All the electricity that powers NVIDIA’s global operations will come from renewable sources by the end of January.
It’s the right fuel for the company’s mission: to help customers and partners harness AI and accelerated computing for sustainable growth.
This week, world leaders gather in New York City to mark Climate Week 2024. It’s a good time to take stock of recent gains in energy efficiency in fields from manufacturing and cloud computing to healthcare and beyond.
Foxconn Saves Energy With Digital Twins
Foxconn, the world’s largest electronics manufacturer, is using accelerated computing and AI to build a digital twin of a new factory in Mexico, where it will train its robots and define production processes.
“Our digital twin will guide us to new levels of automation and industrial efficiency, saving time, cost and energy,” said Young Liu, chairman of Foxconn, which estimates a 30% annual energy savings.
It’s the latest company to employ software from the Siemens Xcelerator portfolio including Teamcenter and NVIDIA Omniverse, a platform for developing 3D workflows and applications based on OpenUSD.
“We will revolutionize how products and experiences are designed, manufactured and serviced,” said Roland Busch, president and CEO of Siemens AG. “In collaboration with NVIDIA, we will bring accelerated computing, generative AI and Omniverse integration across the Siemens Xcelerator portfolio.”
Siemens and NVIDIA executives will talk about their sustainability efforts in a Sept. 25 Climate Week event moderated by TIME magazine.
Cloud Services Increase Efficiency
Cloud computing services are also racking up advances in energy efficiency.
An e-commerce website uses NVIDIA AI to connect hundreds of millions of shoppers a day to the products they need. After migrating from CPUs to GPUs, it achieved significantly lower latency with a 33x speedup and nearly 12x energy-efficiency improvement.
A popular video conferencing application captions several hundred thousand virtual meetings an hour. By switching from CPUs to GPUs, throughput increased from just three to 200 queries per second — a 66x speedup and 25x energy-efficiency improvement.
To meet sustainability goals, service providers and their customers are increasingly turning to renewable energy sources.
For example, Equinix is the first global data center provider to publish a science-based climate neutral target and 100% renewable energy goal by 2030. Currently, renewables address 96% of the energy it uses, giving customers access to technologies such as NVIDIA DGX systems for sustainable, private AI.
The DGX systems use NVIDIA H100 Tensor Core GPUs, which offer up to 4x more energy efficiency than previous models. Equinix employs advanced liquid cooling to reduce power and water consumption.
The evolution of the data center will be discussed at a Climate Week event hosted by NVIDIA and Crusoe Energy Systems, a company that builds and operates clean computing infrastructure.
Decoding Cancer While Saving Energy
Healthcare companies are adopting AI and accelerated computing to help them understand and treat diseases.
For instance, the Wellcome Sanger Institute runs one of the largest genome sequencing facilities in the world. Its cancer program sequences and analyzes tens of thousands of cancer samples.
Using NVIDIA Parabricks software on NVIDIA DGX systems reduces its run times by 1.6x, capital costs by 24x and energy consumption by up to 42x. That amounts to potential savings of $1 million and 1,000 megawatt-hours a year.
The latest version of Parabricks, released in mid-September, can further enhance such work.
“The Sanger Institute handles hundreds of thousands of samples annually,” said Jingwei Wang, a principal software developer at Wellcome Sanger in a profile describing its work. “With NVIDIA GPUs and Parabricks accelerated genomic analysis software, we will save considerable time, cost and energy analyzing them.”
Expanding the Sustainability Ecosystem
In addition to enabling efficiency gains for customers, NVIDIA is nurturing a broad ecosystem focused on sustainability.
For example, a collaboration with the United Nations announced last year has already trained more than 14,000 data scientists on how to design AI models for early flood detection.
Entrepreneurs increasingly see opportunities here, too.
Sustainable Futures, an initiative of the NVIDIA Inception program, now includes more than a thousand startups pioneering new paths to energy efficiency. They apply a mix of AI, accelerated computing and open platforms like NVIDIA Earth-2 to speed climate and weather predictions.
Savings Span Global Industries
These examples provide a taste of what’s happening and what’s to come.
“The generative AI revolution is poised to impact every industry and enable a new era of productivity and sustainability by unlocking efficiencies and resource optimization across sectors,” said NVIDIA founder and CEO Jensen Huang in the company’s latest sustainability report.
It’s a reality independent analysts recognize.
“Even if the predictions that data centers will soon account for 4% of global energy consumption become a reality, AI is having a major impact on reducing the remaining 96% of energy consumption,” said a report from Lisbon Council Research, a nonprofit formed in 2003 that studies economic and social issues.
45,000x Efficiency Gain in AI Inference
NVIDIA works continuously to expand and optimize its full stack of technologies to enable greater efficiencies.
From 2016 to 2025 when the NVIDIA Blackwell platform will ship in volume, NVIDIA AI computing will have achieved approximately 10,000x more energy efficiency for AI training and inference.
To put that into perspective, if the efficiency of cars improved as much as NVIDIA has advanced the efficiency of AI on its accelerated computing platform, they’d get 280,000 miles per gallon. That’s enough to drive to the moon, and back, on less than a gallon of gasoline.
Speedups over CPUs from 20x to 160x — and their associated energy savings — are available across the gamut of workloads from data processing to computer vision the NVIDIA accelerated computing platform serves. That’s why accelerated computing is often called sustainable computing.
Expanding Climate Week’s Agenda
Rallying support for sustainability, NVIDIA and partners are taking part in a handful of other Climate Week events.
They include a Climate AI Summit with Salesforce and a Climate Science Fair with about 30 startups, many using NVIDIA technologies. The Science Fair will give attendees an up-close look at how Earth-2 enables simulation on a planetary scale.
It’s another small step toward helping people understand how to use AI and accelerated computing to address climate change.
“Accelerated computing is how to meet the massive demand for computing power sustainably and cost-effectively,” Huang said.
Learn more about how NVIDIA’s technologies support corporate sustainability.