Nvidia, a global leader in graphics processing units (GPUs) and artificial intelligence (AI) hardware, is reportedly developing a new AI chip tailored for the Chinese market. This move comes as the company navigates stringent U.S. export controls while aiming to maintain its foothold in one of the world’s largest technology markets. The new chip, based on Nvidia’s latest Blackwell architecture, is designed to comply with U.S. regulations while offering enhanced performance compared to Nvidia’s current China-compliant H20 chip. This article explores the details of Nvidia’s initiative, its implications for the Chinese AI market, and the broader geopolitical context shaping this development.
Background on U.S. Export Controls and Nvidia’s Challenges
U.S. Restrictions on Advanced AI Chips
Since 2022, the U.S. government has imposed increasingly stringent export controls on advanced semiconductors, particularly those used for AI workloads, to limit China’s access to cutting-edge technology. These restrictions stem from national security concerns, as advanced AI chips could potentially enhance China’s military capabilities. In 2023, the Biden administration tightened these controls, prompting Nvidia to develop chips like the H20, L20, and L2, which have reduced computing power to comply with U.S. regulations.
In April 2025, the Trump administration briefly banned sales of the H20 chip to China, forcing Nvidia to write off $5.5 billion in inventory and forgo $15 billion in potential sales. However, recent policy shifts have allowed Nvidia to resume H20 shipments, with the U.S. government granting export licenses and reportedly requiring Nvidia to pay 15% of its China-related revenue to the U.S. government—an unprecedented arrangement.
Nvidia’s Market Position in China
China represents a significant market for Nvidia, accounting for 13% of its revenue in the fiscal year ending January 2025. However, U.S. export curbs have slashed Nvidia’s market share in China from 95% before 2022 to approximately 50% today. Domestic competitors like Huawei, with its Ascend 910B chip, and startups like Enflame are gaining ground, fueled by China’s push for technological self-sufficiency.
Details of Nvidia’s New AI Chip for China
The B30A Chip: A Blackwell-Based Solution
Nvidia is reportedly developing a new AI chip, tentatively codenamed B30A, based on its flagship Blackwell architecture unveiled in March 2024. Unlike the dual-die design of the premium B300 Blackwell GPU, the B30A uses a single-die configuration and is approximately half as powerful. Despite this, it is expected to outperform the H20 chip, which is based on the older Hopper architecture. The B30A will incorporate features like fast data transmission, NVLink support, and high-bandwidth memory, making it a competitive option for Chinese AI developers.
The chip is designed to comply with U.S. export controls, particularly the latest restrictions on GPU memory bandwidth, which cap performance at 1.7–1.8 terabytes per second. The B30A is expected to use GDDR7 memory technology, achieving approximately 1.7 terabytes per second, and will not rely on advanced Chip-on-Wafer-on-Substrate (CoWoS) packaging, simplifying production and reducing costs.
Pricing and Production Timeline
The B30A is projected to be priced between $6,500 and $8,000, significantly lower than the H20’s $10,000–$12,000 price tag. This lower cost reflects its simpler manufacturing requirements and reduced specifications. Mass production is expected to begin in June 2025, with shipments to China planned for the second quarter of 2025. Nvidia is collaborating with Inspur, a major Chinese distributor, to launch and distribute the chip, tentatively named the B20 or possibly the 6000D.
Comparison with Existing Chips
The B30A is positioned as a step up from the H20, which was designed to bypass earlier export controls but faced delays due to integration issues and a temporary ban. While the H20 is now considered “obsolete” by some, including former U.S. President Donald Trump, it remains a critical revenue driver for Nvidia in China, with the company ordering 300,000 H20 chipsets from Taiwan Semiconductor Manufacturing Co. (TSMC) to meet robust demand. The B30A, however, promises improved performance, leveraging the Blackwell architecture’s advancements, such as up to 30 times faster processing for certain tasks compared to its predecessors.
Implications for the Chinese AI Market
Competitive Landscape
The introduction of the B30A could help Nvidia regain some of its lost market share in China, which has been eroded by domestic competitors. Huawei’s Ascend 910B, for instance, is gaining traction, and Chinese companies are investing heavily in homegrown AI infrastructure. Analysts predict Nvidia’s market share in China could drop to 54% in 2025 from 66% in 2024, partly due to competition and regulatory scrutiny. The B30A’s lower price and enhanced capabilities may counter this trend, but Nvidia faces challenges from Beijing’s push to prioritize domestic chips.
Chinese Regulatory Concerns
Chinese authorities have expressed concerns about Nvidia’s H20 chips, particularly regarding potential security risks like “backdoors” that could compromise user data. In July 2025, the Cyberspace Administration of China (CAC) summoned Nvidia to address these issues, and state media have criticized the H20 for lacking technological advancement. The CAC has also urged companies like Tencent, ByteDance, and Baidu to justify their purchases of Nvidia chips and consider domestic alternatives, especially for government-related applications. These actions reflect China’s broader strategy to reduce reliance on U.S. technology.
Geopolitical Dynamics
Nvidia’s development of the B30A occurs amid complex U.S.-China trade dynamics. The Trump administration’s decision to allow H20 sales, coupled with the revenue-sharing agreement, marks a shift from earlier restrictions, possibly driven by trade negotiations and Nvidia’s lobbying efforts. CEO Jensen Huang has argued that export controls have backfired, accelerating Chinese companies’ development of domestic alternatives. The B30A’s introduction could further complicate U.S.-China relations, as critics argue it may provide China with advanced computing power, potentially undermining U.S. national security goals.
Nvidia’s Broader Strategy
Maintaining Market Leadership
Nvidia’s efforts to develop China-specific chips reflect its determination to remain a leader in the global AI market. CEO Jensen Huang has emphasized the importance of competing in China, warning that continued export curbs could drive Chinese customers to rivals like Huawei. By offering compliant yet competitive chips like the B30A, Nvidia aims to balance regulatory compliance with market access. The company’s CUDA platform remains a key advantage, enabling seamless integration of AI clusters, which domestic competitors have yet to fully replicate.
Investment in Domestic Production
Nvidia is also aligning with U.S. policies to boost domestic manufacturing. Following meetings with the Trump administration, Nvidia has committed to investing approximately $500 billion in U.S.-based chip manufacturing infrastructure. This aligns with broader U.S. efforts to reduce reliance on foreign supply chains and could mitigate future trade tensions.
Nvidia’s development of a new Blackwell-based AI chip for the Chinese market underscores its strategic adaptability in a challenging geopolitical landscape. The B30A chip, with its competitive pricing and compliance with U.S. export controls, aims to bolster Nvidia’s position in China while navigating regulatory and competitive pressures. However, the company faces hurdles from Chinese authorities’ scrutiny and the rise of domestic competitors like Huawei. As the U.S.-China technology race intensifies, Nvidia’s ability to balance innovation, compliance, and market demands will be critical to its success.
