New research from McKinsey has revealed that, while semiconductor companies around the world are making significant progress towards sustainability, further commitment is needed in order to reach net zero. Without concerted effort, the 1.5 degrees Celsius goal by 2030 is a pipe dream.
This was highlighted at the 2022 Climate Change conference – COP27 – where semiconductor sustainability was high on the agenda. We’ve also seen the establishment of the Semiconductor Climate Consortium (SCC): 60 likeminded specialists seeking to standardise sustainability within the industry.
It‘s self-explanatory that increasing demand for semiconductors will drive up emissions from production. So, what can be done to avoid this?
The key lies in understanding the importance of implementing sustainable methods of chip manufacturing, focusing on the energy efficiency of the device in which the chip resides.
That requires AI at the edge – the Artificial Intelligence of Things (AIoT).
How AI can contribute to sustainability in the semiconductor industry
The AIoT is the point of convergence between AI, the ability to understand data or stimuli in a more sophisticated way than standard devices; and the IoT, the capacity of electronics and machinery to ‘speak’ to one another, sharing data and instructions. This combination enables previously impossible features for everyday devices that can offer creative, unique solutions to sustainability.
For example, an ordinary smart speaker may use 10 watts of power daily in listening out for keywords while on standby. But a device with the right AI has plenty of tricks, such as AI imaging capable of determining human presence. If no human presence is detected, the speaker could then decide not to switch on its ‘ears’ – using minimal power to understand its environment, rather than maintaining an always-on model.
If every household electronic was to integrate AIoT functionality, the reduction in global carbon emissions would be exponential. Pushing this technology should be a focus in the short-to medium term if we are to achieve optimum semiconductor sustainability at some point down the line.
However, the failure to utilise AI to the fullest extent is only part of the problem. As this is only actionable once the production is complete, we also need to focus on opportunities for sustainability in the manufacturing stage.
Opportunities to identify in the manufacturing stage
According to research from McKinsey, 35% of emissions at a semiconductor factory are caused by process gasses with high global warming potential. This alarming figure combined with pollution caused by other areas of manufacturing, such as the extraction of raw materials and the transporting of finished goods, is obviously a concern.
A good way to start implementing sustainable methods is carrying out an audit of the supply chain. This will allow the company to point their engineers to the areas that need improving the most.
Furthermore, designing sustainability into the device itself is a great effort in reducing electronic waste. Research by Accenture reveals that if a company was to implement remanufacturing, repairing, and refurbishing while manufacturing electronic devices, they would see a 16% increase in operating profits, an 80% reduction in material losses and a 45% reduction in carbon emissions.
Responsible design of the device can make the chip a positive contributor to the health of the environment. Overall, increased sustainability in chip production and deploying these chips into devices that are designed with energy efficiency in mind is the way to go.
But the first big success for sustainability in the semiconductor scene would be the widespread knowledge of the potential that is held by those in the industry.
COP27 was an eye opener for many and undoubtedly will have set a good number of semiconductor companies on the correct path. With a light being shone on how players in this industry can contribute to a greener planet, let’s hope that COP28 will be proof for the world that the semiconductor industry has turned over a new leaf.
