Related Articles
Ask anything about stocks
China-Focused Nvidia Chip Gets Lukewarm Reception From Corporates
Table of Contents
Nvidia’s attempt to make AI chips for China under stricter U.S. export rules appears to be running into cold water. The RTX6000D, a China-tailored AI inference chip, is getting only modest demand from big tech firms. Corporates are holding off, waiting for more powerful models, and evaluating whether the steep price is worth it.
Nvidia Chips in China: what is RTX6000D
Nvidia designed the RTX6000D to comply with U.S. export restrictions. It is aimed mainly at AI inference tasks rather than heavy training. It ships with conventional GDDR memory and has bandwidth just under the 1.4 TB/s limit imposed by regulators. It carries a hefty price tag of about 50,000 yuan (≈ US$7,000).
The RTX6000D was among the first of Nvidia’s “degraded but compliant” chips meant for China after export curbs cut off access to top-tier models like the RTX5090. The idea: supply Chinese demand within regulatory limits. But corporates aren’t rushing in. (investing.com)
Why did Nvidia design special chips for China?
U.S. export restrictions force Nvidia to build versions of its chips that comply with rules, so it can sell in China without violating the law.
Reception lukewarm: demand and corporate hesitation
Major Chinese firms like Alibaba, Tencent, ByteDance are reportedly opting not to place orders for the RTX6000D (or placing only small ones). Some of the reasons cited:
- Price vs performance: The RTX6000D is seen as expensive for what it offers. It underperforms the RTX5090 on key benchmarks. Yet the older RTX5090 is still available via grey market channels, often at less than half the price.
- Awaiting H20 and B30A models: Corporations are waiting for shipments of the H20 chip, which got U.S. approval in July but has not fully resumed delivery. Even more anticipation surrounds the B30A, Nvidia’s next-gen chip that may offer far higher performance.
- Regulatory and political uncertainty: Chinese regulators are asking firms to clarify purchases of Nvidia hardware, especially H20. There are concerns about “backdoor” risks, export compliance, and whether newer chips will indeed pass U.S. licensing.
How are Chinese firms reacting to Nvidia Chips?
Cautiously, many are sitting on the sidelines until more powerful and approved chips arrive or price/performance improves.
Why Nvidia Chips are struggling in China: deeper causes
Several forces combine to make the reception lukewarm:
- Grey market access to prohibited chips: Even though the RTX5090 is officially banned, it remains available through unofficial channels. Firms compare performance and cost and often find the banned model more attractive for price/performance.
- High cost and limited gains: RTX6000D offers lower bandwidth and simpler memory than top models. For inference workloads, some may accept that, but firms producing large models or complex AI tasks expect more. If performance is limited, paying a premium is less justifiable.
- Domestic competition: Chinese firms are developing their own chipsets. Alibaba, Baidu, Huawei, Cambricon, and others are pushing local designs. Use of home-grown models is increasing.
- Regulatory friction: Even when U.S. export licenses are granted, delays in shipping, uncertainty over compliance, and scrutiny over security concerns slow adoption. Uncertainty weighs heavily in corporate procurement decisions.
Nvidia Chips: waiting for H20, anticipation for B30A
RTX6000D is not the only chip in play. Chinese tech firms are especially focused on:
- H20 chip – allowed again by U.S. regulators in July. Companies want assurance that orders will be processed and shipments resumed.
- B30A chip – expected to deliver significantly higher performance than the H20 (estimates suggest up to six times more powerful) if approved. It is based on Nvidia’s Blackwell architecture. If B30A clears exports, it may shift demand away from RTX6000D.
These models carry both promise and risk: promise of better power; risk of further delays, higher cost, or regulatory setbacks.
What analysts say, and the competitive landscape
Sell-side analysts had earlier projected strong demand: JPMorgan estimated Nvidia might produce 1.5 million units in the second half of 2025; Morgan Stanley thought as many as 2 million units of RTX6000D might be in the pipeline. The weak uptake shocks those expectations.
Competition is heating up. Chinese domestic chip makers are improving rapidly. Some firms prefer local or alternative GPUs that work well enough, especially when coupled with cost, regulatory ease, and software compatibility.
What does this mean for the global AI chip market?
Nvidia’s dominance may face pressure in China if it cannot deliver better chips that meet both performance and regulatory compliance. Local alternatives may grow faster.
Broader implications: global AI dominance and strategic risk
Nvidia’s business depends heavily on global markets. China alone represented ~13 percent of its revenue in its latest fiscal reports. Problems in China weaken Nvidia’s growth story. Also, export restrictions from the U.S. and regulatory pushback in China mean Nvidia must navigate legal, political, and technological constraints simultaneously.
Investors worry that weak sales of model RTX6000D may erode margins, especially if lower-cost domestic chips capture more of China’s demand. If Nvidia fails to meet expectations on H20 and B30A, local chipmakers’ appeal grows stronger.
What does this mean for the global AI chip market?
Regional restrictions may fragment the market, giving an advantage to firms that can produce compliant yet competitive chips or those that dominate within countries.
Forward-looking: Can Nvidia regain momentum in China
For Nvidia to turn the tide, it must act on several fronts:
- Improve value offer: lower the price or enhance the performance of compliant chips.
- Speed up shipping: ensure H20 chip shipments resume and B30A gets approved smoothly.
- Strengthen domestic trust: address regulatory fears, reassure about security, and build relations with Chinese firms and regulators.
- Adapt to competition: work to integrate with local AI ecosystems, and possibly partner with local chip makers or license technology.
If it succeeds, Nvidia could re-establish steady growth in China. If not, it risks losing ground to local competitors, missing out on a critical market in AI development.
FAQs
Why are Nvidia Chips less popular in China right now?
Because the RTX6000D is seen as expensive compared with its performance, it is compared with banned chips like RTX5090 via grey market channels. Also, firms are delaying until higher spec chips like H20 or B30A receive full regulatory clearance.
What is the Nvidia RTX6000D chip made for?
It is built for AI inference tasks under U.S. export restriction limits, with bandwidth just under 1.4 TB/s, using GDDR memory. Its design sacrifices some performance in training and high-throughput applications in favor of regulatory compliance.
Are Chinese companies adopting Nvidia Chips widely?
Not yet, large tech firms are holding back on bulk orders until more powerful models arrive. Some are using domestic chips, especially for non-core workloads, to reduce dependency.
How do U.S. export restrictions affect Nvidia’s sales in China?
They force Nvidia to reengineer its chips to fall below performance and bandwidth thresholds. Approvals, licenses, and compliance issues create delays and raise costs for both Nvidia and its Chinese clients.
What alternatives are Chinese firms exploring instead of Nvidia Chips?
They are developing chips from companies like Huawei, Cambricon, and others, improving domestic hardware and pushing self-reliance. They also leverage older banned models through grey markets or use downgraded versions that comply with regulations.
Disclaimer:
This is for information only, not financial advice. Always do your research.