Lepton AI
Cloud-native AI inference platform built by Caffe creator Yangqing Jia. Acquired by NVIDIA in May 2025 to power the inference cloud strategy.
Acquired Lepton AI
Acquired · May 2025
Lepton AI was a cloud-native AI inference platform founded in 2023 by Yangqing Jia, the creator of Caffe and former leader of Alibaba's PAI ML platform. The company built high-performance serving infrastructure for LLMs and multi-modal models, competing with Together AI and Fireworks. NVIDIA acquired Lepton in May 2025, joining OctoAI as another inference-platform consolidation under the chip vendor. Lepton's tech and team now power NVIDIA's enterprise inference cloud strategy alongside NIM. The standalone product is being sunset as customers migrate to NVIDIA's broader offerings.
Acquired by NVIDIA.
What to use instead
OpenRouter
Unified API and marketplace for the best LLMs at the best prices for any prompt.
Freemium
4.84
360
Fireworks AI
High-speed, cost-efficient generative AI for product innovation with advanced fine-tuning capabilities.
Paid - Inquire
4.92
420
RunPod
Globally distributed GPU cloud for AI tasks.
Paid - Inquire
4.91
410
Overview
Lepton AI: AI Inference Platform (Acquired by NVIDIA)
Lepton AI was Cloud-native AI inference platform built by Yangqing Jia (Caffe creator, ex-Alibaba PAI). Lepton AI was acquired by NVIDIA in May 2025, joining OctoAI (acquired Sept 2024) as another inference platform folded into NVIDIA's enterprise stack. Yangqing Jia and the team brought deep expertise from Caffe / PyTorch / Alibaba PAI to NVIDIA's inference cloud roadmap. The standalone Lepton product is no longer accessible — customers were migrated to NVIDIA NIM and related services.
Key Features
- Cloud-native AI inference platform built by Yangqing Jia (Caffe creator, ex-Alibaba PAI)
- High-performance LLM and multi-modal model serving
- Acquired by NVIDIA in May 2025
- Now part of NVIDIA's enterprise inference cloud strategy
- Originally backed by CRV and Fusion Fund
- Founded 2023; rapid scaling in the AI inference category alongside Together AI and Fireworks
- Yangqing Jia previously created Caffe (one of the foundational deep-learning frameworks) and contributed to PyTorch
Historical Use Case
Historical reference for the AI inference platform consolidation around chip vendors. Lepton's NVIDIA acquisition was the second major inference acquisition in 2025 after OctoAI.
What Happened to Lepton AI
Lepton AI was acquired by NVIDIA in May 2025, joining OctoAI (acquired Sept 2024) as another inference platform folded into NVIDIA's enterprise stack. Yangqing Jia and the team brought deep expertise from Caffe / PyTorch / Alibaba PAI to NVIDIA's inference cloud roadmap. The standalone Lepton product is no longer accessible — customers were migrated to NVIDIA NIM and related services.
FAQ
Q: Why did NVIDIA acquire Lepton AI? A: NVIDIA continues to consolidate inference platforms — Lepton joined OctoAI as a 2024-2025 inference acquisition under NVIDIA.
Q: What about the Lepton product? A: Lepton's standalone product is being sunset; customers migrated to NVIDIA NIM and inference services.
Q: Founder pedigree? A: Yangqing Jia created Caffe (foundational deep-learning framework), contributed to PyTorch, and led Alibaba PAI before founding Lepton.
Q: Acquisition price? A: Not publicly disclosed.
tl;dr
AI inference platform acquired by NVIDIA May 2025. Built by Caffe creator Yangqing Jia.
Related
Looking for more options? Browse the AI Infrastructure directory or read our best AI infrastructure tools listicle. Lepton AI is also tracked on Crunchbase.
Why Use Lepton AI
