The Symbiotic Relationship between Energy, Big Tech, AI

We recently recorded a fascinating conversation with Ray Dalio, which we’ll soon share on the Giant Ideas Podcast. In it we discussed his latest book, Principles for Dealing with the Changing World Order, where Ray explained that he believes that there are five forces changing our world: debt, conflict, populism, climate change and the power of technology. He believes that understanding these timeless forces can help us position ourselves for what’s ahead.  

The conversation made me think about three powerful forces shaping the technological landscape now: the relationship between Energy, Big Tech, and Artificial Intelligence (AI). The AI boom, propped up by Big Tech, has catalysed sizable infrastructure build-out of data centres, driving up the demand for energy. This article considers the interconnected dynamics and significance for future innovation.

Data centres driving the AI demand

Right now, we are witnessing one of the largest infrastructure build outs in history. The need for computational power to sustain AI’s growth is doubling roughly every 100 days, resulting in a significant surge in data centre construction around the world. In the first half of this year, AWS announced $50B invested in new data centre projects, with Amazon committing $100B-$150B over the next 15 years. Microsoft reportedly now has 5GW of energy capacity and is set to double new data centre capacity this year. The smallest of the three cloud providers, Google, has also announced $3.5B invested - in Indiana, Finland, and Iowa.

The energy consumption of these data centres is staggering. For example, OpenAI’s ChatGPT requires about 2.9 watt hours of electricity per query, almost ten times that of a Google search (0.3Wh). As AI continues to evolve, and with models like GPT-4 requiring thousands of GPUs to work simultaneously, the need for data centres will only grow. This is bringing energy - a previously invisible force in Europe and the US - into the forefront.

Big Tech getting bigger (and the climate challenge) 

Google, Microsoft, and Amazon are certainly leading the race in AI development; they are major purchasers of GPUs and foundation models, and are spearheading the construction of energy-efficient data centres. They are also growing to unprecedented size. With some of the so-called “Magnificent Seven” already crossing the 3 trillion mark, are we going to see these companies cross the ten trillion-dollar mark due to this AI wave?

It sounds far fetched, but so did a trillion-dollar company a few years ago. Or is this reflexive demand going to unravel slowly and then all at once? In “The Engines that Move Markets” - an excellent book on technology investing - Nairn argues that there is nearly always an overbuild of the underlying infrastructure in a first wave of enthusiasm. But there’s no doubt in my mind that ultimately this infrastructure will continue to expand and roll out for a while longer.

It will also be those Big Tech juggernauts that will shape the future sustainability of this wave. Google’s greenhouse gas emissions have surged 48% in the last five years due to their data centre expansion, and Microsoft’s rose a third since 2020. Many of them acknowledge the challenge and are ostensibly trying to reduce these emissions. Debate rages, even between these tech giants, on the right way to account for emissions - especially when it comes to energy usage in the US where many state grids are still heavily dependent on fossil fuels. The Financial Times recently reported that: “companies including Amazon, Meta and Google have funded and lobbied the Greenhouse Gas Protocol, the carbon accounting oversight body, and financed research that helps back up their positions”. Should they be able to buy renewable energy certificates (RECs), or do these techniques distract from the true surge in tech company emissions? As pressure rises, the accounting debate will continue - rules are only expected to be finalised in 2026 - and there will be an incentive to adopt new innovations to solve the issue.  

Hungry data centres’ ripple effect on energy innovation

It is true that tech giants have been taking steps on energy solutions for over a decade. Fourteen years ago Google invested in two wind farms, and the Economist reported earlier this year that tech companies are by far the biggest buyers of green energy via power-purchase agreements (PPAs). These are “long-term contracts between a buyer (like a data-centre company) and a power producer (like a solar or wind-farm operator) to buy electricity at a predetermined price, instead of bulk-buying from utilities.” These PPAs give essential security for new renewable projects to get built. Big Tech is also driving new types of energy demand: last year, Microsoft signed an agreement with Helion to provide them with electricity from its first fusion power plant in 2028. Amazon Web Services recently bought a nuclear-powered data centre from Talen Energy for $650M.

The energy demand will also provide a meaningful market pull to innovate. The solar industry is exploring using robotics to speed up the construction process - responding to the so-called “dawn of the solar age”, where we are seeing solar capacity doubling roughly every three years. There are also ripple effects in innovation across the energy spectrum from the grid, power lines, generation, to distribution. Innovative energy solutions have emerged in companies like Crusoe Energy Systems which designs, builds and operates modular data centres and uses stranded energy (like flare gas) to drive a 60 – 80% energy cost advantage. At the same time, companies like Veir have built cutting-edge technology for superconducting power lines that offer 5-10x greater transfer capacity. A few ambitious entrepreneurs are even proposing building Data Centres in space.

AI supercharging the climate fight

Somewhat conversely, AI as a technology also enables solutions for the energy transition. AI-driven predictive analytics are essential for energy storage and improved grid stability, and AI is essential for infrastructure asset planning and maintenance, like assessing cables underground without digging them up or assessing parts and locations while planning a wind farm. In an ouroboric way, the rise of data centres both creates climate challenges, with the use of water, energy and land, but also enables potential solutions - both by catalysing renewable energy build out, and also by improving legacy industries - for example expediting natural crop evolution (Avalo), improving battery R&D (Byterat), and even working with Big Tech like Meta and Amazon directly to mitigate water risk (Waterplan).

AI considerations at Giant

At Giant, we are committed to investing behind the AI wave, underpinned by the belief that further innovations on hardware, storage, cooling and energy technology will continue to reduce the climate impact of AI. AI has the potential to radically disrupt some industries, and expedite others. We wrote previously about how we are excited about how LLMs will reimagine operations and workflows across healthcare and education, and we have made some recent investments from our seed fund that reflect this thesis. 

There is a huge data centre build out happening due to demand for AI, a buildout comparable to what we saw in innovations like the railroads. In the short term, this investment will spike energy demand and increase carbon footprint, but long term, it looks possible that it will both drive innovation in energy, provided emissions aren’t gamed, and provide the technological firepower required for vital climate solutions. Big Tech is clearly driving the AI wave and will shoulder a lot of the responsibility to ensure that the technology is net positive. These three major forces in the world economy, - AI, Big Tech, and Energy - will shape our world for the coming decade, and we should follow them closely as the impact unfolds. 

For more energy insights, listen to  Octopus Energy CEO Greg Jackson or OVO Energy CEO Stephen Fitzpatrick on the Giant Ideas podcast, or follow the show here.