Revolutionising AI data centre cooling with next-generation cryogenic power solutions
Enabling scalable digital infrastructure by reducing energy loss and heat at source
Zap Power is tackling one of the most critical problems facing advanced AI infrastructure: the growing energy and cooling demands of AI and data-intensive workloads.
As AI systems and high-performance computing scale, conventional power delivery approaches are no longer sufficient. Operating costs are increasing, alongside a rapidly expanding environmental footprint. Energy-intensive cooling, high electricity demand and growing water use are placing increasing strain on grids and emissions targets.
Zap Power’s technology introduces a fundamentally different way of managing power in high-density AI systems, reducing energy losses at source and easing the cooling burden placed on modern data centres.
Our solution delivers up to 75% greater energy efficiency than conventional power systems, addressing one of the most urgent constraints in the global digital economy.
A critical power challenge for the next phase of AI growth
The rapid growth of AI workloads is pushing data centres beyond the limits of conventional power and cooling systems.
As rack-level power density rises, the inefficiencies of traditional AC architectures become increasingly pronounced. Energy losses compound across the power chain, generating heat that must be actively removed, with cooling already accounting for up to 50% of total electricity consumption in many facilities.
Energy efficiency is no longer a secondary optimisation but a limiting factor in the deployment and scaling of AI infrastructure.
Against this backdrop, the historical debate between AC and DC power distribution is re-emerging. Modern compute hardware operates natively on DC power, yet most facilities continue to distribute AC, requiring repeated conversion stages that introduce loss, heat and complexity. As AI workloads intensify, these inefficiencies are becoming increasingly costly, driving growing alignment across the AI ecosystem around DC power distribution as a more efficient foundation for future data centres.
Cryogenically cooled power architecture designed for the AI era
Most efforts to improve data-centre efficiency focus on managing heat after it has already been generated. Zap Power takes a fundamentally different approach by addressing energy loss at its origin.
Zap Power’s system operates at ultra-low temperatures, below –150°C, enabling far more efficient power performance in the environments demanded by advanced AI infrastructure.
At the core is the world’s first cryogenically cooled DC network. This solution enables power electronics to operate at low temperatures, dramatically reducing network electrical resistance and unlocking much higher efficiency and power density. Compared to conventional systems, it delivers up to 25% greater energy efficiency, fundamentally changing how power is converted and managed in AI factories.
Zap Power is proud to be part of the NVIDIA Inception programme, a global ecosystem of companies shaping the next generation of AI and accelerated computing. The programme’s focus on foundational technologies underscores the importance of efficient power delivery within AI-driven infrastructure.
A foundational technology with global scale
Power conversion sits at the heart of every data centre, every accelerator and every high-performance AI system. As AI and data-intensive workloads scale, efficiency gains at this layer translate into significant system-level impact.
Zap Power addresses a global, multi-billion-pound opportunity within advanced AI infrastructure. The initial focus is AI data centres, where energy and cooling costs are most acute, with clear potential to extend into adjacent sectors, including renewables, fusion and quantum computing, where efficiency under extreme conditions is critical.
Zap Power’s technology has relevance across regions, operators and system architectures. As demand for resource-heavy digital systems continues to grow, the value of reducing energy loss at source compounds over time.
Interested in finding out more?
If you are exploring next-generation power and cooling solutions, we would welcome a conversation.