Jensen Huang confirms there are no plans to ship Blackwell GPUs to China right now, chipmaker at Beijing's mercy — Nvidia CEO says shipments haven't been approved by Chinese authorities

News
By published

"There are no active discussions."

Jensen Huang
(Image credit: Getty / Woohae Cho)

Jensen Huang is in Taiwan for TSMC's sports day event right now. Before attending, he spoke to local media, giving insight on various ongoing developments surrounding Nvidia and its relationship with AI constituents — including the prospect of selling Blackwell chips to China. President Trump and U.S. Treasury Secretary have already made it clear that Blackwell is reserved for America first, and Huang has just echoed those sentiments by clarifying that "there are no active discussions" surrounding this matter.

That said, beyond the White House, Huang reminds us that it's ultimately up to China's discretion whether it wants to deal with Nvidia. Beijing has recently banned its firms from dealing with the chipmaker to focus on local, homegrown silicon as the region grows toward AI independence. Blackwell still makes its way into the country, however, through illicit channels.

The previous-gen, Hopper-based H20 is the top-end AI GPU Nvidia sells in China, and there are reports of a Blackwell-based B30A model launching soon. If that comes to fruition and China allows Nvidia to sell it, the chipmaker will have to share 15% of the revenue from all sales with the White House, and that's after export licenses have been granted.

Nvidia

(Image credit: Nvidia)

A dense layer of red tape covers this AI race, which is informing semiconductor trade policy at the moment. Meanwhile, Huang has even claimed that China is aggressively tailing the U.S. in artificial intelligence — if not already ahead — and that most of the popular open-source models in the world today are from China. The region also produces some 50% of all AI researchers globally, implying the competition is stiff.

Even though neither Washington nor Beijing wants to engage with each other on cutting-edge AI silicon, both heads of state struck a historic trade truce last month that's said to last a year. This positive development briefly led the markets to think that a chip breakthrough might be on the cards, though it's now clear the two nations don't even want to talk about it.

Google Preferred Source

Follow Tom's Hardware on Google News, or add us as a preferred source, to get our latest news, analysis, & reviews in your feeds.

Hassam Nasir
Hassam Nasir
Contributing Writer

Hassam Nasir is a die-hard hardware enthusiast with years of experience as a tech editor and writer, focusing on detailed CPU comparisons and general hardware news. When he’s not working, you’ll find him bending tubes for his ever-evolving custom water-loop gaming rig or benchmarking the latest CPUs and GPUs just for fun.

2 Comments Comment from the forums
  • oicu812
    The Chinese already have ways to import chips into the country bypassing the embargo. They don't need permission from anyone nor do they care.
    Reply
  • Moxylite
    China/Alibaba currently has a homegrown H20 Hopper equivalent (T-head) for 40% cheaper right now. My understanding regarding AI; quantity over quality can make a bigger difference than usual?
    FeatureNVIDIA Hopper ArchitectureNVIDIA Blackwell ArchitectureTransistor Count80 billion transistors208 billion transistorsManufacturing ProcessTSMC 4N processCustom-built TSMC 4NP processTransformer EngineFirst-generationSecond-generationDecompression EngineNoYesEnergy EfficiencyImproved over the previous generation25x more energy-efficient than HopperInterconnect TechnologyFourth-generation NVLinkFifth-generation NVLinkChip-to-Chip Interconnect900 GB/s10 TB/sApplicationsGenerative AI, LLMs, Data processing, Quantum computingAccelerated Computing, AI, LLM
    ""With the T-Head PPU, Alibaba aims to capture market share by matching the performance profile of the most powerful Nvidia hardware currently permitted in China: the H20 (Hopper). The specs show that the PPU’s Bill of Materials (BOM) cost per card is “40% lower… compared to the Nvidia H20.”This operational advantage enables Alibaba Cloud to reduce public cloud inference prices by “50%,” directly translating the hardware efficiency into a market advantage.This domestic hardware strategy has already secured significant market validation. State-owned telecommunications operator China Unicom is utilizing the PPU as the primary AI chip in its new $390 million data center project in Xining. Despite the focus on replacing Nvidia’s chips, Alibaba has taken a pragmatic approach to the software ecosystem. The T-Head PPU features compatibility for existing Nvidia software frameworks, a clear reference to the ubiquitous CUDA platform.""
    Reply