Editorial illustration for Meituan trains 1.6 trillion-parameter LongCat-2.0 on Chinese chips, no Nvidia
Meituan's LongCat-2.0: 1.6T-Parameter AI on Chinese Chips
The numbers are cartoonishly large. Forget billions. Meituan built LongCat-2.0 with 1.6 trillion parameters, and they did it on over 50,000 homegrown Chinese chips. Their announcement was a single, clipped sentence that changes everything: "We now have the capability to train large-scale models on domestic computing clusters." Nvidia’s grip just loosened.
The message to Washington is hard to miss. Despite US export controls in place since 2022, China appears to have produced its first competitive trillion-parameter model trained entirely on domestic hardware.
That this team didn't exist in 2022 is staggering. Assembled last year, they’ve already processed 35 trillion tokens. On SWE-bench Pro and Multilingual leaderboards, LongCat-2.0 now beats Google's Gemini 3.1 Pro and OpenAI's GPT-5.5.
It doesn't beat Anthropic's Claude Opus. Not yet. But the gap is technical, not existential.
The dependency on American silicon was a global assumption. Meituan just proved it was an assumption, period. Every strategic plan in every capital and boardroom now requires a rewrite.
Related Reading
Business & Startups Nordic pilot adds Gemini for Education, NotebookLM to boost AI literacy
Business & Startups Kling launches Video O1, all-in-one model with MVL bridge using transformer
Business & Startups DeepSeek Seeks More Capital Weeks After USD 7B Funding Round
Related Topic NVIDIA and Google Cloud let developers scale AI from prototype to production
Market Trends
Related Topic NVIDIA NeMo powers telco reasoning model for autonomous network workflows
Industry Applications
From Archives Google launches Nano Banana 2 Lite image model and Gemini Omni Flash video API
From Archives Deloitte tells consultants AI will pressure billable‑hour model, says Manstof
From Archives OpenAI unveils Jalapeño custom inference chip, challenging Nvidia's AI dominance
Research & Benchmarks
From Archives Prerequisites for NVIDIA AI‑Q Blueprint on OCI: Cluster and Volume Limits
Research & Benchmarks
Common Questions Answered
What is the significance of Meituan training LongCat-2.0 on Chinese chips instead of Nvidia?
Meituan's training of the 1.6 trillion-parameter LongCat-2.0 model on Chinese chips marks a major milestone in reducing dependence on Nvidia hardware. This achievement demonstrates that domestic Chinese AI chips can handle large-scale model training, challenging Nvidia's dominance in the AI chip market.
How many parameters does Meituan's LongCat-2.0 model have?
Meituan's LongCat-2.0 model has 1.6 trillion parameters, making it one of the largest AI models trained entirely on Chinese chips. This scale is comparable to leading global models like GPT-4, but achieved without using Nvidia hardware.
What is the significance of Meituan training LongCat-2.0 on Chinese chips instead of Nvidia?
Training LongCat-2.0 on Chinese chips demonstrates that domestic alternatives can handle massive AI workloads, reducing reliance on Nvidia amid export restrictions. This achievement highlights China's progress in developing self-sufficient AI infrastructure and chip technology.
Further Reading
- China's Meituan says its new AI model was trained on domestic chips — The Next Web
- Meituan claims China's biggest AI model trained on local chips — South China Morning Post
- Meituan open sources LongCat-2.0, the 1.6T, near-frontier agentic coding model that's been leading OpenRouter trained entirely on Chinese chips — VentureBeat
- China's Meituan says new AI model trained on domestic chips — Reuters
- LongCat-2.0, a large-scale MoE model with 1.6T total and 48B Active — Hacker News
