Chinese AI Chip Maker Biren Technology Plans $623 Million Hong Kong IPO: Inside China’s Billion-Dollar Bet on Semiconductor Self-Sufficiency

On December 23, 2025, Shanghai Biren Technology began bookbuilding for a Hong Kong initial public offering aimed at raising up to $624 million, positioning itself to become the first mainland GPU developer to list in Hong Kong when trading begins January 2, 2026. The offering—pricing 247.7 million shares between HK$17 and HK$19.60—has already secured commitments from 23 cornerstone investors totaling $372.5 million, including heavyweight firms like Qiming Venture Partners, Ping An Group, UBS, and Digital China.

This IPO isn’t merely another technology listing. It represents a critical test of China’s strategy to build semiconductor self-sufficiency despite—or perhaps because of—American export controls that have systematically targeted Chinese AI chip development. Biren joins Moore Threads Technology and MetaX Integrated Circuits, which recently debuted on Shanghai’s STAR Market with first-day gains of 425% and 693% respectively, as part of what industry observers call China’s “four little dragons” in the GPU sector: companies racing to challenge Nvidia’s dominance in AI accelerators.

The timing is no accident. As US export restrictions tighten, China has doubled down on domestic alternatives, channeling tens of billions through its National Integrated Circuit Industry Investment Fund (the “Big Fund”) and prioritizing semiconductor independence as national strategic imperative. Biren’s journey from founding in 2019 to this $624 million IPO—while operating at massive losses and navigating US Entity List sanctions—illustrates both the extraordinary challenges and the political will driving China’s semiconductor ambitions.

This analysis examines Biren Technology’s IPO in context: the company’s technical capabilities and financial realities, the regulatory gauntlet of US export controls it has navigated, the remarkable investor enthusiasm for Chinese GPU makers despite consistent losses, China’s multi-hundred-billion-dollar semiconductor self-sufficiency strategy, and the geopolitical implications of a bifurcated global chip industry where both superpowers lose, just differently.

The Four Little Dragons: China’s Answer to Nvidia

China’s push to develop indigenous GPU capabilities has coalesced around four companies, each founded by former employees of American semiconductor giants, each backed by state-linked capital, and each now pursuing public listings to fund expansion despite operating losses. Together they represent China’s most credible attempt to break Nvidia’s stranglehold on AI computing.

Moore Threads: “China’s Little Nvidia”

Beijing-based Moore Threads was founded in 2020 by Zhang Jianzhong, former general manager of Nvidia’s China business, who left to build China’s answer to his former employer. The company’s founding team includes multiple former Nvidia engineers, giving it deep institutional knowledge of GPU architecture and software ecosystems.

Technical Achievements: Moore Threads has launched four GPU architectures—Sudi (2021), Chunxiao (2022), Quyuan (2023), and Pinghu (2024)—demonstrating rapid iteration. The company claims its GPUs match performance of Nvidia’s RTX 3090 and RTX 4090 gaming chips (older generation consumer cards, not Nvidia’s AI-focused H100). Moore Threads announced in 2025 it can run DeepSeek and Qwen AI models on its GPUs, a critical validation of software compatibility.

Recent Product Launch: On December 21, 2025, Moore Threads unveiled its latest architectures—Huashan for AI workloads and Lushan for gaming—promising what the company describes as substantial performance improvements over previous generations. The Huashan chip reportedly targets Nvidia’s Hopper architecture, though independent benchmarks remain unavailable.

IPO Performance: Moore Threads raised 8 billion yuan ($1.1 billion) in its Shanghai IPO and saw shares surge 425% on the December 5, 2025 trading debut, becoming the second-best performing Chinese IPO of 2025 (after MetaX). The company attracted 4.82 million retail subscriber applications.

Financial Reality: Despite market enthusiasm, Moore Threads posted 701 million yuan in revenue against 271 million yuan in losses in H1 2025. The company remains deeply unprofitable, burning cash on R&D, with most of its $1.1 billion in fresh capital likely to earn interest rather than immediately fund chip development.

US Sanctions: Added to the US Entity List in October 2023, Moore Threads lost access to TSMC’s advanced manufacturing nodes, forcing it to redesign chips for less advanced processes—a significant technical setback but one the company has managed to work around through domestic alternatives and design modifications.

MetaX Integrated Circuits: The AMD Lineage

Shanghai-based MetaX was founded in 2020 by Chen Weiliang, former senior director at AMD, alongside CTOs Peng Li and Yang Jian, also AMD alumni. This founding team brings deep expertise in GPU design, though from a company historically positioned as number two to Nvidia.

Technical Capabilities: MetaX’s flagship C600 GPU features HBM3e memory and FP8 precision aimed at high-performance AI training workloads. The company positions the C600 as competitive with high-end AI accelerators, though again independent performance validation remains limited.

Revenue Trajectory: In 2024, MetaX’s training-and-inference GPU series became the core revenue driver, with GPU boards accounting for 68.99% of total revenue and GPU servers 28.29%. By Q1 2025, GPU board revenue from this series surged to 97.55% of total sales—dramatic concentration signaling both product-market fit and dangerous dependency on a single product line.

IPO Performance: MetaX raised 4.2 billion yuan ($596.3 million) from selling 40.1 million shares in its December 18, 2025 Shanghai IPO, making it the fifth-largest mainland flotation of the year. Shares soared approximately 693% by close of debut session, outdoing even Moore Threads’ spectacular first day. The IPO attracted 5.17 million retail subscriber applications, exceeding Moore Threads and indicating even stronger retail enthusiasm.

Production Challenges: MetaX has suffered production shortages. After running out of its most advanced GPU, the company, along with chip startup Enflame, had to downgrade designs so TSMC would continue manufacturing—a compromise forced by US export controls on providing leading-edge nodes to Chinese AI chip companies.

Path to Profitability: MetaX projects reaching break-even in 2026, an aggressive timeline given current losses but backed by rapidly growing revenue and strong order books. Whether this optimistic projection materializes depends on execution, competition, and continued government support.

Enflame Technology: The Tencent-Backed Dark Horse

Founded in 2018 by former AMD executives Zhao Lidong and Zhang Yalin, Enflame operates with lower public profile than Moore Threads and MetaX but maintains significant backing from Tencent Holdings, China’s internet giant. Enflame designs chips for data centers focused on AI training and inference.

Strategic Positioning: Unlike Moore Threads (which emphasizes gaming capabilities) and MetaX (which focuses on training workloads), Enflame has positioned itself primarily for cloud computing applications, partnering with several local data centers to deploy its chips.

Manufacturing Access: Critically, Enflame is not on the US Entity List, meaning it retains access to TSMC’s advanced manufacturing nodes—a significant competitive advantage over Moore Threads, Biren, and MetaX. This access enables more sophisticated designs without the compromises forced on sanctioned competitors.

IPO Pipeline: Enflame kicked off its tutoring process (mandatory step for Chinese IPO applicants) in August 2024, signaling serious intent to list publicly. The company has not announced specific IPO timing but likely awaits optimal market conditions following the successful Moore Threads and MetaX debuts.

Biren Technology: The Fallen Favorite

Shanghai-based Biren was co-founded in 2019 by Zhang Wen (Wall Street veteran, Harvard Law School graduate, former president of AI company SenseTime) and Jiao Guofang (former Qualcomm and Huawei employee). The founding team also included Xu Lingjie, a former Nvidia engineer who served as president before departing in 2023 following US sanctions impact.

The BR100: Technical Ambition: Unveiled in August 2022, Biren’s BR100 GPU represented China’s most ambitious AI chip to date—a 77 billion transistor, dual-die chip manufactured on TSMC’s 7nm process using 2.5D CoWoS packaging. Specifications included:

• 300MB of on-die cache and up to 64GB of HBM2e memory
• 2.3 TB/s bandwidth
• PCIe Gen5 and CXL 2.0 compatibility
• Peak performance: 256 FP32 teraflops (compared to Nvidia A100’s 19.5)—though comparing raw teraflops across different architectures is notoriously misleading
• Claimed advantage: 2.6x speedup over Nvidia A100 in workloads including computer vision, natural language processing, and conversational AI (based on tests by Shanghai AI Lab)

The BR100’s specifications positioned it as competitive with Nvidia’s A100, though trailing the newer H100. Biren also developed BR104, a single-die PCIe variant for easier integration.

The Sanctions Catastrophe: Biren’s fortunes collapsed in October 2022 when US export controls forced TSMC to halt all manufacturing of Biren chips. The company had strangely released a press statement in September 2022 declaring its slower BR104 chip beat the Nvidia A100 on MLPerf benchmarks—less than two weeks after the US government ordered stoppage of A100 exports to China. Highlighting that their chip exceeded US performance thresholds was catastrophic tactical error.

In October 2023, the US Department of Commerce formally added Biren (along with Moore Threads) to the Entity List, cementing the manufacturing blockade. This action limited Biren’s access to advanced foundries, US chip design software, and critical components.

The Aftermath: The sanctions forced Biren to cut its workforce by one-third and attempt to modify the BR100 specifications to fall below US restriction thresholds—a compromise that undermined the chip’s competitive positioning. Co-founder Xu Lingjie left in 2023. The company secured emergency funding including 280 million USD from Guangzhou government-backed investors and 1.5 billion yuan from Shanghai State-owned Capital Investment (SSCI) in March 2025.

Biren 106 Series: Despite setbacks, Biren launched the Biren 106 series in 2025, supporting inference for Alibaba’s Tongyi QWQ-32B model and offering full-stack training capabilities for DeepSeek-V3. These partnerships with leading Chinese AI companies demonstrate that despite US sanctions, Biren maintains viability for domestic Chinese AI workloads.

Current Deployment: Biren’s GPUs are deployed in intelligent computing centers operated by China Mobile, China Telecom, ZTE, and the Shanghai AI Laboratory—major domestic customers providing revenue base, though many have themselves faced US sanctions or investigations.

Financial Performance: The stark financial reality contrasts sharply with technical ambitions:

• 2023 Revenue: Company began generating revenue from intelligent computing solutions
• 2024 Revenue: 336.8 million yuan ($47.8 million)—some sources cite approximately 400 million yuan for the full year
• H1 2025 Revenue: 58.9 million yuan—a dramatic decline suggesting significant operational challenges
• H1 2025 Losses: Nearly 9 billion yuan—a 32.3% increase year-over-year
• Total Losses Since Inception: Losses have widened consecutively, reflecting heavy R&D spending, software development costs, commercialization expenses, and the 108.7 million yuan hit from US sanctions in 2023

Despite these massive losses, Biren reports pending and contracted orders totaling approximately 2.1 billion yuan, which the company claims will support near-term growth as deployments scale—though converting orders to revenue amid US restrictions remains uncertain.

Valuation: Before a mid-2025 funding round, Biren was valued at approximately 14 billion yuan ($2 billion), though the IPO pricing suggests market valuation closer to $2.2 billion. Since inception, Chinese media estimate Biren has raised more than 5 billion yuan ($709 million).

Why the “Four Dragons” Matter

These four companies represent China’s most credible GPU development capability. Each leverages former employees from Nvidia, AMD, and Qualcomm—engineers who understand GPU architecture, software ecosystems, and market positioning. Combined, they’ve raised billions in private and state-linked capital. Their rapid progression from founding to IPO (Moore Threads and MetaX listed within 4-5 years of founding) demonstrates prioritization and resource mobilization that only state backing enables.

However, critical weaknesses remain. All four operate at significant losses. None has demonstrated GPU performance matching Nvidia’s latest H100 or upcoming Blackwell chips. More critically, all face the same existential challenge: breaking CUDA’s ecosystem dominance.

The CUDA Problem: Ecosystem vs. Hardware

Nvidia commands approximately 86% of the AI GPU market not primarily because its chips are technically superior—though they are—but because of CUDA, the software platform developers use to program Nvidia GPUs. CUDA represents 20+ years of ecosystem development: mature libraries, extensive documentation, millions of lines of code, and most critically, an installed base of developers comfortable with the platform.

Chinese GPU makers face a chicken-and-egg problem. Developers won’t adopt their platforms unless the software ecosystem rivals CUDA’s maturity and breadth. But ecosystems don’t mature without developer adoption. Attempting to convince developers to rewrite code for unproven platforms—abandoning CUDA’s massive library of optimized algorithms—is extraordinarily difficult.

Biren’s BIRENSUPA: The Software Challenge

Biren developed BIRENSUPA, its proprietary software platform and programming model. According to the company, “developers familiar with CUDA can easily write code for SUPA,” suggesting compatibility or similarity. BIRENSUPA supports PyTorch, TensorFlow, and PaddlePaddle frameworks and provides an OpenCL compiler.

However, “easily write code” understates the challenge. Even assuming BIRENSUPA achieves CUDA-like functionality (a massive assumption), asking developers to migrate represents friction that inhibits adoption. Enterprises have invested billions optimizing AI workloads for CUDA. Banks, pharmaceutical companies, autonomous vehicle developers, and cloud providers have massive codebases tuned to Nvidia’s architecture. Rewriting for Chinese alternatives—even with good compatibility—requires substantial investment with uncertain payoff.

The Forced Migration: Geopolitical Reality vs. Technical Choice

Yet Chinese GPU makers may succeed not through technical superiority but through geopolitical necessity. If Chinese technology companies lose access to Nvidia chips—either through US export controls or Chinese government requirements prioritizing domestic alternatives—they have no choice but to migrate.

ByteDance invested in Moore Threads and stockpiled approximately one million Nvidia H20 GPUs (China-specific chips designed to comply with earlier US restrictions) before April 2025 export control tightening. Tencent backed Enflame Technology and purchased 2 billion yuan worth of GPUs in Q1 2025 for its ChatGPT-like app Yuanbao. Alibaba’s been deploying custom chips since early 2025, with employees claiming these chips match H20 performance.

Chinese authorities have instructed technology companies to pause H20 purchases pending “national security review” and require data centers to prioritize domestic chips. When Alibaba Cloud, Tencent Cloud, and Baidu Cloud deploy domestic GPUs at scale in 2026-2027, Chinese developers have no choice. The migration is coming whether technology is ready or not.

This represents export controls’ actual impact: Not stopping Chinese AI development—China continues advancing rapidly, as evidenced by DeepSeek’s breakthrough models—but fragmenting global AI infrastructure. American companies lose a $10+ billion annual Chinese market. Chinese companies get inferior chips but maintain domestic AI capability. Both sides lose efficiency and scale economies, though arguably China gains strategic autonomy it values more than efficiency.

The $623 Million Test: Biren’s IPO Strategy

Biren’s December 23, 2025 bookbuilding launch targets raising up to HK$4.85 billion ($624 million) through offering 247.7 million shares priced HK$17-HK$19.60 per share. The company secured $372.5 million in cornerstone investor commitments—representing approximately 60% of the total raise before public bookbuilding even began.

The Cornerstone Investors: Validation Through Capital

Twenty-three cornerstone investors have committed to subscribing $372.5 million worth of shares with six-month lock-up periods, including:

Venture Capital Heavyweights:

• Qiming Venture Partners: Major Chinese VC with extensive semiconductor portfolio
• IDG Capital: Early backer that invested during Biren’s initial rounds
• Hillhouse Venture (HongShan Capital): Powerhouse Chinese PE/VC firm managing hundreds of billions

Insurance and Financial Institutions:

• Ping An Group: One of China’s largest insurance conglomerates, $165 billion market cap
• Domestic mutual funds and insurers: Multiple Chinese asset managers seeking exposure to strategic industries

International Investors:

• UBS: Swiss banking giant’s participation signals some international confidence despite US sanctions risks
• Lion Global Investors (Singapore): Southeast Asian asset manager
• Eastspring Investments: Asian investment management firm

Strategic Corporate Investors:

• Digital China: State-linked technology distributor and systems integrator
• Russia-China Investment Fund: Geopolitical alignment reflected in capital

This diverse investor base combining state-linked entities, private VCs, insurance companies, and select international investors demonstrates sophisticated capital structuring. State-linked investors provide political backing and patient capital. Private VCs bring discipline and growth expectations. Insurance companies seek long-duration assets matching liability profiles. International investors, though limited, provide market credibility.

The Hong Kong Listing Choice: Regulatory and Strategic Calculus

Biren chose Hong Kong over Shanghai’s STAR Market (where Moore Threads and MetaX listed) for several strategic reasons:

International Capital Access: Hong Kong provides access to international investors who may be restricted or hesitant regarding mainland exchanges. UBS, Lion Global Investors, and Eastspring Investments’ participation exemplifies this advantage.

Regulatory Framework: While Beijing maintains ultimate control, Hong Kong’s common-law legal framework and international accounting standards provide perceived governance advantages that attract international capital.

US Sanctions Mitigation: Listing in Hong Kong rather than Shanghai creates slight additional distance from mainland regulatory environment, potentially making international investors more comfortable despite Biren’s Entity List status.

Currency Flexibility: Hong Kong dollar pegged to US dollar provides currency stability. Proceeds raised in HKD are easily convertible, facilitating international transactions if needed.

Beijing’s Preference: Chinese authorities prefer Hong Kong listings for strategically important technology firms, providing international capital access while remaining within China-aligned regulatory framework. This “offshore but onshore” positioning serves multiple objectives.

The January 2 Trading Debut: First IPO of 2026

Biren’s planned January 2, 2026 trading debut positions it as the first new listing of the new year—symbolically important timing suggesting government prioritization. Chinese authorities clearly greenlit the IPO rapidly (China Securities Regulatory Commission approval came in mid-December), indicating political support.

The January 2 date also capitalizes on momentum from Moore Threads’ and MetaX’s spectacular December debuts. If Biren shares surge similarly—repeating the 400%+ first-day gains—it validates the entire Chinese GPU investment thesis and could trigger additional IPOs from companies like Enflame and others waiting in the wings.

Use of Proceeds: R&D, Manufacturing, and Working Capital

While specific breakdown awaits final prospectus details, Chinese AI chip companies typically allocate IPO proceeds across several categories:

Research and Development: 40-50% of proceeds typically fund R&D including next-generation chip architectures, software ecosystem development (critical for CUDA competition), and AI algorithm optimization.

Manufacturing and Equipment: 20-30% funds production, though as fabless company Biren doesn’t own fabs. This allocation covers wafer orders from foundries (now entirely domestic given US restrictions), testing and packaging, and production ramp-up as designs mature.

Working Capital: 15-25% provides operational runway given massive ongoing losses. Biren burns through hundreds of millions quarterly; IPO proceeds extend runway until next funding round or profitability (whichever comes first—almost certainly the former).

Software Ecosystem Development: 10-15% funds BIRENSUPA platform improvements, developer tools, training programs, and partnerships with software vendors to build out ecosystem necessary for long-term success.

The Profitability Question: When, If Ever?

Biren’s prospectus cites “global political and economic factors as ongoing risks to financial position and path to profitability, including U.S. trade restrictions.” This boilerplate language understates the fundamental uncertainty. Biren faces multiple paths, none guaranteed:

Optimistic Scenario: China’s domestic AI market grows explosively. Government mandates requiring domestic chips in government and state-owned enterprise data centers guarantee demand. Biren achieves economies of scale, software ecosystem matures, and revenue grows faster than costs. Path to profitability by 2027-2028, following similar trajectory as Cambricon (which turned profitable for first time in 2024).

Moderate Scenario: Biren survives as viable but marginal player. Government provides continued subsidies and guaranteed orders maintaining operations but never achieving true commercial viability. Company operates semi-permanently at losses or break-even, serving strategic function rather than profit maximization. Many Chinese state champions follow this path.

Pessimistic Scenario: US restrictions tighten further, cutting off critical manufacturing or components. Domestic alternatives prove insufficient. Chinese AI companies find workarounds importing smuggled or gray-market Nvidia chips. Biren’s order book evaporates, IPO provides temporary life extension, but company ultimately fails or gets absorbed by larger Chinese semiconductor player.

Currently, Biren sits uncomfortably between optimistic and moderate scenarios. Revenue exists ($47.8 million in 2024) but declined substantially in H1 2025 ($58.9 million for full year 2025 would be barely above 2024 despite supposedly growing market). Pending orders of 2.1 billion yuan sound impressive but converting orders to revenue amid sanctions remains uncertain. Losses of 9 billion yuan in H1 2025 alone dwarf the $624 million IPO raise—this capital provides perhaps 12-18 months runway unless revenue accelerates dramatically.

The Big Fund: China’s Multi-Hundred-Billion-Dollar Semiconductor Wager

Biren’s IPO exists within context of China’s unprecedented state-directed investment in semiconductor self-sufficiency. Understanding the broader strategy illuminates why investors remain enthusiastic despite consistent GPU maker losses and why Chinese authorities prioritize domestic chip capabilities regardless of economics.

Made in China 2025: The Strategic Foundation

China’s semiconductor ambitions crystallized in 2015 with the “Made in China 2025” plan, setting aspirational goal of 70% self-sufficiency in semiconductors by 2025. Though official public references disappeared after US Trade Representative launched Section 301 investigation in 2017 (China learned to avoid publicizing targets Americans might oppose), the underlying strategy persists and intensified.

China’s 14th Five-Year Plan (2021-2025) explicitly designated semiconductors as strategic tech priority requiring “whole-of-society effort” to achieve self-reliance. This language signals that commercial viability takes backseat to strategic autonomy—profitability is secondary to ensuring China can produce necessary chips domestically even if less efficiently than buying from abroad.

Current Self-Sufficiency Rate: As of 2025, China has reached approximately 30-40% semiconductor self-sufficiency (depending on how measured), falling well short of the 70% goal. However, this still represents dramatic progress from approximately 5% in 2018—a six-fold increase in seven years. China’s domestic semiconductor industry achieved unprecedented 30.6% annual growth rate, with total sales reaching $39.8 billion in 2020 and continuing upward.

The gap between target and reality underscores challenge: Semiconductors are perhaps the most complex products humans manufacture, requiring integration of advanced materials science, quantum physics, precision engineering, and cutting-edge software. Catching up to companies like TSMC, ASML, Nvidia, and Intel—who spent decades developing capabilities—in single generation is extraordinarily difficult regardless of capital deployed.

The Big Fund: Phases I, II, and III

Central to China’s semiconductor strategy is the National Integrated Circuit Industry Investment Fund—universally known as “Big Fund”—established 2014 to inject state capital into domestic semiconductor companies. The fund operates under Ministry of Industry and Information Technology (MIIT) and Ministry of Finance, adopting two-tier structure where board sets strategy and approves major projects while Sino IC Capital manages investments.

Big Fund Phase I (2014-2019):

• Capital: Raised $21.8 billion initially, ultimately deploying $39 billion
• Focus: 60-70% allocated to chip manufacturing, with only ~1% each to equipment and materials
• Strategy: Building foundry capacity to manufacture chips domestically
• Key Investments: SMIC (Semiconductor Manufacturing International Corporation), Hua Hong Semiconductor, YMTC (Yangtze Memory Technology Corp)
• Results: Established domestic foundry base capable of mature-node production (28nm and above), though trailing-edge versus TSMC’s 3nm capabilities

Big Fund Phase II (2019-2024):

• Capital: Exceeded $35 billion in state-backed financing
• Focus: Broader than Phase I, included equipment, materials, design, and packaging
• Strategy: Building complete ecosystem beyond just manufacturing
• Key Investments: Expanded to semiconductor equipment manufacturers, EDA software companies, packaging and testing firms
• Results: Reduced (but didn’t eliminate) dependencies on foreign equipment and materials; improved capabilities in photoresist, etching, CMP (chemical mechanical planarization)

Big Fund Phase III (2024-present):

• Capital: 344 billion yuan ($47.5 billion)—larger than Phases I and II combined
• Focus: “Low self-sufficiency segments”—advanced equipment, 3D integration technology, high-bandwidth memory (HBM), AI chips
• Strategy: Pivoting from “broad-based” to “selective” investment, targeting bottlenecks
• First Public Investment: September 2024, Guotou Jixin (99.9% funded by Big Fund III) committed up to 450 million yuan to Piotech Jianke, focusing on 3D integration equipment
• AI Emphasis: HBM for AI applications and AI chip development explicitly prioritized given geopolitical competition

Big Fund III’s $47.5 billion exceeds the $39 billion in direct incentives Washington introduced via US CHIPS and Science Act—China is outspending America on semiconductor industrial policy. However, US maintains significant advantages: More mature industry, deeper expertise, control of critical chokepoints (especially lithography), and alliance network enabling supply chain redundancy.

National AI Industry Investment Fund: Launched January 2025 with initial capital of 60 billion yuan ($8.2 billion), this separate fund specifically targets AI infrastructure including GPU development, AI data centers, training platforms, and foundation model development. Biren and the “four little dragons” are prime beneficiaries.

Complementary State Support Beyond Big Fund

Big Fund represents only most visible component of China’s semiconductor support. Complementary mechanisms include:

Provincial and Municipal Funds: More than 15 local government IC funds totaling $25 billion dedicated to semiconductor companies. Guangzhou, Shanghai, Shenzhen, and other cities compete to attract semiconductor firms with preferential financing.

Below-Market Loans: OECD 2019 study found China’s four leading state-backed semiconductor companies received $4.85 billion in below-market loans from state-owned banks 2014-2018—representing 98% of below-market borrowing by leading global semiconductor companies. Favorable lending terms provide substantial subsidy difficult to quantify but massive in scale.

Tax Breaks: Reduced income taxes, VAT rebates, and other fiscal incentives for semiconductor firms. Some estimates suggest tax benefits equal 20-30% of government financial support.

Reduced Utility Rates: Semiconductor manufacturing requires enormous electricity and water. State-owned utilities provide preferential rates to strategic fabs.

Free or Discounted Land: Local governments provide factory sites at below-market rates or free to attract semiconductor investments.

Forced Technology Transfers: Requirements for foreign companies operating in China to partner with domestic firms and share technology, enabling knowledge transfer. Diminishing as Western firms pull back but historically significant.

Import Substitution Programs: Government procurement preferentially favors domestic semiconductors, guaranteeing demand even when foreign alternatives are superior.

Combined, these mechanisms represent support likely exceeding $100 billion since 2014—more than three times the Big Fund’s direct investments. Total registered capital in Chinese semiconductor industry is approximately $51 billion, meaning 43% is directly or indirectly government-funded—extraordinary state ownership even by Chinese standards.

The Strategic Logic: Buying Optionality

China’s semiconductor investment defies conventional economic logic. Return on invested capital is negative. Massive overcapacity in mature nodes is emerging. Most companies operate at losses despite billions in subsidies. From pure profit-maximization perspective, China would be better off buying chips from Taiwan, South Korea, and the United States—letting those countries bear R&D costs while China focuses on sectors where it has comparative advantage.

However, viewing semiconductor strategy through commercial lens fundamentally misunderstands Chinese government calculus. China isn’t building profitable semiconductor industry—at least not primarily. China is buying optionality in what President Xi Jinping identified in 2014 as “core technology” that must be produced domestically.

The Stranglehold Problem: America demonstrated through Huawei sanctions that control of semiconductor supply chain provides geopolitical leverage. When US cut Huawei off from TSMC-manufactured chips, Huawei’s smartphone business collapsed—from global second-largest to also-ran. This object lesson demonstrated vulnerability of technological dependence.

China’s strategic calculation: Better to invest $100+ billion building domestic semiconductor capability—even if less efficient—than risk technological stranglehold in crisis scenarios. If US-China tensions escalate to conflict over Taiwan, semiconductor supply chain becomes critical vulnerability. Domestic production capability, even if second-rate compared to TSMC, provides strategic resilience.

The AI Dimension: Semiconductors matter broadly, but AI chips matter existentially. Artificial intelligence is transforming military capabilities, economic productivity, and social governance. Nation that leads in AI likely leads 21st century. AI leadership requires access to advanced compute—GPUs capable of training and deploying large language models, computer vision systems, autonomous weapons, and other applications.

When US restricts Chinese access to Nvidia’s H100 and other advanced AI chips, it directly threatens China’s AI ambitions. Developing domestic alternatives like Biren, Moore Threads, MetaX, and Enflame—despite technical inferiority to Nvidia—ensures China can continue advancing AI even if completely cut off from Western chips. This strategic logic justifies enormous subsidies for companies that may never turn profit commercially.

The Comparison to America’s AI Infrastructure Spending: Chinese semiconductor strategy appears less irrational when compared to American private sector AI investment. Meta spent $14.3 billion acquiring 49% stake in Scale AI primarily to poach talent for its “Superintelligence” team—despite uncertain returns. Google, Microsoft, and Amazon collectively spend over $200 billion annually on AI infrastructure with no profitability guarantee.

If American companies justify massive AI spending based on future strategic positioning rather than immediate returns, China’s government applying similar logic to semiconductor self-sufficiency makes strategic sense. Both represent bets on technological leadership determining future competitive position.

Geopolitical Implications: The Bifurcated Chip World

The confluence of Chinese state investment and American export controls is creating what analysts call “technological decoupling”—separate, incompatible technology stacks in AI and semiconductors. This bifurcation carries profound implications for global trade, innovation, and security.

The US Export Control Regime: A Primer

US restrictions on Chinese semiconductor access have evolved through multiple iterations, becoming increasingly comprehensive:

October 2022 Controls: Initial rules restricted export of advanced logic chips (Nvidia A100, H100, AMD MI250) to China and prohibited US semiconductor manufacturing equipment (SME) sales to Chinese fabs producing advanced chips.

October 2023 Controls: Expanded restrictions to block Nvidia A800 and H800 (China-specific chips designed to comply with 2022 rules) and added Biren and Moore Threads to Entity List, cutting them off from TSMC manufacturing and US design software.

December 2024 Controls: Further tightened restrictions on high-bandwidth memory (HBM)—critical component for AI chips—with country-wide ban on advanced HBM sales to China and tight end-use/end-user controls on less advanced HBM.

Entity List: Over 600 Chinese entities now listed, including semiconductor companies, AI labs, universities, and data centers. Any company (American or foreign) using American technology must obtain export licenses (typically denied) before providing goods/services to listed entities.

Key Impact: Controls don’t ban all chip sales to China—mature-node chips (28nm and above) remain unrestricted. Strategy aims to choke off China’s access to cutting-edge AI and HPC chips while minimizing economic damage to US semiconductor industry from losing Chinese sales.

Dutch and Japanese Alignment: Netherlands (home to ASML, only manufacturer of extreme ultraviolet lithography machines necessary for cutting-edge chips) and Japan (Tokyo Electron, other critical equipment makers) have implemented parallel export controls. This multilateral coordination prevents China from circumventing US restrictions through alternative suppliers.

The Chinese Response: Stockpiling and Smuggling

Before export controls fully tighten, Chinese technology companies engaged in massive stockpiling:

ByteDance: Ordered approximately 1 million Nvidia H20 GPUs (China-specific variant) before April 2025 restrictions—representing over $5 billion in Nvidia chips. This stockpile provides computing capacity for several years of AI development.

Alibaba: Purchased billions in Nvidia chips for Alibaba Cloud before restrictions, while simultaneously developing custom chips internally.

Tencent: Bought 2 billion yuan worth of GPUs in Q1 2025 for ChatGPT-like app Yuanbao and other AI applications.

This stockpiling created one-time windfall for Nvidia—$10 billion annual Chinese sales concentrated into shorter period—but represents future sales pulled forward. Once stockpiles exhaust (likely 2026-2027), Chinese demand for US chips collapses, costing Nvidia and others billions in annual revenue permanently.

Gray Market and Smuggling: Reports indicate some Nvidia chips reach China through third countries—shipping to Singapore, Malaysia, or other jurisdictions then re-exported to China, potentially with specifications modified to appear compliant with export controls. US authorities are tightening enforcement, but gray market trade persists.

Chinese Government Review: Beijing announced “national security review” of foreign AI chips, ostensibly checking whether imported Nvidia chips pose security risks. This provides cover for slowing imports and pressuring Chinese companies to adopt domestic alternatives—turning US restriction strategy back against American companies.

The Innovation Impact: Diverging Tech Stacks

Technological decoupling creates inefficiencies and innovation costs on both sides:

Chinese Innovation Continues: Despite export controls, Chinese AI development continues at remarkable pace. DeepSeek’s reasoning models achieved performance comparable to OpenAI’s o1 and o3 while claiming dramatically lower training costs—suggesting that algorithmic innovation, not just compute scale, drives progress. Chinese researchers publish prolifically, and Chinese companies deploy AI commercially across massive domestic market.

But With Inefficiencies: Training models on inferior domestic GPUs requires more time, energy, and money than training on Nvidia H100s. Chinese AI companies compensate through algorithmic efficiency, but compute handicap creates real competitive disadvantage in scale-dependent applications.

American Companies Lose Market: $10 billion annual Chinese semiconductor sales disappear for US companies—not just Nvidia but also Applied Materials, Lam Research, KLA, and other equipment makers. These sales historically subsidized R&D benefiting global customers. Lost Chinese revenue means higher prices or reduced innovation for American and allied customers.

Fragmented Ecosystems: Two separate AI hardware ecosystems are emerging—CUDA-based (Nvidia, AMD to lesser extent) in West, and domestic alternatives (Biren, Moore Threads, Huawei Ascend) in China. Software developers must maintain dual codebases to serve both markets. Research collaboration becomes harder when Chinese researchers use different hardware/software stacks than Western counterparts.

Europe and Rest of World Squeezed: Countries trying to maintain relationships with both US and China face pressures. Can European companies use Chinese chips without triggering US sanctions? Can they sell to Chinese customers without violating US export controls? These ambiguities create compliance costs and risks deterring beneficial commerce.

The Taiwan Wildcard: TSMC Caught Between Superpowers

Taiwan Semiconductor Manufacturing Company occupies uncomfortable position as world’s most advanced foundry while sitting geographically and geopolitically between US and China:

US Pressure: America demands TSMC stop manufacturing for Chinese chip companies on Entity List (which TSMC complies with) and potentially reduce all advanced-node China business. US views TSMC as strategic asset that must be denied to China.

Chinese Market: China represents significant revenue for TSMC—estimates suggest 10-20% of total sales though exact figures murky given complex supply chains. TSMC wants to maintain business relationships while complying with US restrictions.

Invasion Risk: If China invades Taiwan, TSMC’s fabs become Chinese-controlled or destroyed in conflict—either outcome catastrophic for global semiconductor supply. This scenario motivates both American push for TSMC to build US fabs (though these lag Taiwan fabs technologically) and Chinese determination to develop domestic alternatives.

SMIC as Alternative: China’s Semiconductor Manufacturing International Corporation (SMIC) trails TSMC by several generations (stuck at 7nm through heroic efforts versus TSMC’s 3nm), but Entity List companies like Biren increasingly have no choice but to work with SMIC. This accelerates SMIC development and reduces Chinese dependence on TSMC—unintended consequence partially undermining US strategy.

Who Wins the Chip War?

Neither side is winning cleanly:

American Tactical Success: Export controls have slowed Chinese AI chip development. Biren, Moore Threads, MetaX all make chips inferior to Nvidia’s latest. China missed 70% self-sufficiency target by wide margin. Chinese AI companies handicapped by compute disadvantage.

American Strategic Questions: But China continues advancing AI rapidly despite handicaps. Chinese semiconductor industry is growing, not collapsing. American companies lose billions in sales. Allied countries uncomfortable with pressure to choose sides. And if restrictions successfully contain China for decade, what happens when controls eventually fail? Will Chinese industry that spent decade learning to work around US technology be permanently less dependent—worse long-term outcome than if US had maintained interdependence?

Chinese Tactical Setbacks: Biren forced to redesign chips. TSMC manufacturing cut off. Talent gap persists—despite massive salaries, convincing top Nvidia engineers to move to China becomes harder as geopolitical tensions rise. Ecosystem lock-in to CUDA remains.

Chinese Strategic Gains: But strategic autonomy increases. Domestic alternatives exist where none did five years ago. Government support galvanizes whole-of-society effort creating innovation ecosystem. Younger generation of Chinese engineers gaining experience on cutting-edge problems. And China’s enormous domestic market provides scale to sustain domestic champions even if globally second-rate.

The likely outcome is bifurcated world: Western countries use Nvidia/AMD/Intel chips running CUDA and similar ecosystems. China uses domestic chips running separate stacks. Both continue advancing AI—West probably faster in frontier capabilities, China potentially better at efficient deployment given compute constraints. Neither achieves decisive advantage, but collaboration and scale economies suffer, making humanity worse off overall.

Biren’s IPO as Bellwether: What Success or Failure Signals

Biren Technology’s January 2, 2026 trading debut will serve as critical market test. Will shares replicate Moore Threads’ 425% or MetaX’s 693% first-day surges? Or will investor enthusiasm finally meet reality of massive losses, US sanctions, and uncertain commercialization?

Bull Case: Why Shares Could Soar

Precedent Performance: Moore Threads and MetaX set extraordinary precedent. If Biren trades similarly, investors buying at IPO price and selling first day could triple money overnight. This speculation alone drives demand.

Strategic Importance: Chinese government clearly prioritizes semiconductor self-sufficiency. Investors betting government ensures Biren survives—even if commercially unviable—through continued subsidies, guaranteed orders, and other support. “Too strategically important to fail” becomes investment thesis.

Revenue Growth Potential: Despite H1 2025 decline, Biren reports 2.1 billion yuan pending orders. If these materialize, revenue could increase dramatically from current ~$50-60 million run rate to several hundred million—validating growth story.

Manufacturing Workarounds: Some investors believe Chinese foundries like SMIC will successfully produce Biren chips at acceptable quality despite US restrictions, enabling commercialization and fulfilling order book.

Ecosystem Momentum: If Chinese technology giants (ByteDance, Alibaba, Tencent) commit to deploying domestic GPUs at scale as government pressures them, Biren captures share of enormous TAM (total addressable market) as one of four viable domestic suppliers.

Technical Progress: Biren 106 series supporting DeepSeek-V3 training and Alibaba Tongyi QWQ-32B inference demonstrates technical credibility for real Chinese AI workloads—not just marketing claims.

Hong Kong Premium: Some investors view Hong Kong listing as less risky than Shanghai, providing better governance and exit options. International participation (UBS, etc.) validates investment thesis beyond purely domestic enthusiasm.

Scarcity Value: Only four Chinese GPU companies exist. Three have listed. For investors wanting exposure to Chinese AI chip development, Biren may be last chance to buy at IPO before dramatic first-day surge.

Bear Case: Why Shares Could Disappoint

Fundamentals Disconnected from Valuation: At $2.2 billion valuation, Biren trades at 46x 2024 revenue ($47.8 million) and 37x 2025 revenue if full-year matches H1 2025 run rate. For money-losing company with declining revenue, these multiples defy financial logic.

US Sanctions Reality: Unlike Enflame (not on Entity List with TSMC access), Biren remains under US sanctions limiting access to advanced manufacturing and components. This handicap is structural and worsening, not improving.

Q1 2025 Revenue Collapse: H1 2025 revenue of $58.9 million suggests annual revenue may barely exceed 2024—potentially declining. This contradicts growth story and suggests order book may not convert to revenue as claimed.

Losses Accelerating: 9 billion yuan H1 2025 losses represent 32.3% increase year-over-year. At this burn rate, even $624 million IPO provides only 12-18 months runway. Continuous dilution likely as company returns for more capital.

Co-Founder Departures: Xu Lingjie’s 2023 departure and reported tensions within founding team raise execution concerns. Leadership instability at critical growth stage historically predicts problems.

CUDA Lock-In: No credible path identified for breaking Nvidia’s ecosystem dominance. BIRENSUPA claims compatibility but developers skeptical. Without ecosystem adoption, hardware becomes commodity competing solely on price—terrible business model for loss-making company.

Speculative Bubble Signs: When shares of money-losing companies with declining revenue increase 400-700% in single day, market exhibits classic bubble characteristics. Moore Threads and MetaX first-day pops may represent peak euphoria rather than sustainable valuations.

Chinese Retail Investor Behavior: Chinese retail investors often engage in extreme speculation, particularly in government-favored sectors. IPO “pop” may reflect FOMO (fear of missing out) and greater fool theory rather than fundamental analysis.

Comparable Company Warning: Cambricon Technologies listed in Shanghai 2020 with similar government support and strategic importance, saw enormous early enthusiasm, then declined 60%+ from peaks as commercialization disappointed. Cambricon only turned profitable in 2024—four years post-IPO. Biren may follow similar trajectory.

Geopolitical Risks: If US-China tensions escalate, foreign investors could face sanctions risks for holding Chinese strategic technology companies. Even Hong Kong listing doesn’t eliminate this risk—Entity List status creates legal ambiguity for international capital.

Most Likely Scenario: Pop Then Fade

Most probable outcome combines elements of both: First-day pop of 100-300% as retail speculation and momentum from Moore Threads/MetaX successes drive buying. This validates “strategic importance” thesis for short-term traders and provides exit opportunity for insiders and cornerstone investors after lock-up expires.

But over 12-24 months, shares drift lower as operational reality becomes apparent. Revenue growth disappoints. Losses continue. US sanctions prevent manufacturing at advanced nodes. Software ecosystem doesn’t materialize. Eventually, shares trade at substantial discount to IPO valuation, wiping out gains for anyone who didn’t sell during initial euphoria.

Government support ensures Biren survives—company doesn’t go bankrupt. But survival isn’t thriving. Biren joins ranks of Chinese state champions that exist indefinitely in semi-viable state: Too strategically important to let fail, but never achieving commercial success justifying inflated valuations.

This pattern describes many Chinese technology champions: Alive due to government support, strategically useful, but terrible investments for anyone other than political insiders or short-term speculators.

Conclusion: The $624 Million Bet on Strategic Autonomy

Biren Technology’s Hong Kong IPO transcends typical technology listing. It represents crystallization of China’s multi-hundred-billion-dollar bet on semiconductor self-sufficiency, the market’s attempt to price strategic importance versus commercial viability, and most tangibly, bellwether for whether investor enthusiasm for Chinese AI chips can survive contact with operational reality.

The fundamentals suggest caution: Biren operates at massive losses, faces worsening US sanctions, showed declining H1 2025 revenue, lost key co-founder, and lacks credible path to profitability. By traditional metrics, $2.2 billion valuation for company generating ~$50 million declining revenue defies logic.

Yet Moore Threads and MetaX—facing similar fundamentals—saw 425% and 693% first-day gains respectively, suggesting fundamentals take backseat to strategic narrative. Chinese retail investors demonstrate willingness to disregard conventional valuation when government signals strategic importance. For speculators, the question isn’t whether Biren is good business, but whether shares appreciate enough to profitably exit.

For Chinese government, Biren’s success or failure as investment is largely irrelevant. What matters is whether Biren (alongside Moore Threads, MetaX, Enflame, Huawei, and others) collectively provide Chinese AI companies sufficient domestic compute options to maintain AI advancement regardless of US restrictions. By this metric, the “four little dragons” are succeeding—China continues publishing AI research prolifically, deploying commercial AI applications broadly, and advancing toward what Chinese President Xi Jinping calls “technological sovereignty.”

For American policymakers, Biren’s IPO provides uncomfortable reality check. Export controls have not stopped Chinese GPU development—they have ensured it occurs domestically, subsidized by Chinese government, ultimately reducing long-term American leverage while costing American companies billions in lost sales. Whether this trade-off advances American interests depends on time horizons and strategic assumptions impossible to resolve definitively.

For global technology industry, Biren represents bifurcation everyone hoped to avoid. The elegant efficiency of globally integrated semiconductor supply chains—where chips designed in California, manufactured in Taiwan using Dutch equipment, assembled in Malaysia, and sold worldwide enabled unprecedented pace of innovation—is fracturing into regional blocs optimizing for resilience over efficiency. Humanity will advance technologically slower as scale economies and collaboration suffer. Both superpowers will field capable AI systems, just at higher cost and lower performance than if they cooperated.

Biren’s January 2 trading debut will tell us whether markets have priced in this new reality or remain in speculative euphoria disconnected from fundamentals. If shares surge 500%+ like predecessors, bubble dynamics likely prevail—ultimately painful for retail investors left holding when music stops. If shares rise modestly or decline, markets may be ahead of government in recognizing that strategic importance doesn’t equal commercial success.

Either way, Biren Technology’s $624 million IPO marks milestone in what will be decade-long restructuring of global semiconductor industry from integrated whole to rival camps. The company may never turn profit. It may never build GPUs competitive with Nvidia’s. But it serves function Chinese government values: Ensuring that when geopolitical crisis comes—whether over Taiwan, South China Sea, or issues yet unknown—China controls semiconductor supply chain foundational to AI leadership and national security.

By that standard, the $624 million IPO represents extraordinarily cheap insurance premium on multi-trillion-dollar economy. Whether investors buying shares on January 2 will feel similarly sanguine when reviewing portfolios in 2027 remains entirely different question.

Sources

1. South China Morning Post. (December 23, 2025). Chinese AI chipmaker Biren kicks off bookbuilding for US$624 million Hong Kong IPO. https://www.scmp.com/business/banking-finance/article/3337272/chinese-ai-chipmaker-biren-kicks-bookbuilding-us624-million-hong-kong-ipo
2. Tom’s Hardware. (December 22, 2025). China’s premier GPU maker Biren kicks off Hong Kong IPO — GPU startups vying for Nvidia’s crown race to fund AI chip development. https://www.tomshardware.com/tech-industry/biren-kicks-off-hong-kong-ipo
3. Asia Tech Daily. (December 23, 2025). Biren’s Hong Kong IPO Signals a New Phase for China’s Homegrown AI Chipmakers. https://asiatechdaily.com/birens-hong-kong-ipo-signals-a-new-phase-for-chinas-homegrown-ai-chipmakers/
4. South China Morning Post. (December 16, 2025). Biren Technology’s Hong Kong IPO gets green light as China’s AI chip firms rush to list. https://www.scmp.com/tech/tech-trends/article/3336658/biren-technologys-hong-kong-ipo-gets-green-light-chinas-ai-chip-firms-rush-list
5. Wikipedia. Biren Technology. https://en.wikipedia.org/wiki/Biren_Technology
6. Rest of World. (September 24, 2025). China chip startups race to replace Nvidia amid U.S. export bans. https://restofworld.org/2025/china-chip-startups-nvidia-us-export/
7. Tom’s Hardware. (October 17, 2023). US Prohibits Exports of Nvidia’s A800 and H800 to China, Blacklists Chinese GPU Developers. https://www.tomshardware.com/news/us-prohibits-exports-of-nvidias-a800-and-h800-to-china-blacklists-chinese-gpu-developers
8. The Korea Times. (October 19, 2023). US sanctions on Biren and Moore Threads strike strong blow to China’s GPU champions. https://www.koreatimes.co.kr/www/world/2023/10/672_361495.html
9. CNBC. (December 17, 2025). MetaX and Moore Threads’ IPOs underscore Chinese chipmakers’ growing challenge to Nvidia. https://www.cnbc.com/2025/12/17/metax-moore-threads-chinese-rivals-nvidia-ai-chips.html
10. TrendForce. (October 7, 2025). China’s GPU Trio Rise as NVIDIA Retreats — Decoding Moore Threads, MetaX, and Cambricon. https://www.trendforce.com/news/2025/10/07/news-chinas-gpu-trio-rise-as-nvidia-retreats-decoding-moore-threads-metax-and-cambricon/
11. South China Morning Post. (December 17, 2025). MetaX soars nearly 700% in debut as traders snap up second Chinese GPU maker to go public. https://www.scmp.com/business/china-business/article/3336690/metax-soars-frenzied-debut-traders-snap-second-chinese-gpu-maker-go-public
12. Wikipedia. China Integrated Circuit Industry Investment Fund. https://en.wikipedia.org/wiki/China_Integrated_Circuit_Industry_Investment_Fund
13. Centre for Emerging Technology and Security. China’s Quest for Semiconductor Self-Sufficiency. https://cetas.turing.ac.uk/publications/chinas-quest-semiconductor-self-sufficiency
14. Economics Observatory. (May 28, 2025). What’s happening in China’s semiconductor industry? https://www.economicsobservatory.com/whats-happening-in-chinas-semiconductor-industry
15. Sino Southeast Initiative. (July 27, 2025). China’s Semiconductor Investment Defies Economics—But Makes Perfect Strategic Sense. https://sinosoutheastinitiative.com/2025/07/26/chinas-semiconductor-investment-defies-economics-but-makes-perfect-strategic-sense/
16. TrendForce. (September 26, 2025). China’s “Big Fund III” Invested RMB 450 Million in Semiconductor Equipment. https://www.trendforce.com/news/2025/09/26/news-chinas-big-fund-iii-invested-rmb-450-million-in-semiconductor-equipment/
17. Semiconductor Industry Association. (July 13, 2021). Taking Stock of China’s Semiconductor Industry. https://www.semiconductors.org/taking-stock-of-chinas-semiconductor-industry/
18. China Britain Business Council. (June 26, 2025). China’s Semiconductor Sector. https://focus.cbbc.org/chinas-semiconductor-sector/
19. CSIS. (December 16, 2024). Understanding the Biden Administration’s Updated Export Controls. https://www.csis.org/analysis/understanding-biden-administrations-updated-export-controls
20. Akin Gump. (October 2023). Commerce Imposes Significant New Controls on Advanced Semiconductors. https://www.akingump.com/en/insights/alerts/commerce-imposes-significant-new-controls-on-advanced-semiconductors
21. ServeTheHome. (August 22, 2022). Biren BR100 GPU for Datacenter Compute and AI Workloads. https://www.servethehome.com/biren-br100-gpu-for-datacenter-compute-and-ai-workloads/
22. HPCwire. (August 23, 2022). Chinese Startup Biren Details BR100 GPU. https://www.hpcwire.com/2022/08/22/chinese-startup-biren-details-br100-gpu/
23. Jon Peddie Research. (November 24, 2022). China’s Biren to feel the heat of US sanctions.https://www.jonpeddie.com/blog/chinas-biren-to-feel-the-heat-of-us-sanctions/