Touring Capital and Kindred Ventures co-lead $32M round to scale Parasail’s AI Supercloud, an inference and training platform built for AI agents
SAN FRANCISCO, April 15, 2026 /PRNewswire/ — Parasail, a company building an AI Supercloud, an inference and training platform for deploying and scaling AI agents, today announced it has raised $32 million in Series A funding, bringing total funding to $42 million. The funding round was co-led by Touring Capital and Kindred Ventures, with participation from Samsung NEXT, Flume Ventures, Banyan Ventures and existing investors.
The new capital will be used to expand Parasail’s AI Supercloud, a fabric of global compute resources, which automatically optimizes model endpoints for speed, performance, and cost. With the additional funding, Parasail will deepen the orchestration and inference optimization, accelerate go-to-market efforts and to strengthen strategic partnerships across the GPU and data center ecosystem.
The world is currently rebuilding the entire cloud around AI, with trillions of dollars building data centers and filling them with GPUs. Yet, developers are still constrained by access to this infrastructure and the challenges of standing up and running AI models quickly and efficiently for their products. From vertically-focused enterprise agents to new consumer-focused personal agents to broad-based agent SDK platforms, Parasail powers inference and reinforcement learning environments for the surging wave of AI agents upending the legacy application paradigm.
Parasail is delivering the AI Supercloud to satisfy the insatiable need for customized, instant, and dependable inference and continuous training. Customers can now set up massively scalable AI in less than five minutes, powered by the world’s supply of GPU compute.
AI as a Component for Modern Software
The market for developer-controlled AI is rapidly expanding into what analysts estimate will exceed $100 billion as startups and enterprises move toward aggressive performance targets, custom models, and distributed compute ecosystems. A vast software, hardware, and data center ecosystem is forming to support that shift.
Companies are discovering that AI has become a component of modern software that they can deploy themselves, leading to an explosion of open source and specialized models. To capture the potential of AI, software companies need greater control over cost, latency, and customization, so they are moving away from black-box APIs and towards AI endpoints they can operate themselves. But the infrastructure behind that shift remains fragmented: GPU supply is constrained and inconsistent, inference optimization is complex, and scaling often requires contracts, sales negotiations, and months of integration work.
Parasail removes that friction and delivers a seamless customer experience for AI agent deployment with production-ready AI endpoints in minutes, scaling to massive traffic without contracts or infrastructure management. With five lines of code, developers can launch custom models, hit aggressive targets on latency, throughput, and tokens per second, and scale effortlessly through large surges in traffic when products go viral.
“AI builders shouldn’t have to become infrastructure experts to ship great products,” said Mike Henry, founder and CEO of Parasail. “AI is becoming the core infrastructure for modern software. But the infrastructure layer itself hasn’t kept up. We built Parasail so teams can deploy custom AI at massive scale without negotiating contracts, managing fragmented GPU supply, or hiring performance engineering teams.”
A Structural Layer for AI Inference
Parasail is not another single-cloud provider. It operates a programmable deployment network that abstracts away supply fragmentation and inference optimization.
Its differentiation is structural and this model allows startups and growth-stage companies to scale from zero to enterprise-grade workloads without rewriting infrastructure as demand increases:
- Developer-first deployment: Production AI endpoints live in minutes with minimal code, abstracting significant backend orchestration complexity.
- Economics as a first principle: Rather than competing on isolated speed benchmarks, Parasail optimizes workloads for cost efficiency at scale.
- Aggregated hardware supply: The AI Supercloud connects Parasail’s internal GPU fleet with a diverse set of compute providers, unlocking elastic capacity beyond any single cloud’s availability and bringing new hardware platforms with disruptive performance to market quickly.
- Automated performance optimization: Parasail replaces manual kernel tuning and performance engineering with automated systems that continuously optimize inference across the network.
“AI infrastructure is moving beyond single-cloud models,” said Samir Kumar, General Partner at Touring Capital. “As inference workloads scale, companies need flexibility across hardware, geography, and cost structures. Parasail has built the control layer that makes that possible. The team combines deep systems expertise with a clear product vision, and we believe they are well positioned to define how modern AI applications are deployed.”
“The main product construct of this AI wave is the agent – replacing the notion of the manually-operated application world of the last thirty years,” said Steve Jang, Managing Partner at Kindred Ventures. “These agents are directed but can operate autonomously, call multiple models at runtime, and will require massive amounts of tokens. This new world and its developers need powerful customized inference and reinforcement learning capability that are flexible, instant, and dependable. Parasail offers the first agent-focused inference and training solution which simplifies the model and compute complexity of today’s dynamic generative AI market.”
Since launching in April 2025, Parasail processes over 500 billion tokens per day and customers include Elicit, mem0, Gravity, Kotoba, and Venice with 30% MoM revenue growth, leading the pack of the second-wave inference providers.
About Parasail:
Parasail is delivering the AI Supercloud—a programmable deployment network that gives AI builders production-ready endpoints in minutes, scaling to massive traffic without contracts, infrastructure management, or vendor lock-in. By aggregating global GPU supply and automating inference optimization, Parasail enables companies to deploy custom AI models with aggressive performance targets on latency, throughput, and cost. For more, visit: https://www.parasail.io/
SOURCE Parasail/BAM Communications