Chipmakers should come together and share knowledge on how they can improve sustainability for organisations, which is critical as artificial intelligence (AI) adoption continues to scale.
Collaboration is essential to address a major challenge the semiconductor industry faces, which is sustainability, said Henri Berthe, vice president for global semiconductor at Schneider Electric. The French vendor specialises in digital automation and energy management, providing a range of connected products including sensors and digital twins.
Market players in the semiconductor ecosystem today are not utilising the tools available to achieve better energy efficiency, said Berthe, in a video interview with FutureCIO.
This could be due to various reasons, such as a lack of competencies and concerns about the impact on capital expenditure, he said. Sustainability also may not be a primary focus for the chip manufacturer, he added.
In addition, the semiconductor industry is typically guarded and do not like to share their knowledge. So there may be a lack of understanding around best practices.
Berthe urged the need for closer cooperation across the semiconductor supply chain and ecosystem, so the industry can better prepared for growing AI adoption.
Deeper engagement can help plug the current lack of knowledge within the supply chain about what can be done technically to drive better energy consumption within the fab, he said.
For example, there is insufficient exchange of best practices on how AI can be tapped to drive sustainability, he noted.
Schneider Electric hopes to change this through its Catalyze program, he said, which aims to bring together semiconductor players and their suppliers in their transition towards renewable energy sources.
The initiative provides participants access to knowledge resources and expert consultation to help them deploy renewable energy. Smaller suppliers, for instance, can leverage collective purchasing power via the program and establish better power purchase agreements.
Catalyze’s sponsoring companies currently include Intel, Applied Materials, and Google, enabling any company that supplies to these sponsors to join the program.
There now also are similar efforts from industry consortiums and other market players to fuel a collective focus on improving sustainability, Berthe said.
Sustainability needs to stay on the top agenda of these ecosystems, he stressed, especially as it is increasingly a criteria for customers.
Designing data centres that need less power
And with AI chip clusters running much hotter, achieving better sustainability requires data centres to be redesigned and more efficiently deployed.
This means starting from the architecture, including figuring out what is the most optimal design and how it can be managed in the most optimal way, Berthe said.
“We need to start by designing new data centres that are more sustainable,” he said, adding that the biggest bottlenecks in datacentre sustainability today are cooling and power.
He noted that energy consumption at data centres has been growing exponentially and will continue to climb in coming years.
“Energy scarcity will be a major challenge in future...so we should be building systems that are more sustainable [and] designing [semiconductor] fabs that are more sustainable and that consume less power,” he said.
Efforts should also encompass implementing solutions on existing fabs, to enable them to operate more sustainably, he added.
AI, too, can be tapped to monitor equipment and predict potential system failure, Berthe said. The aim here is to minimise downtime and extend asset lifetime, he said.
AI can analyse how energy is consumed across systems and buildings, and offer insights on how consumption within fabs can be optimised and reduced, he noted.
This can generate significant savings, since energy contributes between 15% and 30% of a fab’s overall operating expenditure, he said.
And it appears that organisations are keen to leverage AI to fuel their green objectives.
Some 76% regard AI and cloud computing as essential tools to achieve sustainability outcomes, according to a March 2025 study from Alibaba Cloud, which polled 1,300 business leaders in Asia, Europe, and the Middle East.
These figures are highest, at 83%, amongst respondents in Asian markets, with 91% in the Philippines, 84% in Singapore, and 81% in Thailand as well as in Indonesia keen on the potential of AI and cloud in driving sustainability.
Another 82% globally believe it is critical these technologies are developed sustainably, the survey found.
However, 59% say they lack understanding of how digital technology can help achieve their sustainability goals. This figure is highest in Asia, at 63%, followed by 61% in Europe and 45% in the Middle East.
Globally, 62% describe their organisation as lagging in adopting cloud and AI to drive their sustainability targets. In Singapore, this number is at 80%, with 77% of their peers in the Philippines and 75% in Japan acknowledging that their organisation is behind in tapping AI and cloud for better sustainability.
Data silos still a struggle
Companies also are facing barriers applying AI in product design and manufacturing.
Fragmented data and systems are a key challenge, Berthe said, noting that enterprises commonly struggle with many disparate tools and siloed data.
Without a unified view, it is difficult to extract useful insights or to automate any process effectively, he explained.
AI also requires new skillsets and capabilities, he added.
Engineers and data scientists are skilled primarily in IT and data, respectively, with the former strong in processes but not necessarily in the data field, and vice versa.
These two knowledge areas need to be combined to better understand, for instance, how energy consumption can be optimised, Berthe said.
There also are clashing values in AI, where the technology can be leveraged to improve energy efficiency, but at the same time, consumes significantly more energy to run.
This can prove a major roadblock, Berthe said.
In addition, organisations must look closely at change management because AI will alter how people work, he said.
There will always be resistance to change and this needs to be addressed, he added.
Companies will have to establish trust and transparency, he noted, stressing the importance of training.
