Why trust Greenbot
Nvidia plans to slash prices on a new artificial intelligence (AI) chip designed specifically for China, cutting costs by up to 35% compared to its recently banned model as the company fights to maintain access to a $50 billion market.
The graphics chip maker will launch the new processor in June at $6,500 to $8,000. This price sits well below the $10,000 to $12,000 cost of its H20 model that U.S. authorities blocked in April. Mass production begins as early as next month, according to sources who spoke to Reuters.
Nvidia built the chip using its newest Blackwell architecture but stripped out advanced components to satisfy American export rules. The processor uses standard GDDR7 memory instead of high-speed memory and avoids Taiwan Semiconductor’s advanced packaging technology.
“Until we settle on a new product design and receive approval from the U.S. government, we are effectively foreclosed from China’s $50 billion data center market,” an Nvidia spokesperson told Reuters.
This move marks Nvidia’s third attempt to create China-specific chips since trade limits began in 2022. Export controls have damaged the company’s Chinese operations, forcing Nvidia to write off $5.5 billion in inventory and abandon $15 billion in potential sales.
Nvidia’s Chinese market share has collapsed from 95% before 2022 to just 50% today. CEO Jensen Huang warned last week that continued limits could push more Chinese customers toward local competitors like Huawei. Huawei makes the competing Ascend 910B chip.
China made up 13% of Nvidia’s total revenue last year. This makes it a critical market that he company cannot afford to lose entirely. The new chip’s memory speed will hit about 1.7 terabytes per second. This stays just within the 1.8 terabyte limit set by U.S. regulators.
“Domestic Chinese technologies like Huawei are expected to catch up with the computing performance of downgraded versions within one to two years,” said Nori Chiou, an investment director at Singapore-based White Oak Capital Partners.
Nvidia maintains an advantage through its CUDA programming platform, which engineers worldwide use to build AI applications. This software system keeps developers tied to Nvidia chips even when competitors offer similar hardware performance.
The company plans to release a second Blackwell-based chip for China in September, though specifications remain undisclosed. Both processors represent Nvidia’s strategy of compliance over confrontation as U.S.-China tech tensions continue escalating.
Recent export limits target memory speed specifically. Officials recognize its importance for AI training and data processing tasks. The new limits force Nvidia to work around technical constraints while keeping enough performance to compete against Chinese alternatives.
Huang recently called the export limits ineffective. He argued that they only speed up China’s development of local chip technology. His company now faces the challenge of staying profitable in a market where both prices and performance face artificial limits.
