Chinese semiconductor self reliance strategy

What India can learn from China in building semiconductor Supply Chain 1. Policy-Driven Push • 2000s: China was primarily an assembly hub, dependent on foreign ICs. • 2014: Launch of the National IC Industry Investment Fund (Big Fund) → tens of billions of dollars flowed into local fabs, packaging houses, EDA startups, and equipment players. • Post-2019 (Huawei ban): The U.S. tech restrictions accelerated the push for self-reliance, leading to even more subsidies, talent programs, and local procurement mandates. 2. Design Ecosystem • Fabless companies: HiSilicon (Huawei), UNISOC, and newer AI chip startups have thrived, giving China one of the world’s largest fabless ecosystems. • EDA Tools: Still dominated by Synopsys, Cadence, Mentor, but local alternatives like Empyrean and X-Epic are maturing, though they lag behind in cutting-edge nodes. 3. Manufacturing (Foundries & Packaging) • Fabs: SMIC, Hua Hong, Nexchip, YMTC (memory) are key. SMIC has achieved 7 nm-class chips with DUV-based workarounds. • Packaging & Testing: China is very strong here—JCET, Tongfu Microelectronics, Huatian are among global leaders in OSAT (Outsourced Semiconductor Assembly and Testing). • 3D packaging & advanced packaging have become a focus since lithography restrictions limit leading-edge scaling. 4. Materials • Wafers: Still heavily reliant on Japan, Taiwan, South Korea for high-purity 300 mm wafers. However, National Silicon Industry Group (NSIG) and Zhonghuan Semiconductor are scaling up domestic wafer production. • Chemicals & Gases: Rapid expansion of domestic suppliers (e.g., Peric, Huate Gas, Yonghua Chemical). For photoresists and EUV materials, dependence on Japan and Korea persists. 5. Equipment • Historically dominated by ASML, Applied Materials, Tokyo Electron, Lam Research. • Local players like Naura, AMEC, Kingsemi, Hwatsing are catching up in etch, deposition, cleaning, and CMP equipment. • Lithography is the weakest link—SMEE makes i-line and KrF steppers, but far behind ASML’s EUV. 6. Talent and Research • China has built chip design courses in 200+ universities, attracting returnees from Taiwan, US, and Singapore. • Heavy state-backed recruitment (“Thousand Talents” program) helped, though brain drain due to geopolitical tensions is a risk. 7. Current Position • Strengths: • World’s largest semiconductor consumer (~60% of global demand). • Strong in packaging, legacy-node fabs, memory (YMTC in NAND). • Fast-growing domestic equipment and materials base. • Weaknesses: • Still dependent on foreign IP, EDA, lithography, advanced wafers, and some specialty chemicals. • Leading-edge (<5 nm) manufacturing blocked by export controls. ✅ In summary: China has grown a broad, localized supply chain in semiconductors—design, OSAT, some fab capacity, materials, and equipment—but remains constrained at the bleeding edge by foreign choke points, particularly in lithography, EDA, and high-purity materials.

Last week, China barred its major tech companies from buying Nvidia chips. This move received only modest attention in the media, but has implications beyond what’s widely appreciated. Specifically, it signals that China has progressed sufficiently in semiconductors to break away from dependence on advanced chips designed in the U.S., the vast majority of which are manufactured in Taiwan. It also highlights the U.S. vulnerability to possible disruptions in Taiwan at a moment when China is becoming less vulnerable. After the U.S. started restricting AI chip sales to China, China dramatically ramped up its semiconductor research and investment to move toward self-sufficiency. These efforts are starting to bear fruit, and China’s willingness to cut off Nvidia is a strong sign of its faith in its domestic capabilities. For example, the new DeepSeek-R1-Safe model was trained on 1000 Huawei Ascend chips. While individual Ascend chips are significantly less powerful than individual Nvidia or AMD chips, Huawei’s system-level design to orchestrate how a much larger number of chips work together seems to be paying off. For example, Huawei’s CloudMatrix 384 system of 384 chips aims to compete with Nvidia’s GB200, which uses 72 higher-capability chips. Today, U.S. access to advanced semiconductors is heavily dependent on Taiwan’s TSMC, which manufactures the vast majority of advanced chips. Unfortunately, U.S. efforts to ramp up domestic semiconductor manufacturing have been slow. I am encouraged that one fab at the TSMC Arizona facility is operating, but issues of workforce training, culture, licensing and permitting, and the supply chain are still being addressed, and there is still a long road ahead for the U.S. facility to be a viable substitute for Taiwan manufacturing. If China gains independence from Taiwan manufacturing significantly faster than the U.S., this would leave the U.S. much more vulnerable to possible disruptions in Taiwan, whether through natural disasters or man-made events. If manufacturing in Taiwan is disrupted for any reason and Chinese companies end up accounting for a large fraction of global semiconductor manufacturing capabilities, that would also help China gain tremendous geopolitical influence. Despite occasional moments of heightened tensions and large-scale military exercises, Taiwan has been mostly peaceful since the 1960s. This peace has helped the people of Taiwan to prosper and allowed AI to make tremendous advances, built on top of chips made by TSMC. I hope we will find a path to maintaining peace for many decades more. But hope is not a plan. In addition to working to ensure peace, practical work lies ahead to multi-source, build more fabs in more nations, and enhance the resilience of the semiconductor supply chain. Dependence on any single manufacturer invites shortages, price spikes, and stalled innovation the moment something goes sideways. [Original text: https://lnkd.in/gxR48TK8 ]

China’s memory strategy is entering a new phase, and it is unfolding at exactly the moment global supply constraints are reshaping the economics of DRAM and NAND. According to reporting by Nikkei Asia, ChangXin Memory Technologies, Inc. (CXMT) and Yangtze Memory Technologies (YMTC) see the current shortage as a window for “emerging players” to close the gap with Samsung Semiconductor, SK hynix and Micron Technology. What is striking is not just the scale of the investment, but the strategic intent behind it: capacity is being built primarily to secure China’s domestic demand for foundational memory, from servers and PCs to automotive and increasingly AI infrastructure. CXMT’s expansion in Shanghai is designed to dwarf its Hefei operation, with total output expected to be two to three times larger once equipment move-in begins in the second half of 2026 and mass production ramps in 2027. The roadmap extends beyond DDR4 and LPDDR4 into DDR5-grade technology, alongside parallel investment in HBM lines to support China’s growing AI and cloud ecosystem. Yole estimates already place CXMT as the world’s fourth-largest DRAM producer by capacity, with share projected to rise from around 11% today toward nearly 14% by 2027, underpinned by major domestic customers such as Alibaba Cloud and a planned IPO targeting roughly 29.5 billion yuan. YMTC is pursuing a similarly ambitious path. Its third Wuhan fab, due to come online in 2027, is expected to split capacity between NAND and DRAM, while also exploring domestic HBM packaging partnerships. Since being added to the US Entity List in 2022, the company has reoriented around localisation, alternative tooling paths and proprietary development, regaining momentum as China’s policy and market environment increasingly favours domestic supply chains. Market share projections for its NAND business alone point to growth from around 12% in 2025 to 15% by 2028. The deeper implication is structural. If China succeeds in scaling DDR5 and mainstream memory production in the way it previously industrialised silicon manufacturing, it positions itself as a dominant supplier of foundational memory. Commercial logic is already pushing Samsung Semiconductor, SK hynix and Micron Technology further up the value curve toward high-margin, high-performance HBM for AI and advanced compute. A bifurcated market begins to emerge, with China anchoring volume DRAM and NAND at scale, and the incumbent leaders concentrating on the high-end frontier. For the global supply chain, this has long-term consequences for diversity, resilience and pricing power. Capacity, capital and policy are aligning behind memory in the same way they have behind compute and fabrication. The result is not just more chips, but a rebalancing of who controls the foundations of the digital economy. #Semiconductors #MemoryMarkets #AIInfrastructure #SupplyChains #Geopolitics #DDR5 #HBM #ChinaTech

China’s Chip Strategy Is Evolving—Faster Than Expected As someone who serves on the boards of two public semiconductor companies, I found @Liza Lin’s WSJ piece to be clear, rational, and timely. It’s a valuable watch for anyone tracking the future of global semiconductor supply chains and its impact on AI. U.S. export controls aimed at slowing China’s chip development have had a complex impact. One unintended consequence: a renewed push for self-sufficiency. SMIC, despite restrictions, is now producing 7nm chips—technology that powers Huawei’s latest smartphone and was once thought inaccessible without Western semiconductor manufacturing equipment. Beyond the technical achievement, China is investing billions in its semiconductor ecosystem—from equipment and fabrication to talent pipelines. Local firms are shifting procurement strategies, reinforcing domestic capacity. This is more than just a semiconductor story. It cuts across global supply chains, national security, and AI development. A critical inflection point—worth watching closely. https://lnkd.in/gfhvpytF #Semiconductors #AIstrategy #TechSupplyChain #BoardLeadership

The argument for allowing exports of advanced NVIDIA chips is centered on the key assumption that it will keep China ‘addicted’ to US technology. That assumption is deeply flawed and ignores history, Chinese strategy and mercantilist industrial policies, and technological realities. Dmitri Alperovitch lays out this case in his excellent letter to the editor in The Wall Street Journal “Claims that Chinese firms will become dependent on NVIDIA chips and their CUDA architecture ignore what is happening across the tech industry. Every major U.S. hyperscaler—Google, Amazon, Microsoft, Meta—is reducing its reliance on Nvidia. The latest iterations of the two leading frontier models, Claude and Gemini, were reportedly trained largely on Amazon’s Trainium and Google’s TPU chips, respectively, not Nvidia GPUs. If U.S. companies can switch architectures, why would Chinese firms be any different? Rather than grow dependent, China will take Nvidia chips while they are available, use them to train models to compete with American frontier variants and continue to invest heavily in domestic alternatives like Huawei’s Ascend chips. When those are good enough, the firms will drop Nvidia—and quickly. The notion that restrictions “accelerate Beijing’s move toward alternatives” misses a critical reality: Technological self-sufficiency is a Xi Jinping mandate. He isn’t going to allow China to rely on an American tech stack. The Communist Party is already investing in an alternative supply chain and will limit Nvidia imports if needed to ensure sufficient domestic demand for Huawei. The question isn’t whether China pursues self-sufficiency; it’s whether we hand it advanced capabilities during its years-long catch-up period.” #china #technology #innovation #semiconductor #nvidia #gpu #google #amazon #meta #microsoft #cuda https://lnkd.in/ePH2kMFn

China can’t access EUV lithography. Yet its semiconductor firms are posting breakout profits. That contradiction is the real signal. AI demand plus policy pressure is reshaping China’s chip stack faster than expected. This isn’t about catching up overnight. It’s about building viable domestic layers under constraint. One number stands out: +247% revenue growth at a local GPU supplier. BREAKDOWN • GPUs: Cambricon delivered 2.2B yuan profit, its first ever → AI demand is monetizing. • Memory: Flash prices rising 15%+ → supply crunch rewards local storage players. • Equipment: Domestic etch and lithography firms saw 35–84% profit growth → capex stays inside China. Export controls didn’t stop the buildout. They redirected it. China’s semiconductor strategy is no longer just defensive. It’s becoming economically sustainable. Is this the early proof that tech self-reliance can work under sanctions? #Semiconductor #AIInfrastructure #Geopolitics #SupplyChain #ChipManufacturing #ChinaTech #EDA #GPUs

China Orders State-Funded Data Centres to Drop Foreign AI Chips Beijing Tightens Control Over AI Infrastructure China has issued sweeping new guidance mandating that state-funded data centres use only domestically made AI chips, a move that could reshape the global semiconductor market and deepen the technological rift with the United States. Projects less than 30% complete must remove foreign chips, while those further along will face case-by-case reviews. Who’s Hit Hardest Nvidia, AMD, and Intel face sharp revenue losses. Nvidia, which once held 95% of China’s AI chip market, has now fallen to zero, according to company data. The order also applies to Nvidia’s H20 chips—the most advanced models allowed for sale in China—and bars grey-market versions of higher-end processors like the B200 and H200. A $100 billion wave of Chinese data centre projects launched since 2021 could now pivot entirely toward local suppliers. Strategic Goals and Political Context The directive underscores Beijing’s determination to achieve AI chip self-sufficiency, part of its “Made in China” drive to eliminate reliance on U.S. technology. It follows earlier bans on Micron products and prior restrictions discouraging purchases of Nvidia chips for national security reasons. The policy emerged just days after President Trump’s talks with Xi Jinping, in which Trump said Washington would “let them deal with Nvidia—but not the most advanced chips.” Winners and Risks at Home Domestic chipmakers such as Huawei, Cambricon, MetaX, Moore Threads, and Enflame are poised to gain market share. However, the move could slow China’s overall AI progress, as local manufacturers face limitations in high-end fabrication due to ongoing U.S. export controls on chipmaking equipment. Developers, long reliant on Nvidia’s ecosystem, may struggle to migrate software and training infrastructure to Chinese chips. Why It Matters This directive marks one of Beijing’s boldest steps yet toward technological independence. While it strengthens China’s strategic control over AI infrastructure, it also widens the global AI computing divide, potentially shifting the balance of innovation from competition to isolation. I share daily insights with 31,000+ followers and 11,000+ professional contacts across defense, tech, and policy. If this topic resonates, I invite you to connect and continue the conversation. Keith King https://lnkd.in/gHPvUttw

I have become a huge fan of Nikkei Asia. This story encapsulates all the reasons why I am finding far better coverage of global affairs, particularly in matters related to international supply chains, from media outside the Western world. https://lnkd.in/gxFkzXF8 Most articles published here in the US concerning China are either ringing alarms of a perceived threat or celebrating its impending demise, which has apparently been unfolding for over 20 years with no end in sight. Very rarely do I see fair, balanced, detailed, and objective coverage on issues vital to our national security that do not end up highly editorializing into either the China-threat or China-doom narrative. This article from Nikkei Asia is worth at least 50 articles I've read on China's progress in semiconductors from outlets such as WSJ and NY Times. It provides a detailed breakdown of all the different steps in how lithography functions, the Western incumbents, the Chinese startups and competitors, as well as choke points of the technology. It asks an open-ended question, without issuing judgment. In the end, I walked away with an independent sense that China was forced to develop a long list of domestic alternatives due to export controls, and while these domestic alternatives are collectively not yet mounting a challenge to ASML, they are beginning to directly compete against their Western alternatives individually at the global stage. Unsurprisingly, TSMC had to pledge to not use any Chinese suppliers in building its US fabs, implying that emergent Chinese alternatives are already replacing their Western counterparts in TSMC's fabs elsewhere. I hope you all enjoy this article as much as I did.