The Carbice and DarkNX agreement positions Carbice as a system-level solutions partner for global AI data center projects. Carbice will contribute advanced interface materials, thermal expertise, and long-term performance validation. At the same time, the partnership aims to strengthen AI data center performance and sustainability as both companies support next-generation, high-density infrastructure.
As artificial intelligence workloads intensify, operators face growing reliability challenges. Therefore, the companies are targeting critical thermal, mechanical, and electrical interfaces that directly affect uptime and hardware lifespan. Through this collaboration, engineers will improve performance at both the chip and system levels.
Supporting High-Density AI Infrastructure
Under the agreement, Carbice will assist DarkNX and its customers in designing and deploying AI data centers. The Carbice Lab will provide engineering insights and performance data. The lab specializes in thermal interface reliability and real-world validation testing.
In addition, Carbice will work with other technology providers on DarkNX’s 300-megawatt data center campus in Ontario, Canada. DarkNX designed the facility to support high-density AI workloads, where cooling efficiency and interface stability remain essential.
By addressing interface degradation early in the design phase, engineers can reduce throttling and downtime. Consequently, hardware can maintain stable performance over longer operating cycles.
Advancing Sustainability Goals
Sustainability remains central to the Carbice DarkNX partnership. Carbice states that its cooling technologies can prevent nearly 30 gigawatt-hours of unnecessary electricity use over five years in a single 300MW campus. That reduction equals roughly 10,500 metric tons of avoided CO₂ emissions.
Moreover, Carbice designs its materials to limit lifecycle inefficiencies. By preventing degradation-related losses, operators can lower operating costs and reduce carbon intensity. Over time, companies could save millions of dollars each year in energy and reliability expenses.
Science-Backed Interface Engineering
Baratunde Cola, CEO and founder of Carbice, said interface reliability increasingly defines AI infrastructure performance. He stressed that teams must evaluate how systems perform not only at launch but also years later.
Carbice manufactures its interface solutions in the United States and engineers them to resist degradation. The materials maintain elasticity and stability, which strengthens long-term durability.
DarkNX will integrate Carbice’s Dow-Carbice SW and SA Series Pads across CPUs, GPUs, and power systems. OEMs have already qualified these technologies for multiple applications. Additionally, engineering teams will use the Carbice Lab to test solutions against datasets that replicate real-world operating environments.
Strengthening Full-Stack Reliability
The collaboration also extends beyond compute hardware. Carbice addresses torque loss and mechanical inconsistencies in switchgear and busbar joints. By resolving these issues, DarkNX can deliver stronger reliability across the full data center stack.
Isaac Islam, CEO of DarkNX, said the company takes a technology-agnostic approach when selecting solutions. He added that Carbice’s materials science and system-level expertise strengthen DarkNX’s ability to deliver durable AI infrastructure at scale.
