Realtime AI News
China's tech giants reportedly order over $35B of Nvidia GPUs to work around US curbs
Chinese tech giants including ByteDance, Alibaba and Tencent are placing huge Nvidia GPU orders to work around US export limits, Chosunbiz reported on October 21. The orders center on the export-compliant RTX PRO 5500 workstation card and reportedly exceed 35 billion dollars, though supply may cover only a fraction of the demand.

Chinese technology giants are flooding Nvidia with GPU orders, using the purchase volume of export-compliant parts to offset the compute they cannot buy at the top of the range, Chosunbiz reported on October 21. ByteDance, Alibaba and Tencent are named as the main buyers behind the wave.
The orders center on the RTX PRO 5500, a workstation card that fits inside current US export rules. A Reuters exclusive put the combined value of the three companies' orders at more than 35 billion dollars.
The volume is the striking part. Reports cited by Tom's Hardware described roughly three million cards, while TrendForce was more cautious, estimating that only five to ten percent of the orders could actually be filled. The gap between paper demand and delivered hardware is enormous.
The workaround is straightforward in logic and painful in practice: buy the best export-compliant card available, then buy millions of them. Where cutting-edge data center accelerators are walled off, scale becomes the substitute for per-chip performance, and clusters are assembled out of parts that were never designed for the job.
Nvidia's own framing matters here. According to The Next Web, the company described the curbs as outdated, arguing that the current thresholds no longer reflect the real technology and market landscape. That is both a complaint and a bid for more policy room to sell compliant products into China.
For the supply chain, the surge creates two problems at once. Global capacity has to be split across other buyers, so Chinese orders are unlikely to be prioritized, and a compliant purchase of this size inevitably tests how much slack the existing rules actually contain.
What to watch next is delivery pace, whether Chinese cloud providers steer these cards toward inference and industry-specific models rather than frontier training, and whether regulators respond by tightening the list to cover workstation-class parts as well. For domestic chip suppliers, the shortage window is also an opening.
Why it matters
The order book shows export controls being diluted by sheer volume, while shifting the real constraint to supply and policy: how much can be delivered, and whether regulators broaden the restricted list.
Nearby Updates
All10/21, 17:30
GTC Berlin 2026 Opens With Huang's Keynote, and Agents and Robots at the Top of the Agenda
NVIDIA opened GTC Berlin 2026 on October 21 with a keynote from founder and CEO Jensen Huang at the Tempodrom, as the three-day event hosts more than 100 sessions on agentic AI, robotics, physical AI, open models and accelerated computing. Journalists, analysts and partners gathered for the main address, which set the tone for how NVIDIA wants to be positioned in Europe.
10/21, 18:00
NVIDIA Expands Its Open-Model Family With Nemotron 3.5 Super VL, Bolt and a New Retriever Line
NVIDIA has expanded its open-model family across language, robotics, biology and quantum computing, adding the Nemotron 3.5 Super VL and Nemotron 3.5 Bolt models plus a new Nemotron Retriever line for embedding and reranking. The company frames the release as a way for developers, researchers and enterprises to build applications tuned to local languages and industries, with Europe as the explicit backdrop.
10/21, 18:02
NVIDIA Publishes OpenShell Reference Architecture to Govern AI Agent Workloads at Scale
NVIDIA's developer blog detailed an OpenShell Reference Architecture that pairs an open-source runtime with Vera CPUs and BlueField-4 hardware isolation to control large-scale AI agent workloads. Its argument is that agents cannot govern themselves, so permissions and sandbox boundaries must be enforced by infrastructure.
10/21, 13:45
Yelp sells a $30-a-month AI agent that answers and qualifies leads
Yelp is charging $30 a month for an AI agent that answers inquiries from potential customers and qualifies the resulting leads, according to PPC Land. The low subscription price moves agent software out of pilots and into a line item small local businesses can buy like any other SaaS add-on.