China’s sweeping campaign to build a self-sufficient semiconductor ecosystem is translating into massive financial windfalls for domestic hardware designers. Cambricon Technologies, the country’s leading artificial intelligence chipmaker, reported a 108 percent surge in first-half revenue, capitalizing on a nationwide rush to replace restricted foreign technology. The financial results underscore how U.S. export controls, intended to constrain China’s AI development, have inadvertently accelerated the commercial maturation of its domestic champions.
According to a stock exchange filing cited by the South China Morning Post, Cambricon’s revenue for the first six months of 2026 reached 6 billion yuan ($890 million), up from 2.88 billion yuan during the same period last year. The top-line growth was accompanied by an even steeper rise in profitability, with net profits jumping 122.6 percent year-on-year to 2.3 billion yuan. The momentum held steady through the second quarter, where the firm posted 3.1 billion yuan in revenue, perfectly aligning with the 3 billion yuan consensus estimate from a Bloomberg poll, and generated 1.3 billion yuan in net profit.
The Rush for Domestic Replacements
Cambricon attributed its explosive growth to a steady, unrelenting increase in demand for AI computing power across China. The company noted in its filing that it had leveraged its core competitiveness to deepen cooperation with leading enterprises in the finance and internet sectors. This demand is being artificially supercharged by geopolitical friction. With Washington largely cutting off China’s access to advanced AI accelerators from industry leaders like Nvidia and Advanced Micro Devices, domestic tech giants have no choice but to turn inward.
The shift in procurement strategy is stark. According to a June 2026 survey of Chinese technology executives by Bloomberg Intelligence, companies expect to spend 46 percent of their AI accelerator budgets on locally made chips over the next 12 months, a significant jump from the current 30 percent allocation. The survey explicitly noted that AI accelerators produced by Cambricon and Hygon Information Technology were being evaluated by a large pool of respondents. As Chinese customers transition from training massive frontier models to integrating AI into enterprise applications, the economic calculus of deployment is changing.
(Related: Huawei’s Ascend 910C Demand Surges 20% as DeepSeek V4 Pushes to Break CUDA Dependence)
A Broader Sector Boom
Cambricon is not the only beneficiary of this state-mandated import substitution drive. The broader Chinese AI chip sector is experiencing a rising tide. Up-and-coming competitors like Shanghai Iluvatar CoreX Semiconductor are expected to see their sales triple in the first half of the year, while Beijing-based Moore Threads Technology recently projected its revenue for the same period would grow by as much as 149 percent. Even with the U.S. clearing sales of Nvidia’s tailored H200 chips to China, American officials acknowledged last month that the actual volume of processors shipped has been “trivial.”
The consensus among industry analysts is that Chinese chips, while still trailing Nvidia in single-unit peak performance, are becoming viable alternatives through scale and aggressive pricing. “Purchasing decisions are increasingly driven by deployment economics rather than absolute peak silicon performance,” Morgan Stanley analyst Charlie Chan wrote in a recent research note. Chinese chipmakers are compensating for the performance gap by linking tens of thousands of chips together in massive server racks, prioritizing total system throughput over individual processor speed.
The Long-Term Infrastructure Play
The financial boom for companies like Cambricon is secured by Beijing’s ambitious, state-directed infrastructure plans. China is currently executing a massive data center buildout, aiming to spend 2 trillion yuan ($295 billion) over the next five years. This project is designed to rely on local suppliers for at least 80 percent of its core technology, including AI chips.
This guaranteed pipeline of state and enterprise demand is structurally altering the market. Morgan Stanley now estimates that China’s self-sufficiency level in AI chips will reach 70 percent by the end of the decade, a dramatic increase from 42 percent in 2025. “Most people would privately agree that having access to Nvidia would always be ideal, but in the absence of that, the trend I’m seeing is more Chinese chips being used,” Morningstar analyst Phelix Lee told Bloomberg. For Cambricon, this trend translates directly into the triple-digit growth reported this week, transforming the company from a promising designer into a cornerstone of China’s technological independence.
(Related: China’s AI Chip Self-Sufficiency Reaches 41%, Morgan Stanley Projects 76% by 2030)
Cambricon’s Product Roadmap and Competitive Positioning
Beyond the headline earnings figures, Cambricon’s strong first-half performance reflects a deliberate product strategy focused on the inference market rather than the training market. While Nvidia’s dominance in high-end training accelerators remains largely intact for now, the inference segment, where models are deployed to serve end users, is where Chinese chipmakers are finding the most traction. Inference workloads are less computationally intensive than training, making them more amenable to the current generation of Chinese chips, and they represent the fastest-growing segment of the AI hardware market as enterprise adoption accelerates.
Cambricon has positioned its MLU series of AI accelerators specifically for inference deployment in data centers, targeting the finance, internet, and cloud computing sectors that have been most aggressive in adopting domestic alternatives to Nvidia. The company’s ability to offer competitive token-generation costs, combined with deep integration with Chinese software frameworks and strong after-sales support, has made it a preferred supplier for domestic cloud providers building out AI inference infrastructure.
As the Chinese AI ecosystem matures and the volume of inference workloads grows exponentially, Cambricon’s strategic positioning in this segment should continue to generate strong revenue growth well into the second half of the decade.