Nvidia has tightened its sales terms for Chinese customers seeking its H200 artificial intelligence chips, requiring full upfront payment as the U.S. chipmaker moves to shield itself from regulatory uncertainty surrounding Beijing’s approval of the shipments, according to two people briefed on the matter who spoke to Reuters.
Under the new conditions, Chinese buyers must pay the entire order value in advance, with no option to cancel, seek refunds, or change chip configurations after orders are placed. In limited cases, customers may substitute cash payments with commercial insurance or asset-backed collateral, one of the people said. The sources spoke on condition of anonymity because the policy is not public.
While Nvidia has long required advance payments from Chinese clients, the H200 terms mark a sharp tightening. Previously, some customers were allowed to secure orders with deposits rather than full payment. For the H200, Nvidia has enforced stricter conditions due to uncertainty over whether Chinese regulators will ultimately approve the imports, one of the people said.
Register for Tekedia Mini-MBA edition 19 (Feb 9 – May 2, 2026): big discounts for early bird.
Tekedia AI in Business Masterclass opens registrations.
Join Tekedia Capital Syndicate and co-invest in great global startups.
Register for Tekedia AI Lab: From Technical Design to Deployment (next edition begins Jan 24 2026).
The tougher stance comes as demand in China for the H200 has surged. Chinese technology firms have placed orders for more than 2 million of the chips, priced at about $27,000 each, according to a Reuters report last month. That volume far exceeds Nvidia’s current inventory of roughly 700,000 units, highlighting the scale of unmet demand.
The H200 has become a focal point for Chinese companies racing to build and train large AI models. Domestic alternatives, including Huawei’s Ascend 910C, have made progress, but their performance still trails Nvidia’s H200 in large-scale AI training tasks, making Nvidia’s hardware highly sought after among China’s internet and technology giants.
Regulatory uncertainty on the Chinese side remains a key obstacle. Bloomberg reported on Thursday that Beijing plans to approve some H200 imports as early as this quarter, allowing purchases for selected commercial uses while barring the military, sensitive government agencies, critical infrastructure operators, and state-owned enterprises due to security concerns. In the meantime, Chinese regulators have asked some tech firms to temporarily pause H200 orders while officials determine how many domestically produced chips each buyer must purchase alongside each Nvidia order, one of the sources said. The Information first reported that pause earlier this week.
Nvidia’s payment structure effectively transfers the risk of regulatory delays or rejections from the seller to the buyer. Chinese customers are being asked to commit significant capital without certainty that approvals will be granted or that the chips can be deployed as planned. For large orders, that can mean billions of dollars tied up while regulators deliberate.
The company’s caution is shaped by recent experience. Last year, Nvidia wrote down $5.5 billion in inventory after the Trump administration abruptly banned sales of its H20 chip to China, then the most powerful product it was permitted to sell there. Although that U.S. restriction was later reversed, China subsequently banned H20 shipments, leaving Nvidia exposed. That episode appears to have informed the company’s decision to harden its commercial terms for the H200.
The policy shift also unfolds against a rapidly changing geopolitical backdrop. The Biden administration had imposed sweeping restrictions on advanced AI chip exports to China. President Donald Trump reversed that stance last month, allowing H200 sales to China subject to a 25% fee payable to the U.S. government. While that opened the door for renewed trade, it left final approval squarely in Beijing’s hands.
Nvidia chief executive Jensen Huang said on Tuesday that demand for the H200 was “quite high” and that the company had “fired up our supply chain” to increase output. Huang added that he did not expect China’s government to issue a formal declaration approving the chips.
“If the purchase orders come, it’s because they’re able to place purchase orders,” he said, underscoring the informal nature of the approval process.
The company plans to fulfil initial Chinese orders from existing stock, with the first batch of H200 chips expected to arrive before the Lunar New Year holiday in mid-February, Reuters reported previously. Nvidia has also approached Taiwan Semiconductor Manufacturing Co about ramping up H200 production, with additional manufacturing expected to begin in the second quarter of 2026 to meet Chinese demand.
Expanding capacity, however, is not straightforward. Nvidia is in the midst of transitioning from its current flagship Blackwell architecture to the more advanced Rubin platform, while competing with customers such as Alphabet’s Google for cutting-edge production slots at TSMC. That competition for manufacturing capacity adds another layer of complexity as Nvidia tries to balance global demand with political and regulatory risk.
Currently, Nvidia’s strict upfront payment requirement signals its determination to pursue Chinese sales while insulating itself from sudden policy reversals.