NVIDIA DGX Cloud Lepton

NVIDIA DGX Cloud LeptonSince 2023-08

NVIDIA DGX Cloud Lepton (formerly Lepton AI, acquired by NVIDIA) is an AI-centric compute marketplace and model serving platform that connects developers to tens of thousands of GPUs across a global network of NVIDIA Cloud Partners (including CoreWeave, Lambda, and tier-1 clouds). Founded by Yangqing Jia (creator of Caffe) and acquired by NVIDIA, DGX Cloud Lepton functions like a high-performance compute marketplace for AI engineering teams. It allows developers to discover available GPU compute across regions and seamlessly deploy, fine-tune, and scale AI workloads with zero Kubernetes overhead.

DGX Cloud Lepton integrates directly with the full NVIDIA enterprise software stack, including NVIDIA NIM (Inference Microservices), NeMo, and NVIDIA Cloud Functions. Developers use Python Photons and simple CLI commands to turn arbitrary PyTorch scripts into auto-scaling microservices running on NVIDIA H100, H200, and Blackwell B200 clusters. The platform provides heterogeneous multi-cloud abstraction, automatic load balancing, scale-to-zero serverless runtimes, and distributed key-value storage, giving enterprise teams instant access to reserved and spot GPU capacity with guaranteed NVIDIA driver and CUDA acceleration.

Fireworks AISince 2023-09

Fireworks AI is an enterprise AI inference and model serving platform built to run open-weights LLMs, vision models, and multimodal architectures with lightning-fast speeds and lowest cost. Created by former Meta AI and PyTorch infrastructure engineers, Fireworks powers millions of daily AI requests with sub-100ms time-to-first-token (TTFT) and high token throughput. Fireworks allows developers to seamlessly deploy, fine-tune, and serve models like Llama 3.1/3.3, DeepSeek-R1/V3, Mixtral, Qwen 2.5, and Flux.1 with zero cold starts. It uniquely supports instant LoRA fine-tuning switching on shared GPU infrastructure, allowing thousands of custom fine-tuned adapters to run without paying for dedicated hardware.

The Fireworks AI engine utilizes proprietary GPU compilation optimizations, speculative decoding, dynamic kernel fusing, and custom tensor-parallel kernels to maximize memory bandwidth and FLOPS efficiency on NVIDIA H100 and B200 clusters. Fireworks provides a fully OpenAI-compatible REST and streaming API alongside native function calling, JSON schema guarantees, and multimodal image input. Its FireAttention technology drastically cuts KV-cache memory overhead, enabling massive concurrency and context lengths up to 128k tokens while maintaining deterministic latency SLAs.