Artificial intelligence infrastructure startup Parasail Inc. announced it has raised $32 million in early-stage funding. The Series A round was jointly led by Touring Capital and Kindred Ventures, with participation from Samsung Electronics’ startup investment arm.
Parasail operates an inference-optimized cloud platform, the AI Supercloud, allowing customers to purchase GPU capacity on a pay-per-token basis without long-term contracts. This model addresses the impracticality of lengthy procurement agreements for startups and small-scale AI projects.
The platform supports advanced GPUs, including Nvidia's H200, with capacity hosted in internally-operated clusters and by partners across 40 data centers in over 15 countries.
Parasail streamlines AI inference by enabling developers to deploy workloads with minimal code and automating administrative tasks like kernel configuration. The platform offers serverless hosting options and dedicated endpoints for tailored performance.
Dedicated endpoints allow customization of GPU capacity scaling and support neural network quantization to reduce inference costs. A batch processing service is also available for cost-efficient, large-volume data processing.
"We built Parasail so teams can deploy custom AI at massive scale without negotiating contracts, managing fragmented GPU supply, or hiring performance engineering teams," said CEO Mike Henry. The company will use the new capital to enhance platform features, strengthen its partner ecosystem, and invest in go-to-market strategies.