At the same time, it completed two rounds of share repurchases totaling nearly 600 million yuan, all used for equity incentives, deeply binding core R&D backbones and technical teams, and closely tying talent returns to the main line of computing power transformation.

On the capital side, it has built an A+H dual financing platform, opening up long-term capital space. In April 2026, it listed on the Hong Kong Stock Exchange (03296.HK), raising approximately HK$4.549 billion, with the public offering oversubscribed 531 times; cornerstone investors cover long-term financial funds and core capital in the industrial chain, linking upstream resources such as wafers, PCBs, copper-clad laminates, and sensors. Beyond financing, it is simultaneously building a solid industrial collaboration ecosystem.

Operating cash generation capability has seen a substantive reversal. In the first half of 2025, net operating cash flow was -1.522 billion yuan; in the first half of 2026, it turned positive to 643 million yuan, an improvement of more than 2.1 billion yuan; in the first three quarters, net cash flow from operating activities was 781 million yuan. For the relatively low-margin ODM model, positive cash flow means that scale expansion already has self-generated cash capability and no longer relies heavily on external financing for support, and the company's overall resilience against cycles has significantly strengthened.

The continued repair of cash flow is also corroborated by profitability quality indicators. The average ROE for the three years from 2023 to 2025 was 15.7%; in the first three quarters of 2026, annualized ROE reached 19.3%, up 1.9 percentage points year on year, continuing to maintain a relatively high level of profit returns. In the medium to long term, as the proportion of high-value-added products such as super nodes, self-developed switches, and commercial PCs continues to rise, combined with economies of scale diluting fixed costs, the group's gross margin has clear room for upward movement, and ROE is expected to operate stably within the high-quality range of 15%–20% indicated by management's long-term target.

Conclusion

Computing power infrastructure is moving from standardized standalone machines to highly integrated complete-rack system solutions. The focus of industry competition is shifting from manufacturing and delivery to comprehensive system engineering capabilities, and the barriers in the track have risen substantially. Unlike most manufacturers that are limited to server assembly, Huaqin relies on its first-mover mass-production advantage in super nodes, full-stack self-development of computing and networking, standard-setting voice, and its own production capacity to build a scarce system-level moat.

Following the implementation path of domestic GPU adaptation, first launch of super nodes, and volume supply, Huaqin has precisely seized the golden window for reshaping the computing power landscape. The consumer electronics terminal business solidifies the cash flow foundation, while the A+H capital platform and sustained high-intensity R&D further amplify competitive advantages; the three major cycles of continued release of computing power capital expenditure, the spread of on-device AI, and automotive intelligence resonate, opening up the long-term growth ceiling. The 2026 third-quarter report has become an important milestone for Huaqin's new computing power business as it shifts from technology implementation to performance realization.

Source

Jiwei.com