Hua Hong Semiconductor

Hua Hong Semiconductor (华虹半导体) traces its roots to the ‘908 Project,’ one of China's earliest state-directed semiconductor manufacturing initiatives from the 1990s, and was formally established in Shanghai in 1997. Unlike SMIC, which has spent the last two decades racing toward advanced logic nodes despite export-control ceilings, Hua Hong made a deliberate strategic choice to specialize almost entirely in mature process technology — nodes from 55nm down through 350nm — serving analog, power management, and embedded non-volatile-memory applications. This is not a consolation-prize niche: automotive power ICs, industrial control chips, display driver ICs, and MCUs all depend on exactly this class of manufacturing, and demand for them has proven remarkably durable even as headlines fixate on leading-edge AI logic. That specialization has made Hua Hong the world's sixth-largest foundry by revenue, holding roughly 2.6% of global foundry market share — a real position in a market otherwise dominated by TSMC, Samsung Foundry, and, in the mature-node tier specifically, UMC and GlobalFoundries. Recent full-year revenue reached approximately $2.4 billion, up close to 20% year-over-year, and the company guided Q1 2026 revenue of $650–660 million with gross margin in the 13–15% range — modest by leading-edge foundry standards, but consistent with the thinner margins structurally typical of mature-node manufacturing. Hua Hong is dual-listed, trading on the Hong Kong Stock Exchange (1347.HK) as well as Shanghai's STAR Market, giving it access to both offshore and onshore capital pools as it expands. The defining event of Hua Hong's 2026 was its move to absorb Hua Hong Wuxi, known within the group as ‘Huali,’ in a consolidation deal valued at approximately RMB 8.27 billion. The transaction folds in roughly 38,000 wafers per month of additional 65nm, 55nm, and 40nm logic and specialty-process capacity, directly enlarging Hua Hong's scale precisely in the mature and specialty nodes that remain structurally important for automotive, industrial, and power-semiconductor customers. Analysts framed the deal as part of a broader wave of foundry consolidation among China's mature-node players, mirroring the industry logic that has long driven scale mergers among Western legacy-node foundries. A subtler but important 2026 development is that AI-driven demand has begun tightening supply even in the ‘boring’ parts of the chip market that Hua Hong occupies. As fabs across the industry reallocate wafer capacity toward AI-adjacent products — power-management chips for AI servers, high-current MOSFETs for data-center power delivery, and the analog support silicon every AI system still needs — mature-node capacity has tightened enough that both Hua Hong and SMIC moved to raise prices on 12-inch wafers in 2026. This is a useful corrective to the assumption that the AI boom only stresses cutting-edge fabs: it is squeezing legacy-node supply too, with knock-on effects for automotive and industrial customers who have no substitute source. Hua Hong's equipment supply chain reflects its mature-node focus: because it does not need EUV lithography, it draws on the full universal roster of DUV-era tools — ASML immersion scanners, Applied Materials and Tokyo Electron deposition/etch systems, Lam Research etch tools, and KLA inspection equipment — alongside a growing share of domestic Chinese equipment from NAURA and AMEC as China's toolmaking self-sufficiency drive matures. Output is packaged downstream by OSAT partners including JCET, China's largest packaging and test house. Hua Hong sits in a general China-foundry export-control classification similar to SMIC's, reflecting Washington's broad concern about Chinese semiconductor manufacturing capacity regardless of node, even though Hua Hong's mature-node focus makes it less directly implicated in the advanced-AI-chip fabrication debates that dominate coverage of SMIC. Its near-term trajectory depends less on export-control headlines and more on whether AI-adjacent demand for power and analog silicon continues pulling mature-node utilization higher, and on how smoothly the Huali integration converts into usable additional capacity.