Highlights:

  • Many enterprises developing their own AI models rely on cloud-based VS Code and Jupyter Notebook environments—tools commonly used for coding and neural network development.
  • TrueFoundry revealed that its platform is now used by over 30 customers, including Nvidia Corp., which leverages the software to optimize GPU usage for its LLM workloads.

Startup TrueFoundry Inc. secured USD 19 million in Series A funding round. The company specializes in AI workload management.

The round was led by Intel Capital, with participation from Eniac Ventures, Surge, and Jump Capital. Several angel investors also joined, including former Cohesity Inc. CEO Mohit Aron.

Building the infrastructure required to run AI models in production can be a time-intensive process, often taking months. TrueFoundry’s platform aims to simplify this by offering prebuilt features for deploying and maintaining large language models (LLMs), reducing the need for developers to build everything from scratch.

Many enterprises developing their own AI models rely on cloud-based VS Code and Jupyter Notebook environments—tools commonly used for coding and neural network development. To optimize costs, TrueFoundry automatically shuts down these environments when they’re no longer needed.

The platform also streamlines LLM management. TrueFoundry enables developers to convert a newly created LLM into an API with a single click, making integration with other workloads seamless. Behind the scenes, it dynamically adjusts infrastructure resources based on demand.

To further cut costs, the platform utilizes spot instances—cloud-based virtual machines that run on unused server capacity. While spot instances are significantly cheaper than standard virtual machines, they come with technical challenges, which TrueFoundry helps manage.

Additionally, TrueFoundry provides a library of over 30 open-source LLMs, along with a fine-tuning tool that allows customers to optimize these models using custom training data. It leverages a technique called QLoRA, which slightly reduces training speed but significantly lowers memory usage.

Once an LLM is deployed, the platform continuously monitors its performance, tracking key metrics such as inference cost, latency, and error rates to detect and resolve technical issues efficiently.

“Enterprises using TrueFoundry have built and launched their internal AI platforms in as little as two months, achieving ROI within four months – a stark contrast to the industry average of 14 months,” said TrueFoundry CEO Nikunj Bajaj.

TrueFoundry revealed that its platform is now used by over 30 customers, including Nvidia Corp., which leverages the software to optimize GPU usage for its LLM workloads. Over the past year, customers have deployed more than 1,000 AI clusters using TrueFoundry’s technology.

The company plans to use its Series A funding to develop new features and expand its go-to-market strategy.