In the whirlwind of technological advances, two revolutions stand poised to reshape our world: the rise of Artificial Intelligence (AI) and the urgent shift to clean energy.
These concurrent shifts promise to drive economic growth through productivity, employment, and investment. In terms of investment volumes looking ahead until 2030, the energy transition will probably be larger by a factor of ten. Over that period and beyond, these shifts will intersect in ways that could amplify their benefits—or challenges. Both will happen in similar geographies, notably China, North America, the European Union and India. They will also access similar pools of global capital.
AI will be an enabler for cleaner energy deployment
The energy transition needs to be managed in a way that it does not impose high costs on the consumers, while ensuring reliable energy supplies and AI will act as an enabler to achieve both. At ReNew, leveraging AI has not only improved our electricity output by up to 1.5% from existing solar and wind installations but also streamlined maintenance, demonstrating AI's potential to enhance efficiency and reduce costs.
Big data, and innovation in analytics, enable us to measure inputs from satellites, sensors and weather monitoring stations to predict solar radiation and wind speed, accurately forecasting the supply of renewable energy generation. On the side of the equation, AI is accumulating terabytes of historic consumer data to forecast consumer demand for electricity. Balancing supply and demand is critical in preventing supply disruptions and blackouts.
Globally, almost $3 trillion worth investment is being allocated between now and 2030 to lay the wires and infrastructure to transport clean energy from points of generation to the consumers. Several companies are already leveraging AI for strategic decision making in terms of planning which type of grid is suitable to which location, all the way down to the size of the wires. With several of these wires running thousands of miles, it is difficult to inspect and maintain them. New machine learning software predicts anomalies in wiring and failures of transformers, saving time and money. While actual numbers are larger, even 5% savings on capital expenditure for installation and replacement, will result in reduced expenditure of $150 billion in the next 7 years.
However, AI is an enabler for more efficient and sustained fossil fuel driven activities too
Like the internet, AI is a tool that is useful for everyone, including the fossil fuel sector. It is an equalizer. Companies like BP, Shell, Exxon are already using AI to lower the cost of extracting oil and gas. Autonomous vehicles, based largely on AI, are increasing in numbers - most of which run on gasoline. By making travel cheaper and more convenient, autonomy could increase the number of vehicle miles travelled. If media reports are to be believed, General Motors Co.’s Cruise and Alphabet Inc.’s Waymo are likely to begin offering self-driving taxis in San Francisco shortly. The merger between electrification and autonomy of vehicles is likely to take some time, based on improvements in batteries, sensors, and computation capabilities.
AI will also be a huge energy guzzler
Training AI models (meaning setting up AI models to spot patterns in datasets) and delivering inferences (meaning numbers, text, videos, imagery based on the patterns) require huge amounts of computing power and data storage. The energy demands of AI, particularly for data centers, are soaring, potentially rivalling the consumption of entire countries like Brazil, South Korea or Germany. According to the IEA, data centre energy usage stood at around 460 terawatt hours in 2022.
To be candid, I am cautiously optimistic about deploying Artificial Intelligence. Drawing lessons from the energy transition journey, there are three areas where we must collectively pay particular attention, to ensure that AI makes a strong positive contribution to humanity.
Diversification of solution providers: A single US based firm holds around 80% of the high-end AI chip market. The few big tech firms hold most of the computational capabilities, datasets and servers that will enable AI to even function. Much like energy transition, capabilities are concentrated in one or two nations, posing risks of disruptions and trade controls. A wider set of nations, including from the global south, need far more active domestic policies to develop their own AI technologies, solutions, and business models.
Governance to ensure reliability, accountability, and dispute resolution: As the use of AI grows, there will be ever more capturing of data. This makes us prone to errors (due to use of poor-quality data), cyber-attacks and data-theft. These will need appropriate legal provisions, that ensure access by clients to datasets used by service providers and allocation of responsibilities for safety and privacy of the data.
Making it low carbon, before we are locked-in: AI being a sunrise sector, presents an opportunity for being lower carbon right from its early stages. We have the technological solutions to do so and a number of the top 10 companies globally that run data centres have adopted bold targets for achieving net zero emissions. Accountability towards meeting these targets will need to be ensured over the next few years. Large investors, that have many of these companies in their portfolios, currently seem to be focused on understanding and minimizing the social risks of AI. There must also be a focus on the environmental implications. Equally, the biggest clients must start accounting for emissions due to AI services received by them in their Scope 3 emissions, and take steps to reduce them significantly.
As we stand at the crossroads of these technological revolutions, our choices today will determine whether AI becomes a pillar for a sustainable future or a missed opportunity. Embracing these lessons with caution and optimism is not just advisable; it's imperative.