Kyndryl’s Boulder Site Welcomes Argentum AI as Flagship Anchor Tenant

Unite.AI
Read full story· 3 min
Kyndryl and Argentum AI have formed a 20-year, $630M partnership to deploy NVIDIA B300 GPU systems for AI inference at Kyndryl's Boulder data center. Argentum will operate GPU clusters providing high-performance inference capacity to AI customers, leveraging Kyndryl's Tier III infrastructure.

What happened

On November 17, 2026, Argentum AI and Kyndryl announced a 20-year partnership worth $630 million to deploy NVIDIA B300 GPU systems at Kyndryl’s Boulder, Colorado data center. Argentum AI will become the main tenant at the facility, operating GPU clusters for AI inference workloads. The deployment is already in progress, using Kyndryl’s Tier III infrastructure, which is built to support high-power GPU systems. The agreement gives Argentum access to 10 megawatts of data center capacity to serve its AI customers.

Key details

  • 20-year, $630 million agreement between Argentum AI and Kyndryl
  • Deployment uses NVIDIA B300 GPU systems for AI inference
  • Initial deployment covers 10 megawatts of capacity at Boulder, Colorado site
  • Kyndryl’s Boulder facility is Tier III certified and supports high-density hardware
  • Argentum offers customers options from Hardware-as-a-Service to Platform-as-a-Service

Why it matters

AI companies and enterprises seeking high-performance inference capacity now have access to dedicated GPU resources at a major U.S. data center. Argentum’s customers can scale their AI workloads without waiting for new construction or separate financing deals, reducing deployment time and complexity. Kyndryl benefits by securing a long-term anchor tenant and putting its existing infrastructure to use for AI workloads.


More in Industry & Manufacturing