Accelerating AI for everyone. 加速到来你的AI世界
Approaching.ai is a leading AI inference optimization company, dedicated to making enterprise AI deployment accessible, affordable, and efficient.
Founded in late 2023, our team consists of experts from Tsinghua University, with Prof. Weimin Zheng (Academician of Chinese Academy of Engineering) as our chief advisor. We pioneered the "Storage-Computation Trade-off" (以存换算) technology and Heterogeneous Cooperative Inference Architecture, reducing LLM deployment costs by 10x.
- 🔄 Storage-Computation Trade-off: Reuse historical KV cache to reduce computation by 90%
- ⚡ Heterogeneous Cooperative Inference: Fully leverage CPU, GPU, NPU, and storage
- 🖥️ Single-GPU Inference: Run 100B+ models on consumer-grade GPUs
We're proud to be the co-founders of KTransformers together with the KVCache.AI team from Tsinghua University.
🔥 16.6k+ stars on GitHub | Industry-leading performance | Featured on GitHub Trending worldwide
| Product | Description |
|---|---|
| KLLM | Proprietary inference engine with storage-computation trade-off |
| AI Inference Appliance | Hardware-software integrated solution for enterprise deployment |
| AMaaS | Model operations platform with graphical UI and one-click deployment |
- 📅 2023 Q4 - Company founded, seed round from ZhenCapital
- 📅 2024 Jul - Launched KTransformers open-source project
- 📅 2024 Sep - Released enterprise inference appliance
- 📅 2025 Feb - KTransformers reached 10k+ stars, #1 on GitHub trending
- 📅 2025 Mar - Partnership with Huawei Ascend & Kunpeng, Lenovo AI workstation
- 🌐 Website: https://approaching-ai.com
- 📧 Email: quxin@approaching.ai
- 📍 Address: 12A Floor, Xueqing Jia Chuang, Haidian District, Beijing
Building the future of affordable AI inference 🚀