AI Disruption: assumptions on timing and origin thwart

Analysis
The cascade effect of unintended consequences in markets and where the ripples are eventually felt in the lives of ordinary people continues to fascinate. The younger capitalists close to me whose prognosis of the price of energy inputs based on their long-term view that renewables and conventional power infrastructure are unlikely to meet the future requirements of a vibrant tech-rich economy and the energy markets have not had a good year. The year-to-date drop in the spot price of uranium has been around 36%, and this calendar year so far has not been pretty either. A third of the drop has occurred in only the last week as I write; 5% on Monday night 27th January alone: the man’s not happy but he is showing his “mettle” by sticking to it – it’s a long-term game after all, he says. The traders, however, need to make their money now and the necessary short-termism drives the volatility: the Term Price, however, offers another story vindicating those taking a long-term position in the face of investment market disruptions.

The rationale for investing in the uranium market for this n=1 was on the back of the shift to non-emissions producing resources for electricity production, that is, nuclear power. Economic activity is energy-based – it was always thus. Most of human existence we used energy from animal and human labour (maybe with some water-, wind- and fire-power) when it was only efficient to produce one calorie if the people-power in the equation did not expend more than that one calorie in doing so (the true definition of sustainability!).

The extraction of cheap energy that was sequestered over millennia and consumed by the industrialised world has allowed us to use 10 to 15 calories of energy to produce every calorie of food, and we waste far more on trinkets, fast-fashion and low-quality shiny landfill. To enable the continuation of the degree of energy use that in the past would have been defined as ruinous, but now to do so without the pollution and climate damaging costs of fossil fuels, has required some concerted rejigging of our energy markets, thus the bets on uranium.

Another explanation for the increased demand is increased electricity use in transport, communications technology, manufacturing and residential heating/cooling (aggravated by climate change) and other home-based uses. Global electricity demand is expected to increase by an average of 3.4% annually to 2026 according to the International Energy Agency, much of it supported by an improving economic outlook and the drive towards electrification in advanced economies and China to meet decarbonisation targets. To maintain targets, power production needs to shift from 18% share of electricity in 2015 globally to 30% in 2030. Not only does the share of the energy pie increase but the pie itself is growing: electricity consumption in data centres, AI and the cryptocurrency sector is set to double by 2026 – for visualisation purposes, this is similar to the entire electricity consumption of Japan.

China’s manufacturing centres are critical to supply the world with the necessary hardware and technology for the electrification shift (see graph for production growth). China’s electricity demand in 2023 rose by 6.4% and per capita consumption continues to be higher than Europe. Regardless of a slowing economy, high per capita use is forecast to continue, driven by China’s investment in and position as the leading global supplier of clean energy technologies, including electric vehicles. The supply of these products into the local Chinese market will further push up electricity consumption from renewable sources here.

This is the context within which the price of uranium shifted overnight on the 27th January. A shift in any element in the broader context has potential to disrupt forecasts. What happened at the end of January made the market take a very close look at that specific assumption in the energy supply and demand model.

The Chinese AI firm Deepseek, the developer of the R1 model released last month, shook the energy market for a while, more than rattling just the AI industry. The issue was not that efficiencies over time were not expected in AI, it was the timing and the origin of the development that caught the global market off-guard. Deepseek’s lean use of energy and integrated circuits has provided evidence that Chinese technologies no longer rely on borrowed knowhow and copied hardware. This success points to impressive strengths in development expertise. Necessity is the mother of invention, and it was resource constraints in the global landscape combined with Chinese government direction, investment and support for AI research and innovation which have created the right circumstances (note graph for growth in AI investment).

China’s growth in production output in renewables manufacturing. Source: IEA

Power crises and U.S. chip export restrictions among others presented environmental threats, and government investment in chip technology and AI education offered the opportunities. In hindsight, a good ol’ SWOT analysis points towards a potential for this occurring. The plethora of other AI models in China lends credence to the idea that Deepseek’s launch on the global stage is not an aberration.

The AI roll-out will continue to disrupt and not just the energy sector or the semi-conductor supply market. New and existing forms of AI will disrupt the labour market, whether it becomes a substituting technology, increasing unemployment and further exacerbating inequalities (feelings of shame and being unvalued, and placing power and control in the hands of a few), or a complementing technology which enhances productivity and assists people in the quality of their work or conversely risks the de-skilling of the workforce as people rely on AI as a crutch. As an example of ethical concerns, academic publishing is being challenged by a form of plagiarism on steroids and one that is far harder to detect with the current technologies. Journal publishing companies are grappling with a spectre of unknown quantity and nature that cuts to the heart of knowledge creation and the credibility of experts in their field.

The complexities and challenges of AI will be felt in other spheres, including political and cultural ones. MDT reported last week that China has put boundaries around AI for the safety of the community, privacy and ethical concerns. Regulating and being cautious about AI’s influence is imperative so that it remains a tool under our control to help us progress as a society and not become a disruptive vector that gives others control over how we think, what we believe and who we become.  By Leanda Lee, MDT