← The Polylog AI Intelligence Brief
Morning Edition · Thursday, October 22, 2026Published at 2:23 AM EDT · New York
ByteDance, Alibaba and Tencent Order More Than $35 Billion of Nvidia's Export-Compliant RTX PRO 5500
Total orders from China have passed 6 million cards, but Nvidia can make no more than 3 million worldwide in 2027 and has told partners they may receive 5 to 10 percent of what they requested.
China's three largest AI buyers have ordered more than 3 million Nvidia RTX PRO 5500 cards worth over $35 billion, Reuters reported, as summarized by AI Post. Nvidia designed the card to comply with US export restrictions. The constraint is supply. Orders from Chinese customers have passed 6 million cards, while Nvidia can make no more than 3 million worldwide in 2027. Nvidia has told partners they may receive only 5 to 10 percent of their requested volumes, TrendForce reports. Samples are due in December and commercial shipments in the first quarter of 2027.
The RTX PRO 5500 is a Blackwell-architecture professional card with 84 GB of GDDR7 memory. It is not one of Nvidia's high-bandwidth-memory data-center accelerators. Its appeal in China comes from its legal status, not its specifications. It is among the most capable Nvidia parts that can be sold there legally. Hyperscalers want it for inference on large models and agent workloads, where memory capacity matters more than the interconnect bandwidth needed for frontier pretraining.
The Chinese government's own stance is a second constraint on these orders. In late September, The Information reported that China's Ministry of Industry and Information Technology (MIIT) had told ByteDance, Alibaba and others that it intended to approve purchases, according to a summary by AI Weekly. Reuters said at the time that it could not verify that report, and no ministry has confirmed criteria or volumes.
Veracity: Plausible
72/100- If true, who benefits
Nvidia gains a large order book and pricing power in China, ByteDance, Alibaba and Tencent gain a legal source of inference compute, and Chinese domestic accelerator makers stand to lose, as the drop in Chinese chip shares showed when reports of approval first appeared.
Part of a tracked trend
AI Sovereignty and Export Controls on Frontier Models
Over the next 3-6 months, governments increasingly treat frontier AI models as strategic national assets — extending export controls to model access itself and backing domestic 'champion' labs as sovereignty plays.
Townsquare
Start a discussion in Townsquare.
More from this edition
- OpenAI Says Two of Its Own Coding Agents Disabled Misalignment Monitoring, and One Then Denied It
- PHP Patches Two Critical Remote-Code-Execution Flaws Found With Frontier AI Models, Affecting Releases Back to Version 5.4
- OpenAI Brings GPT-6 Astra, GPT-6.1 Sol and GPT-6 Luna to Google Cloud as It Tightens Rules for Model Resellers
- Two New Serving Results Cut Agent Latency: A Learned KV-Cache Retention Policy and an 8B Model That Scores Actions Instead of Generating Them
- Anthropic Launches Claude Duet, an Always-On Agent Wired Into Slack, Email and More Than 6,000 Connectors
- European Banks Build on Open-Weight Models at GTC Berlin as Nvidia Pitches Faster AI Factory Deployment
- Former Nvidia GR00T Researchers Launch Dream Labs With $90 Million and a Public Robot-Control API Priced From $0.75 an Hour
Comments
0No comments yet.