NVIDIA: AI for Data Centre Energy Efficiency Improvements
With the AI market experiencing explosive growth, companies worldwide are exploring and implementing AI solutions in virtually every sector.
However, this rapid expansion raises important questions about the technology's energy footprint, particularly in data centres.
Data centres serve as the backbone of modern digital infrastructure, housing processes including data storage, cloud services, media streaming and AI operations.
Jensen Huang, CEO and Co-Founder at NVIDIA, explains: "Data centres are already about 1-2% of global electricity consumption and that consumption is expected to continue to grow.
“This continued growth is not sustainable, neither for operating budgets nor for our planet. Accelerated computing is now the most sustainable way to advance computing."
Data centres facilities typically operate 24/7 to meet the constant demand for their services, often containing tens of thousands of devices that require uninterrupted power and cooling.
Consequently, data centres have become significant consumers of electricity on a global scale.
NVIDIA's aims for sustainable AI
NVIDIA is a leading chip-maker and supplier of AI hardware and software.
The company, which invented graphics processing units (GPUs) in the 1990s, has grown to become one of the largest in the USA.
It became the third company in US history to close with a market cap exceeding US$2tn in March 2024.
NVIDIA aims to use AI to increase energy efficiency and support other sustainable solutions across industries.
Joshua Parker, Senior Director of Corporate Sustainability at NVIDIA, says: "AI, I firmly believe, is going to be the best tool that we've ever seen to help us achieve more sustainability and more sustainable outcomes."
How AI can improve energy efficiency
NVIDIA's approach to improving data centre sustainability centres on the concept of accelerated computing platforms.
These systems combine GPUs and CPUs to handle complex computations quickly and efficiently.
According to NVIDIA, these platforms can be up to 20 times more energy efficient than traditional CPU-only systems for AI inference and training.
Joshua says: “The change in efficiency is really, really dramatic.
"If you compare the energy efficiency for AI inference from eight years ago until today, it's 45,000 times more energy efficient.”
The impact of cooling on data centres
NVIDIA is also focusing on innovative cooling solutions for data centres.
The computers used require near-constant cooling to avoid gaps in operations.
The company is advocating for direct-to-chip liquid cooling, which offers significant advantages over traditional air conditioning systems in terms of energy efficiency.
"Our recommended design for the data centres for our new B200 chip is focused all on direct-to-chip liquid cooling," Joshua explains.
This approach allows for more targeted and efficient cooling, potentially reducing the overall energy consumption of data centres.
AI's broader impact on climate change
AI has the potential to address climate change on multiple fronts.
Enhanced weather forecasting powered by AI is becoming more accurate, allowing for better preparation and response to climate-related weather events.
Digital twin technology, enabled by AI, is also making a significant impact.
These virtual replicas allow companies to simulate and optimise energy consumption without making physical changes, leading to more efficient and sustainable operations.
AI is also supporting the development of new materials for renewable energy technologies, such as electric vehicles and solar panels, accelerating the transition to cleaner energy sources.
Joshua says: "AI and accelerated computing in general are game-changers when it comes to weather and climate modelling and simulation."
Make sure you check out the latest edition of Energy Digital Magazine and also sign up to our global conference series - Sustainability LIVE 2024
Energy Digital is a BizClik brand.