Data Infrastructure Decarbonisation in AI Era

September 18, 2024

As the fourth industrial revolution powered by the rapid advances in AI, the world faces the most extraordinary techno-capital acceleration.

AI investment is forecast to approach $200BN globally by 2025 and with that, if executed with discipline could power a near $1TRN opportunity for renewable energy.

The problem

The Traditional  Data Centre Model is not fit for purpose to serve the demands of AI

AI's computational demands to train and power the models already account for 2% of the global datacentre footprint, projected to rise to 10% by 2025, according to The Uptime Institute. This is forcing rapid expansion.

The impact of AI on data centres is profound, with rapid adoption and expanding user bases for platforms like ChatGPT. This growth places enormous strain on datacentre infrastructure, necessitating advanced solutions to manage the increased power demand effectively. Continuous advancements in computational power are essential to support the world's growing reliance on sophisticated digital applications

In addition, AI's energy consumption is expected to reach 85 to 134 terawatt hours annually, about 0.5% of global electricity consumption.

The solution

Decentralised Compute is the Key to the Decarbonisation of Data Infrastructure

Traditional ‘centralized’ datacentres have been criticized for their excessive power consumption, and impact on environment, and lack of community contribution.

Compute clusters are modular supercomputers that handle high-performance tasks such as Artificial Intelligence (AI), High Performance Computing (HPC) and Bitcoin Mining.

These clusters rely on renewable energy sources, lowering their carbon footprint and reducing their grid dependency. These clusters have clear advantages such as:

  1. Use blended energy sources and harvest ‘latent’ green energy.
  2. Agile, mobile and distributed geographically increasing resilience and scalability.
  3. Use eco-friendly materials and practices in construction and maintenance.
  4. Can be easily integrated into solutions for utilising waste “heat’ to be used by homes or industry.

Compute Clusters are container-based portable compute ‘centres’. They are configurable based on size, site requirements, energy supply and onward off-take agreements. The diagram below outlines the model.

The potential impact and realised benefits

Compute Clusters can have both a social and environmental impact within as little as six months. We have looked at both the environmental and socio-economic impacts based on megawatt capacity.

A close-up of a graphDescription automatically generated

Why Redsand Ventures

RSV offers a complete lifecycle solution to the future of greening the future of compute in the AI age. We are both Investor/Funders as well as Cluster Operators. RSV builds and invests through its regulated alternative investment fund (RAIF) – called RSV Data Fund I.

We are currently focused on projects in the UK, Europe and the Middle East.

For further information contact:

Nicole Anderson ([email protected])

Sandeep Dama ([email protected])