The American technology company NVIDIA is leading efforts to combat climate change through the use of artificial intelligence (AI) and accelerated computing. The company has announced significant advances in energy efficiency and sustainability, coinciding with Climate Week 2024 in New York.
Commitment to renewable energy
NVIDIA has stated that all the electricity powering its global operations will come from renewable sources by the end of January. This significant step aligns with the company’s mission to help its customers and partners leverage AI and accelerated computing for sustainable growth.
Improvements in energy efficiency
Foxconn saves energy with digital twins
Electronics giant Foxconn is using accelerated computing and AI to build a digital twin of a new factory in Mexico. This technology will allow them to train robots and define production processes, estimating an annual energy savings of 30%.
More efficient cloud services
Cloud computing services are also making significant advances in energy efficiency. For example:
- An e-commerce website using NVIDIA AI to connect hundreds of millions of daily shoppers achieved an energy efficiency improvement of nearly 12 times by migrating from CPUs to GPUs.
- A popular video conferencing app increased its performance from 3 to 200 queries per second, achieving a 25-fold energy efficiency improvement.
Advances in the healthcare sector
The Wellcome Sanger Institute, one of the world’s largest genomic sequencing facilities, uses NVIDIA Parabricks software on NVIDIA DGX systems to reduce runtime, capital costs, and energy consumption. This translates to a potential savings of $1 million and 1,000 megawatt-hours per year.
Expanding sustainability ecosystem
NVIDIA is fostering a broad ecosystem focused on sustainability:
- A collaboration with the United Nations has trained over 14,000 data scientists in designing AI models for early flood detection.
- The NVIDIA Inception program now includes over a thousand startups applying AI, accelerated computing, and open platforms like NVIDIA Earth-2 to accelerate climate and weather predictions.
Efficiency in AI inference
Over the past eight years, the NVIDIA platform has become 45,000 times more energy-efficient in processing large language models. This efficiency improvement is equivalent to a car being able to travel 280,000 miles per gallon of fuel.
Future outlook
Jensen Huang, founder and CEO of NVIDIA, has stated: “The generative AI revolution is poised to impact all industries and enable a new era of productivity and sustainability by unlocking efficiencies and resource optimization across all sectors.”
Accelerated computing, often referred to as sustainable computing, is shaping up to be the key to meeting the growing demand for computational power in a sustainable and cost-effective manner.
Source: Nvidia