The Tool Desk
Outbyte PC Repair FREEClear out junk files and repair common Windows errorsFree Scan →Outbyte Driver Updater FREEScan for outdated or missing drivers - takes under a minuteDriver Scan →iTechGuides is reader-supported. When you buy through links on our site, we may earn an affiliate commission. As an Amazon Associate I earn from qualifying purchases. Learn more
NVIDIA designed the H20 as a data-center AI accelerator for China that could fit within U.S. export restrictions. Its prospects have since been shaped as much by government decisions as by chip performance: the United States required a license for H20 exports in April 2025, NVIDIA announced a conditional return to sales that July, and reporting in June 2026 described its China sales as stalled while Huawei and other domestic suppliers gained ground.
Why did NVIDIA make the H20 for China?
The H20 is a data-center AI integrated circuit designed specifically for the China market under U.S. export constraints. NVIDIA’s strategy was to offer Chinese customers an accelerator that could be sold within the rules then in effect, rather than simply offer the same products available in less-restricted markets.
That made the H20 a commercial product with a regulatory dependency: its market depended on which chips U.S. authorities would license for export and whether Chinese customers could and would buy them. A chip designed for a particular export regime could lose access to its target market when that regime changed.
What happened to H20 exports?
The key events show how quickly the H20’s sales outlook changed. In its fiscal 2026 first-quarter filing, NVIDIA said that on April 9, 2025, the U.S. government informed it that H20 exports to China and certain other destinations required a license.
#1 Best Overall
- Memory Size: 16 GB GDDR6 ECC.
- Memory Bus Width: 128-bit.
- Memory Bandwidth: 200 GB/s.
- CUDA Cores: 1280.
- Peak Single Precision floating point performance: 18 Tflops (GPU Boost Clocks).
| Date | What happened | Why it matters |
|---|---|---|
| Before the new licensing requirement in April 2025 | NVIDIA reported $4.6 billion in H20 sales for fiscal 2026’s first quarter. | This is the company’s reported sales figure for that quarter, not an estimate of future shipments. |
| April 9, 2025 | The U.S. government told NVIDIA that exports of H20 chips to China and certain other destinations required a license. | The requirement restricted shipments; it was not simply a product-performance issue. |
| Fiscal 2026 first quarter | NVIDIA reported a $4.5 billion charge for excess inventory and purchase obligations tied to diminished H20 demand. | The charge reflected inventory and purchase commitments affected by the change; it should not be read as $4.5 billion in lost H20 sales. |
| July 14, 2025 | NVIDIA announced that it would resume H20 sales to China and introduce a fully compliant China GPU. | The announcement indicated a path to renewed sales, subject to regulatory conditions. |
| July 18, 2025 | China’s Ministry of Commerce published a briefing on U.S. approval of H20 sales. | The chip’s availability involved both U.S. export licensing and Chinese policy. |
| July 31, 2025 | China’s Cyberspace Administration said it had summoned NVIDIA over alleged H20 backdoor and security risks and requested explanations and evidence. | The regulator’s allegations were a further point of uncertainty for the chip’s acceptance in China; the allegations themselves do not establish that a backdoor existed. |
| June 29, 2026 | The Associated Press reported that NVIDIA’s China AI-chip sales had stalled while Huawei and other domestic suppliers gained ground. | That reporting describes a changing market, not a permanent outcome. |
The sales and charge figures come from NVIDIA’s reporting for fiscal 2026’s first quarter. They describe different things: reported H20 sales and a charge related to excess inventory and purchase obligations after the licensing change. They should not be combined into a single measure of the ban’s cost.
Did the United States ban H20 sales to China?
On April 9, 2025, the U.S. government told NVIDIA that H20 exports to China and certain other destinations required a license. That imposed a licensing requirement and interrupted the prior sales path; it is more precise to describe the action as requiring a license than to say that all H20 sales were permanently banned.
NVIDIA announced on July 14, 2025, that it would resume sales to China. China’s Ministry of Commerce addressed U.S. approval four days later. Those announcements show that permission to sell could change, but they do not establish that every shipment was licensed, that all Chinese buyers could procure the chip, or that sales resumed at scale.
Recommended Free Tools
Rank #2
- Professional GPU Acceleration: Powered by NVIDIA Tesla K20 architecture designed specifically for high-performance computing and computational workloads, delivering exceptional parallel processing capabilities for scientific simulations, data analytics, and complex calculations
- Generous Memory Capacity: Equipped with 5GB of GDDR5 SDRAM providing ample high-speed memory bandwidth to handle large datasets and memory-intensive applications, ensuring smooth performance for demanding computational tasks and professional visualization work
- High-Speed Connectivity: Features PCI Express 2.0 x16 interface enabling fast data transfer rates between the GPU and system, minimizing bottlenecks and maximizing throughput for time-sensitive computational operations and rendering tasks
- Single GPU Configuration: Contains one dedicated GPU core optimized for parallel processing workloads, making it ideal for workstations and servers requiring accelerated computing performance for engineering, research, and professional applications
- Renewed Quality Product: This professionally renewed graphics card has been inspected and tested to work and look like new, offering a cost-effective solution for adding GPU acceleration capabilities to your computing infrastructure
Can NVIDIA’s H20 compete with Huawei Ascend chips?
There is no single public, independently standardized, apples-to-apples benchmark in the available evidence that proves the H20 beats or loses to every Huawei Ascend configuration. “Compete” also depends on more than peak compute: buyers need to know whether a chip is eligible to import and procure, works with their software, can scale across a cluster, and can be supplied and operated economically.
| Procurement factor | What can be established about H20 | What can be established about Huawei Ascend |
|---|---|---|
| Export and procurement eligibility | U.S. authorities required a license for H20 exports to China and certain other destinations from April 9, 2025. NVIDIA announced a resumption of sales in July, but that announcement does not establish eligibility for every shipment or buyer. | Not stated in the cited public reporting for a like-for-like procurement comparison. |
| Compute and memory resources | Not stated in the cited evidence in a way that supports comparison across configurations. | Not stated in the cited evidence in a way that supports comparison across configurations. |
| Software and developer compatibility | Not stated in the cited evidence in a form that supports a direct comparison. | Not stated in the cited evidence in a form that supports a direct comparison. |
| Interconnect and cluster scaling | Not stated in the cited evidence in a form that supports a direct comparison. | Not stated in the cited evidence in a form that supports a direct comparison. |
| Power, cooling and total cost of ownership | Not stated in the cited evidence in a form that supports a direct comparison. | Not stated in the cited evidence in a form that supports a direct comparison. |
| Supply reliability and political or security acceptance | Sales depended on U.S. licensing, and China’s cyberspace regulator publicly raised alleged security risks in July 2025. | Associated Press reporting dated June 29, 2026, described Huawei and other domestic suppliers as gaining ground; it did not establish a universal advantage in supply or security acceptance. |
For a real procurement decision, the relevant comparison is between specific chip configurations and the systems a buyer can actually obtain, license, deploy, and support. A market-share estimate or a single benchmark cannot answer all of those questions.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.Is Huawei replacing NVIDIA in China’s AI-chip market?
Not conclusively. The Associated Press reported on June 29, 2026, that NVIDIA’s China AI-chip sales had stalled while Huawei and other local suppliers gained ground. The report cited a Bernstein estimate that NVIDIA and Huawei each held about 40% of China’s AI-chip market in 2025. That is an attributed estimate for a particular year, not an official measure of current sales or proof that one company has permanently displaced the other.
Rank #3
- Item Package Dimension -14.7L X 8.8W X 3.4H Inches
- Item Package Weight - 2.4 Pounds
- Item Package Quantity - 1
- Product Type - Video Card
The same reporting supports a narrower conclusion: Chinese domestic suppliers had momentum while NVIDIA’s sales were stalled. Whether that turns into lasting replacement depends on future licensing decisions, procurement policy, customers’ software and infrastructure choices, and the ability of suppliers to deliver systems at scale.
What Jensen Huang said about the trade-off
As quoted by the Associated Press, NVIDIA CEO Jensen Huang said: “We have to have, number one, make sure that we have national security and that we protect our nation, but also simultaneously should go and compete and grow our technology industry and maximize our exports.” His statement captures the policy tension behind products such as H20: the United States seeks to control sensitive technology exports while companies argue that remaining in a major market matters for competition and the technology industry.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
