Tensormesh Secures $20M, Launches Serverless AI Inference Platform

Tensormesh, a company focused on optimizing AI inference, has successfully closed a new funding round, securing $20 million from prominent investors such as AMD Ventures, CoreWeave, and NVentures, NVIDIA's venture capital arm. This latest investment extends its seed round, pushing the total funding to $24.5 million. The capital infusion coincides with the general availability launch of Tensormesh Inference, the company's flagship SaaS platform. The Tensormesh Inference platform addresses a critical and expensive problem in AI: the redundant computation of previously processed data during inference. Traditional GPU workflows often recompute the same inputs from scratch for every inference request, leading to wasted GPU cycles and increased operational costs. Tensormesh tackles this inefficiency by employing KV caching, a method that stores and reuses computed results. This approach can lead to up to a 10x reduction in latency and GPU expenditure. Among its deployment options, Tensormesh offers a serverless inference solution. This provides developers with immediate API access to a curated catalog of frontier models, eliminating the need for provisioning or managing underlying infrastructure. The serverless API is designed to be fully compatible with OpenAI's standards, allowing for straightforward integration into existing tools and workflows. This means teams can quickly move from signup to their first inference request within minutes, making AI development more accessible and cost-efficient. For enterprises with larger-scale AI operations, Tensormesh also provides reserved deployments for dedicated capacity and custom SLAs.