Firmus Technologies has announced a strategic compute partnership with NVIDIA to build a massive AI computing campus, according to a report from telecomtv.com.
The centerpiece is what the companies call an "AI factory" — a data-center campus purpose-built to run artificial intelligence workloads. According to telecomtv.com, the project is anchored by a dedicated 360 MW NVIDIA DSX AI Factory and is designed to house roughly 170,000 GPUs.
GPUs, or graphics processing units, are the specialized chips that power modern AI systems, from chatbots to image generators. Assembling them at this scale is what makes it possible to train and run the largest AI models. The "360 MW" figure refers to electrical capacity — a rough measure of how much power, and therefore how much computing, the site is built to handle.
telecomtv.com reports that the partnership runs through 2034, signaling a long-term commitment rather than a one-off build. The campus is aimed at what the companies describe as global "AI-natives" — organizations whose products and businesses are built around AI from the ground up.
The two sources available for this story both point to the same telecomtv.com report, and specific details beyond the headline figures — such as the campus location, total cost, and construction timeline — were not included.
Why it matters: The scramble to secure enough NVIDIA chips and electrical power has become the defining bottleneck of the AI boom, and a 170,000-GPU commitment stretching to 2034 shows just how much infrastructure companies believe the next decade of AI will demand.