@ green | March-April . 2024

Can fast AI adoption co-exist in a world striving for net zero by 2050 ?

2015 – 2022 : HONEYMOON YEARS FOR GREEN ENERGY

The targets were ambitious and noble from the onset , signalling the start of an accelerated energy transition process expected to bring climate change into control by substantially reducing the use of fossil fuels in the global energy mix .

The media , climate change activists , politicians , government regulators , the financial sector , young people , and many other streams of society embraced the concept of a cleaner and cooler future with great enthusiasm .

Throughout human history , there have been several transitions . While technology and economic benefits underpinned previous transitions , public policy has driven the current energy transition virtually exclusively .

In the early 2020s , the Covid-19 pandemic ravaged the global economy . Affordable energy was the key , but by COP 26 , held in Glasgow , Scotland , in late 2021 , the long-term well-being of humankind once more became a priority . The position of sustainability was restored as the prime element of the Energy Trilemma .

With a boost from Boris Johnson , the then-British Prime Minister , the world was again on the Energy Transition trail . In fact , in a seminal report issued in 2021 before COP 26 which is titled “ Net Zero by 2050 - A Roadmap for the Global Energy Sector ”, the prestigious International Energy Agency ( IEA ) authoritatively charted a Pathway to a Net Zero world by 2050 , boldly claiming the following :

“ There is no need for investment in new fossil fuel supply in our net zero pathway .

Beyond projects already committed as of 2021 , no new oil and gas fields are approved for development in our pathway , and no new coal mines or mine extensions are required .”

Unexpectedly , however , paradigms had to be reset when Russian military forces made an incursion into Ukraine in early 2022 . Since then , energy security has been a critical priority , given that much of the European gas supply traditionally came from Russia .

The energy supply chaos that resulted from the Russia – Ukraine war led to the European Union having to change definitions within its green taxonomy , allowing gas and nuclear as legitimate clean energy sources .

It was already evident that fulfilling the energy transition goal would be complex and costly . Were there further hurdles ahead ?

ENTER THE FORCE

In early 2024 , a design studio for semiconductor chips , Nvidia , having most of its manufacturing outsourced , achieved valuations trading at 25x its revenues and 73x its net profits ( valuation of circa

BY DR KENNETH PEREIRA

US $ 2.18 trillion ).

This valuation surpassed some of the highest metrics seen during the dot-com era . What is so special about Nvidia ? Whilst Nvidia has been around for more than two decades , its chips were recently found to be best suited as a platform for the most cutting-edge technology available - Artificial Intelligence ( AI ).

AI has been conceptually around for nearly a century in one form or the other . Visionaries like Alan Turing envisaged AI in the 1940s . His earliest paper on the subject was published in 1941 . More recently , a form of AI captured the public imagination when , in 1997 , Gary Kasparov , the reigning world chess champion , was defeated by IBM ’ s Deep Blue ( a chess-playing computer programme ).

If forms of AI have been around for over two decades , what has changed to propel it as a modern-day technology force ?

In November 2022 , Open AI began offering Chat GPT for public use . Chat GPT is based on generative ( thinking , so to speak ) AI as opposed to reactive AI ( like Deep Blue ), and “ Open Access ” means that AI quickly captured the attention and imagination of the world .

It therefore came as no surprise that the Chat GPT user base rose dramatically , reaching a staggering 100m users within two months of launch , and it has increased even further since , to about 180 million users today .

This uptake in utilisation can be explained by the fact that Chat GPT could communicate with users . It was novel and made the tool genuinely path-breaking . Several commentators have likened generative AI to a standalone industrial revolution . This is because generative AI ( like the original Industrial Revolution ) can reduce human involvement in processes and activities .

The earlier industrial revolution reduced human effort and transferred those heavyduty activities to coal-powered machines . This time , it is thought that generative AI can be expected to take over even more and more human tasks .

A graphic processing unit ( GPU ) is at the heart of AI technology . Nvidia is the leader in designing these new-generation semiconductor chips , which explains why it stands out among investors . However , a contentious issue is surfacing slowly . AI will touch every aspect of our lives . To do this will require ever-increasing amounts of energy .

There have been some attempts to quantify this new requirement . An article from the reputable Scientific American magazine referred to “ a peer-reviewed analysis by a researcher , Alex de Vries , published in Joule . He is among the first to quantify the new energy demand quickly materialising due to AI .

“ A continuation of the current trends in AI capacity and adoption is set to lead to NVIDIA shipping 1.5 million AI server units annually by 2027 . These 1.5 million servers , running at full capacity , would consume at least 85.4 terawatt-hours of electricity annually - more than what many small countries use in a year , according to the new assessment .”

As a comparison , the 170 million population of Bangladesh used just 79 terawatt-hours of electricity in 2023 , according

Figure 1 : Increasing Trend of Carbon Dioxide Levels Over the Past Fifty Years