Prime Intellect has raised $130 million from leading technology investors, including Nvidia’s NVentures, Intel Capital, Dell Technologies Capital, and Cloudflare CEO Matthew Prince. This financing round values the company at $1 billion, underscoring growing demand for specialized AI training platforms that accelerate model customization and deployment.

The startup’s cloud-based platform focuses on optimizing AI model training by leveraging open-source toolkits — Verifiers and Prime-RL — which it released last year. These tools enable developers to train AI in simulated virtual environments replicating real-world tasks, such as web browsing, code repository interaction, or customer support ticket processing. By reducing setup time and providing a library of ready-made AI environments, Verifiers streamlines the model training workflow.

Prime Intellect’s infrastructure harnesses powerful managed cloud clusters of graphics processing units (GPUs) optimized via Prime-RL, an open-source solution designed for parallelizing large-scale AI training across thousands of GPUs. This system utilizes FSDP2, a PyTorch feature known for its memory efficiency and flexibility in distributing training processes.

The platform supports multiple training methods, including full parameter fine-tuning and LoRA, a lighter-weight approach that enhances models with a small number of specialized artificial neurons tuned for specific tasks. Following training, developers can deploy AI models on inference infrastructure managed by Prime Intellect, accessing GPU clusters optimized specifically for LoRA-enhanced models through a marketplace spanning more than 50 data centers.

Demonstrating its capabilities, Prime Intellect’s platform trained Intellect-3, an AI model with over 100 billion parameters, which reportedly outperformed peers of similar size on various reasoning tests. Clients such as Ramp Inc., a corporate credit card provider, have already employed the platform to create AI models that automate complex tasks like spreadsheet searches, indicating growing commercial adoption.