← Back to live feed1 story
Several Chinese customers have already placed orders for a new AI inference chip that Nvidia will deliver in small batches by the end of 2026. The hardware, a China-specific variant of the company's Language Processing Unit (LPU), uses Groq-licensed technology to comply with U.S. export rules without reducing computing power. This LPU is designed to work alongside Nvidia's GPUs to accelerate AI chatbot processing.
Nvidia modified the chip's software to function with processors available in China as an alternative to its restricted Rubin GPUs. While the product is cleared under U.S. regulations, the company faces a remaining hurdle in Beijing, which could still restrict the purchase of the chips.
Sign in to suggest edits
Key sources
- SOURCE@financialjuice“small-batch shipments of an AI chip specially tailored for Chinese customers by the end of the year”x.com
- SUPPORT@wallstengine“Nvidia modified the software so it can work with processors available in China instead of its restricted Rubin GPUs.”x.com
- SUPPORT@stockmktnewz“a new chip that works alongside Nvidia's GPUs to speed up AI chatbot processing”x.com
- SUPPORT@zerohedge“UNCLEAR IF BEIJING WILL ALLOW ORDERS TO GO THROUGH: INFORMATION”x.com
- SUPPORT@techmeme“small-batch shipments of an LPU tailored for Chinese customers by the end of 2026”x.com