In March 2026, Sail Research emerged from stealth mode, securing $80 million in funding from prominent investors Kleiner Perkins and Sequoia. The startup, co-founded by Naganath and Movva, aims to transform the AI landscape by focusing on long-running AI agents, which require a specialized inference platform. Traditional AI infrastructure is designed for quick interactions, such as chatbots, but this approach has led to skyrocketing enterprise AI costs. As token consumption increases, Goldman Sachs predicts a 24-fold rise by 2030, highlighting the urgent need for more efficient systems.
Sail Research's unique platform is engineered to optimize resource allocation, allowing for significant cost improvements for clients, often between 3x to 10x compared to existing alternatives. Movva emphasizes the importance of efficiency, stating that the company prioritizes throughput over latency, which is a challenging balance to achieve. The startup's technology is not suitable for applications like voice assistants or live chatbots, but it excels in scenarios where agents operate for extended periods.
The founders, both with backgrounds at Apple, have a shared vision of maximizing GPU utilization and minimizing waste. Movva's previous experience at Together AI, a leading open-source model inference provider, further informs Sail Research's approach. The company has already ramped up to processing trillions of tokens weekly, showcasing its capability to handle large-scale AI workloads effectively.
As the demand for AI agents continues to grow, Naganath believes that the inference market will expand significantly, potentially becoming 10x to 100x larger than it is today. This belief underscores the strategic positioning of Sail Research within the rapidly evolving AI landscape, as it seeks to provide a purpose-built solution that addresses the inefficiencies of current AI infrastructure.
