Parasail Raises $32M Series A to Scale Pay-Per-Token AI Inference Cloud Processing 500B Tokens Daily

Parasail Secures $32 Million Series A Funding for AI Inference Expansion

Parasail, an AI infrastructure startup, has closed a $32 million Series A funding round co-led by Touring Capital and Kindred Ventures, with participation from Samsung Electronics' startup investment arm and other investors 1

1

2

2

. The startup funding will fuel the expansion of Parasail's cloud computing service, which currently generates 500 billion tokens per day for developers running AI models 1

1

. Mike Henry, CEO and founder of Parasail, previously built the cloud offering at Groq, the LLM-focused chipmaker, where he recognized that developers building software on AI models needed specialized cloud processing tailored to their needs 1

1

.

Pay-Per-Token Model Eliminates Long-Term GPU Commitments

The AI inference cloud platform, branded as AI Supercloud, addresses a critical pain point in the market by offering GPU capacity on a pay-per-token model without requiring long-term contracts 2

2

. Traditional cloud providers often require companies to sign extended procurement agreements for renting GPUs, which proves impractical for startups with limited resources and enterprises running small-scale AI pilot projects 2

2

. Parasail operates across 40 data centers in more than 15 countries, renting processing time and purchasing additional capacity from liquidity markets to orchestrate cost-efficient AI inference behind the scenes 1

1

2

2

.

Infrastructure Designed for Deploying Custom AI at Scale

Parasail simplifies the deployment of AI workloads by enabling developers to launch models with as few as five lines of code 2

2

. The platform automates administrative tasks such as kernel configuration, streamlining workload management for developers 2

2

. Customers can access GPUs through multiple hosting options, including two serverless configurations that automate cluster management and dedicated endpoints that offer customizable performance 2

2

. The company currently offers Nvidia's H200 as its most advanced graphics card, though newer GPU generations have since been released 2

2

.

Open-Source Models Drive Demand for Cost-Efficient Inference

Parasail's growth strategy relies on the continued proliferation of open-source models and agents outside frontier labs, driven by the increasing cost and friction of using offerings from companies like Anthropic and OpenAI 1

1

. Andreas Stuhlmüller, CEO of Elicit, a startup that raised a $22 million Series A to develop a research assistant for scientific literature, explained that his company has moved toward open models because "it's pretty rough sending 100,000s of requests to an API endpoint" 1

1

. Elicit's customers at top pharmaceutical companies use the LLM-based tool to review and analyze data from tens of thousands of scientific papers, employing a hybrid architecture where open models handle initial screening before frontier AI models provide final answers 1

1

.

Inference Demand Outstrips Supply as AI Agents Proliferate

The proliferation of model queries, as agents become an increasingly common part of software development, is driving investment in companies like Parasail that provide infrastructure for cheap inference 1

1

. Samir Kumar, a partner at Touring Capital, expects inference to account for at least 20% of the cost of building software in the future 1

1

. Steve Jang, a partner at Kindred Ventures, emphasized the market opportunity: "Everyone thought there was an AI bubble. There's no AI bubble. Inference demand is far outstripping supply" 1

1

. Henry argues that Parasail's focus on inference—with no training allowed—and willingness to serve startup customers without long-term commitments distinguishes the company from larger cloud-computing firms focused on enterprise business and better-funded competitors like Fireworks AI and Baseten 1

1

. "AI is becoming the core infrastructure for modern software. But the infrastructure layer itself hasn't kept up," Henry said 2

2

.