The artificial intelligence (AI) boom has had such a profound effect on big tech companies that their energy consumption, and with it their carbon emissions, have doubled in just the last 4 years.
Greenhouse Gas (GHG) Emissions
Data centres currently account for 2% of global GHG emissions (equivalent to the Aviation industry), with projections suggesting global GHG emissions attributed to data centres could reach 14% by 2040.
Google’s latest environmental report stated that it’s GHG emissions in 2023 were 48% higher than in 2019.
Noting that a ChatGPT query takes 10x as much energy to process than a google search (Source: IEA).
Energy Consumption
Emissions are not the only problem.
Data centres also require a lot of power. Massive increases in data centres will put incredible strains on electricity grids that are already grappling with increased power needs from things like plug-in hybrids, electric vehicles, general population growth and sprawl, baseload supply constraints, intermittent renewable technologies while intentionally collapsing fossil fuel fired power infrastructure.
Data centres’ electricity consumption in 2026 is projected to reach 1,000 terawatts, roughly Japan’s total consumption.
Goldman Sachs Research estimates that data centre power demand will grow 160% by 2030.
Industry Solutions
There are several ways big tech is trying to address this current and worsening energy crisis.
First, computing hardware has gotten substantially more energy efficient over the years in terms of the operations executed per watt consumed, however hardware efficiency gains have slowed down substantially, as the industry has hit the limits of chip technology scaling.
Unfortunately, efficiency alone is not going to solve the sustainability problem.
Paradoxically, it is likely to result in further increases of energy consumption in the longer run; with more energy consumed as efficiencies are realised, exploited, and pushed further.
Industry is also looking at new ways of building AI data centres to incorporate flexible computing, where the key idea is to compute more when electricity is cheaper, more available, and greener, and less when it’s more expensive, scarce, and polluting.
Therefore, ideologically dovetailing with the inherent intermittent availability of weather-based energy sources – what could possibly go wrong with that?
If you’re keen to know more, suggest you research Green AI vs Red AI to make your head really hurt, as it eerily mirrors the Renewable vs Fossil Fuel discourse.
Investment, Finance, and Insurance
Following form, it would be logical to conclude that our global financial services sector will now start moving to address this issue through scrutinising big tech’s Environmental, Social and Governance (ESG) credentials, picking them apart and openly chastising and threatening them with withdrawal of support – with any sign of populist dissent a marker to cause them to evaporate entirely like the preverbal rain drop on a hot tin roof?
NVIDA, Apple, Google, Microsoft release their ESG reports and an eyelid is barely moved as the financial sector collectively falls over themselves to ride the gravy train.
Ultimately, we have seen our sector (insurance), finance and investment cruelled in the mining and power sectors while we exponentially pursue an unsustainable energy consumption model across the globe.
Hypocrisy abounds.