QumulusAI (NASDAQ:QMLS) has announced the purchase of 1,632 Nvidia Blackwell B300 GPUs as part of a significant expansion of its AI cloud infrastructure, marking one of the company’s largest capacity investments to date.
The order includes 204 Nvidia HGX B300 systems equipped with the latest Blackwell B300 GPUs, alongside an additional 192 Nvidia RTX PRO 6000 Blackwell GPUs that will be used for AI training, inference and visualisation workloads.
Expansion supports growing AI cloud demand
QumulusAI said the investment is designed to support more than $124 million in three-year AI inference agreements secured earlier this year.
The acquisition was financed primarily through funding arrangements with Technology Finance Corporation and USD.ai. The company sources its computing infrastructure from several original equipment manufacturers, including Supermicro and Lenovo, before deploying systems across its network of colocation facilities and company-owned data centres.
“Meeting AI demand today is a test of access, speed and flexibility, and you can’t win on one without the other two,” said Mike Maniscalco, CEO of QumulusAI.
GPU fleet continues rapid expansion
According to the company, its deployed GPU fleet has grown by more than 450% between June 2025 and June 2026 as demand for AI computing infrastructure continues to accelerate.
The latest investment further expands QumulusAI’s capacity to deliver cloud-based AI services and positions the company to support increasing enterprise demand for high-performance AI inference and training workloads.