How sustainable AI improves the triple bottom line

An investment in green AI enables financial services firms to align people, profit, and planet

By Nick Dale, EVP business development, Verne Global

Green investing is widely regarded as a mega trend, with chief executive Larry Fink of BlackRock, the world’s largest money manager, stating, “Climate change has become a defining factor in companies’ long-term prospects … awareness is rapidly changing, and I believe we are on the edge of a fundamental reshaping of finance.”

The recent seismic shift in public opinion about climate change has not only increased attention on the sustainability and societal impact of investing in a company, it’s also influencing the decisions being made in finance industry boardrooms overall, whether that’s implementing innovative business models or adopting new partnerships and technologies. However, as business leaders strive to make green choices, many are unaware of the hidden environmental costs of the technologies they are employing.

AI in the finance industry

The use of AI has become ubiquitous across industry sectors, and is now an integral part of the technologies being used in financial services, from optimising asset portfolios and underwriting loans to assessing risks.

AI is especially beneficial for things like quantitative trading, which uses large data sets to identify patterns that can then inform strategic trades. AI’s machine learning models can analyse vast and complex data and make predictions accordingly. But AI models are not only data-hungry, they are power hungry.

Power-hungry AI

Supercomputers train and test mountains of data for AI models, and can run 24-hours a day, for hours, days, or even weeks. These applications consume huge amounts of energy, and as AI technology continues to grow and develop, the computations behind it are also increasing in size and complexity. The carbon emissions from training a single AI model for language translation is roughly equivalent to 125 round-trip flights from New York to Beijing (AI Now 2019 Report).

The carbon cost of AI becomes even higher when you factor in the energy required to keep the computing equipment housed in data centres cool – overheating can impact performance and damage equipment. As a result, in a conventional data centre, at least 40% of all energy consumed goes towards cooling.

But sustainable AI is possible if financial services organisations take positive steps to minimise its environmental impact.

Minimising AI’s carbon footprint

Location, location, location

Many tech giants are committing to reducing their carbon footprint, with Amazon pledging to reach 80% renewable energy by 2024, and Google investing in data centres in Nordic countries specifically for better energy efficiency.

This is because in the Nordics, data centres are largely powered by renewable energy sources. Iceland, in particular, uses 100% renewable hydroelectric and geothermal power – with no nuclear power sources – and is connected to a reliable power grid. These renewable energy sources are much less harmful to the environment because, unlike fossil fuels, they don’t cause pollution and don’t generate greenhouse gases. Not to mention, renewable energy is based on natural resources that can be replenished within an average human lifetime, as compared to fossil fuels, which can take thousands—or even millions—of years to replace.

Over 80% of compute doesn’t need to be near the end-user, and in those situations, choosing data centre locations in cool climates has a significant impact on carbon emissions. AI compute can be located in places like Iceland, which can utilise all-year-round, free cooling due to its temperate climate.

Data centres that are located in hot climates, like Arizona in the US, require high-powered cooling systems in operation around the clock. With average high temperatures of 40° Celsius in the summer, these data centres can use up to 4 million gallons of water a day to absorb heat through evaporation into cooling towers. Consequently, when location doesn’t hamper performance or accessibility, housing AI compute in data centres with natural cooling is a no-brainer.

Energy efficient and cost-effective

Many in the financial sector have traditionally viewed sustainability as a trade-off between profit and planet, but when it comes to green AI, financial services firms can have it both ways. By housing the servers that train AI models in data centres powered by renewable energy sources, businesses can substantially reduce energy expenses and benefit from long-term, fixed pricing.

And when renewable energy sources are combined with year-round, cool climates, the energy demands and costs of AI can be dramatically reduced. AI is here to stay, but by making the right choices, companies in the finance sector can still drive profitability whilst making real and measurable progress on sustainability.