The next 3-5 years of the energy transition will dictate what the global energy landscape looks like for decades to come. The grid of 2040 and beyond must look entirely different than it does today and. although building that grid is already underway, it isn’t yet clear whether we will have the grit, patience and thoroughness to build it right. By “right,” I mean to allow for scaled and sustained growth that meets the energy demands of AI and an expanding population in a warming world, well into the latter half of the century.
Energy providers today face a near-impossible task - rebuild the energy grid while simultaneously bringing online unheard-of levels of new energy output.
The impossibly steep drive for AI growth, rising global temperatures and mass electrification of industry are expanding energy demand at a breakneck pace. Currently, this growth in demand far outpaces a feasible growth in energy supply, by any combination of renewable, nuclear or fossil energy.
Why? A few key features:
Energy and grid stabilization projects of this scale cost hundreds of millions of dollars to deliver, requiring major coordination within large energy providers
Procurement lead times alone for this specialized equipment are often 2+ years long, on top of an additional 2-3 years for design and construction
Projects must go through extensive, time-consuming regulatory approvals before they can be connected to the grid
The energy grid itself must be redesigned and supported in key areas - requiring additional multi-year projects - to tolerate the massive energy spikes and dips associated with heavy load customers like datacenters
This essentially means that no matter the level of investment available from the AI sector, we are beginning to see years-long gaps between projected completion dates of datacenters and the ability of energy providers to support them. Furthermore, while the AI boom accounts for the lion’s share of energy demand growth, two other factors are already contributing significantly to grid stress.
Growing populations paired with rising global temperatures - hitting hardest in the summer - are leading to greater and greater use of air conditioning, for more hours each day. Did your energy provider ask you this past summer to cut down on AC use during certain hours? That wasn’t to cut costs - it was because your grid was so maxed out on certain days that it was approaching rolling blackouts.
Finally, heavy industry - like fabrication, mining, manufacturing, etc - has been making a positive push in recent years to electrify their operations and heavy machinery. While reducing the sector’s dependency on fossil fuels and their associated carbon emissions, this move also requires that enormous facilities now be connected to the grid, further increasing energy demand on an already strained system.
Something’s gotta give, right?
Short answer - yes. We need a new approach, a new way of looking at energy, if we are to reconcile these forces of energy supply and demand pulling in opposite directions. The AI and climate change trains have left the station - already making it glaringly obvious that the old system and solutions are broken, that simply throwing money at this challenge will fail.
In the coming weeks, we’ll dive deeper into the major obstacles that must be addressed by a new approach to energy:
Rethinking AI: investing in reuse and efficiency versus raw energy production, and being realistic about what’s possible
Future-proofing the grid for climate change: enabling resilience to rising temperatures and extreme weather events
Electrifying industry and transport: preserving grid stability while phasing in low carbon roadmaps for the most energy-intensive sectors
2026 is one of only a few foundational years that will dictate our success in this space - requiring tact, precision and recalibration of our expectations. This key stage of the energy transition will test the limits of what we are capable of as nations and a global society, as well our ability to be realistic about our limits.