Organisations facing sustainability challenge amid AI ambitions

The sustainability invoice

With tightening legislation regarding emissions, organisations are forced to pay closer attention to the environmental impact of their tech consumption. Nathan noted that companies like NTT Data can now provide “the science of the possible,” thanks to Cisco’s transparency regarding product data.

While many customers are not yet explicitly asking about emissions associated with service production, Nathan observed that “the smart ones get that”. These mature organisations are now looking for a “sustainability invoice” alongside the standard financial invoice, making carbon footprint a key factor in buying decisions alongside features and price.

Cisco has responded by adopting circular design principles. de Wysocki noted that all products introduced this year are more modular, easier to disassemble, and delivered with reduced packaging. However, she stressed that the company’s primary responsibility lies in the electricity consumed by its products in the field, where improved power efficiency directly correlates to a lower total cost of ownership.

Converging IT and energy

A critical aspect of modern sustainability is the ability to respond to signals from energy providers regarding peak demand – for example, by suspending non-critical workloads until a peak has passed.

Sustainable IT is not just about lower power usage and supplier ‘take-back’ arrangements, but the use of technology, including AI, to drive organisational efficiency, said Nathan. He highlighted Cisco’s support of Power over Ethernet (PoE), which helps organisations converge their IT and energy networks.

de Wysocki expanded on this, pointing out that automating the physical workplace can significantly reduce energy consumption. Simple measures, such as automated window blinds that adjust to the sun’s position, can help. She noted that Cisco itself has achieved a strong return on investment by deploying these technologies internally.

Sustainability measures are becoming a standard requirement for leading customers, but they should also be viewed as a matter of risk mitigation. Responding to severe weather incidents, for instance, carries a tangible cost.

However, Nathan suggested that most organisations currently lack the maturity to execute these projects independently. Consequently, many begin with consulting engagements to determine feasibility, costs and projected benefits. The cost of necessary hardware, such as sensors, has fallen sufficiently to justify these initiatives.

Most of these engagements are now conducted with the chief financial officer, Nathan observed. Financial benefits remain the easiest sell, and for large organisations, even a 1% energy saving is significant. However, he warned that the data used in these calculations must be robust to ensure the modelling results are trustworthy.

Going beyond FOMO

Regarding AI, Nathan stressed the importance of clearly defining the business problem and the value of solving it.

A vague desire to become an AI company is insufficient and often a symptom of “fear of missing out” (FOMO), said de Wysocki. She suggested leaders should instead ask: “What is the capability that will add value?” and warned that time to productivity is critical.

Fortunately, most organisations are more similar than they realise, allowing them to apply existing reference architectures. Best practices have already been defined in many cases, Nathan said.

But if an organisation has set itself the goal of monetising AI more efficiently than its peers, that will consume a lot of energy and so it needs to be done efficiently.

Ultimately, energy management is set to become a make-or-break topic. “Lowering your electricity use is a survivability measure,” Nathan said.