Dassault Systèmes Tackles AIâs Rising Energy Demands
AI-related data centre energy use is projected to comprise up to 3% of global power consumption by 2030, with some regions experiencing even higher demand.
With this in mind, the International Organization for Standardization has released ISO/IEC 42005 guidance to help companies evaluate AIâs societal impacts, including its environmental footprint.
For example, in Ireland, a hub for digital service providers, AI data centre energy demands could represent 35% of national power consumption.
Despite these projections, Philippine de TâSerclaes, Chief Sustainability Officer of Dassault Systèmes, remains optimistic about AIâs environmental potential.
The potential of machine learning for data centre sustainability
Machine learning (ML) algorithms hold promise for enhancing grid efficiencies by 15% and increasing battery storage efficiency by 10-20%.
AI might also shorten timelines for new clean energy projects by 20%, potentially saving hundreds of billions of dollars by 2050.
McKinsey suggests that AI and machine learning could expedite 47% of initiatives crucial for achieving the global 1.5-degree pathway under the Paris Agreement.
This technology is showing potential for sustainable innovation.
AI has already contributed to developing lightweight packaging that reduces transport energy needs and advanced materials for more efficient batteries.
Additionally, recent research highlights AI-developed paint coatings capable of reducing building temperatures by 20°C.
However, the current deployment patterns of AI are raising growing concerns related to energy sourcing.
We have what we need to make the infrastructure and ecosystems that power AI start working more effectively today.
Dassault Systèmesâ partnerships for efficiency gains
MIT Technology Review has found that, as applications grow rapidly, data centres are becoming increasingly reliant on carbon-intensive energy sources.
In response, Dassault Systèmes has partnered with Taiwanese server manufacturer Quanta Cloud Technology to enhance efficiency.
Philippine advocates for a 'Frugal AI' approach that prioritises lightweight models and measurable impacts.
Model pruning is one such technique, allowing the removal of unnecessary neural network connections to decrease computational needs without compromising accuracy.
This method reduces energy use while maintaining performance. Optimising data centres presents immediate opportunities for efficiency.
As cooling systems can account for up to 40% of energy consumption, choosing effective systems could boost efficiency by 30%.
At Dassault Systèmes, virtual twins on the 3DEXPERIENCE platform enable sustainable operations via partner collaboration.
These scenarios help perform trade-off analyses between technology configurations to optimise setups and accommodate future technologies.
Dassault Systèmes also collaborates with QCT to simulate data centre heat and airflow patterns, which are critical for designing efficient air conditioning systems.
The company's solutions support Bouygues Construction in modular construction processes and aid Olivier Naar in the design of modular nuclear reactors.
“I don’t see isolated projects,” Philippine says, “I see interconnected, mutually enhancing nodes within a wider value network.
“I see how those same techniques can help build modular data centers. I see how we can power them with electricity that is low-carbon and convenient.”
The importance of broader thinking
Network effects are fostered through initiatives like the Coalition for Sustainable AI, where Dassault Systèmes is an active participant.
However, Philippine notes that technical solutions alone are insufficient for tackling sustainability challenges.
The value derived from processing power is contingent on its application, independent of data centre efficiency.
“AI will be what we make of it,” she says.
“Circular thinking needs to be embedded not just in our processes and products, but also in our methods and in the way we think about the world.”
She concludes: “We have what we need to make the infrastructure and ecosystems that power AI start working more effectively today.
“What we make of it? Well, that’s up to us.”
