So where will this market evolution take us? An immediate turnaround in fortunes for the oversupplied and largely loss-making Chinese solar industry seems unlikely, while battery manufacturers are under similar pressure. Those that survive, perhaps consolidating smaller weaker players along the way, have the prize of a rapidly expanding global market.  

In this scale-up phase, the parallel revolution in battery storage will be a key catalyst for the next wave of solar growth. In more mature solar markets, battery flexibility ensures higher utilisation of solar capacity and improves its financial returns. In emerging solar markets, solar and batteries are now increasingly co-located from the outset. That in turn underpins the continued rapid growth in demand for batteries, as they get more efficient, cheaper and more critical for wider renewable growth.

This is how we get a mutually beneficial virtuous circle, where batteries unlock solar upside and vice versa. The solar and battery revolutions become self-reinforcing, setting the new template for how the world will largely be powered in the future. This comes mainly at the expense of coal and gas, depending on the market in question (each country has a natural inclination to maximise its own natural resource endowment, as China is doing across the whole energy spectrum). 

Seeing beyond volatility 

So, having understood these macro power dynamics, the question remains: are clean energy infrastructure and technology companies currently at risk from an AI-driven financial market bubble? Absolutely. Especially in the US, many have been swept upwards in the AI excitement. But the bigger question is: are they also exposed to a bigger supercycle in global electricity demand? Also absolutely. Both things can be simultaneously true. As the next decade to 2035 unfolds, electrification (of industry, transport and heating) and improving living standards create a rising tide for the clean energy infrastructure supply chain: from batteries and solar PV to grids and electrical hardware to EVs and electrified heating equipment. 

The corollary, then, is to accept the fear of a crash, while continuing to invest steadily for the longer-term. Famous last words? Maybe. Popping of an AI bubble – assuming that is what it is – could well be painful, but for now, remaining invested specifically in the electrification supercycle - not frothy AI tech stocks - still feels like a rational choice, versus being paralysed by ubiquitous dire warnings and attempting to time the market around an unpredictable crash. 

It’s also possible a crash may not arrive for much longer than many expect, or may disproportionately impact tech rather than energy companies. It may turn out that we are only in the very early stages of the AI revolution, evolving into new forms of intelligence, including humanoid robots, fully autonomous vehicles and the rest. No one really knows. And even if it does happen, it may ultimately prove a useful opportunity to get cheaper exposure to the companies that will thrive in the Age of Electricity. Either way, the second half of this decade will undoubtedly be volatile and noisy, as long-held assumptions in the energy world come under increasing challenge. 

Positive feedback loops 

Even with a bigger clean energy investment boom, there’s obviously no guarantee that those companies will outperform wider markets over the next decade. But we do seem to be rapidly approaching or probably already at some key tipping points that provide strong tailwinds for those companies. Examples include: 

• Solar and battery-firmed solar power becoming the default choice for new generation capacity in most of the world 

• Growth in renewable output exceeding electricity demand growth  

• EVs taking majority shares of new car sales in key markets, not least China 

These tipping points are consistent with the expectation of broader, deeper, more diversified clean energy progress over the decade ahead. Falling costs will bring further such positive feedback loops, bringing a surge of renewable and battery installations across different markets, including those perceived as petrostates like the US, Saudi Arabia and the UAE.  

This rising global wave of demand for solar, wind and battery hardware along with all the electrical infrastructure and equipment needed to hook it all up and into people’s lives, is the real context for the recent clean energy equity recovery. Without doubt, some of those supply chains remain in heavy oversupply today, leaving many manufacturers in unsustainable loss-making positions. But the winners of that Darwinian struggle have the potential to win very big indeed, while the installers and users of their products will continue to reap the benefits in the meantime. 

Efficiency and Rebound Effects

One important note of caution is that previous electricity demand growth forecasts have often overshot. Especially when it comes to data centres, let’s not forget about energy efficiency. AI chip computing power is on a rapid efficiency curve: new models already use 99% less energy than in 2008. This extraordinary rate of improvement will undoubtedly continue, even at a slower pace as chip technology matures. 

Jean-Pascal Tricoire, Chairman of Schneider Electric, claims that the energy efficiency unleashed by AI will ultimately outweigh its additional energy demand by up to 4-9 times, thanks to advances it will bring in predictive maintenance, smart charging, demand response, real-time analytics and autonomous operations. In other words, instead of devouring incremental supply, AI will be a grid enhancing technology (more on them soon). This is consistent with past technology booms, notably the internet’s emergence when efficiencies defused similar initial fears about excessive demand growth. 

But even then, the rebound effect means greater efficiencies reduce barriers to adoption, increasing total demand. So even if AI ultimately ends up helping to save energy, this will likely be outweighed by the broader growth in demand for AI services, meaning the overall energy use driven by AI will still rise.