The cobweb model explains why the AI data-centre boom keeps running short of electricians

The cobweb model describes how the AI data centre boom continues to face a shortage of electricians. Currently, AI data centres in the United States are facing a severe shortage of electricians as demand continues to outstrip supply. The $6.7 trillion global data centre investment that McKinsey estimates by 2030 will not be adequately served, given the large number of electricians that will need to come out of the training programmes within the next 10 to 15 years to meet demand.

The cobweb model describes a market in which producers use the price from the previous year's market to determine their current year's production. The lag time between receiving the price signal for a good and being able to supply that product creates cyclical fluctuations in both price and quantity around an equilibrium price.

The cobweb model is so-named because of the spiral path traced on the supply and demand diagram generated by the model. The cobweb model has been used to explain the price fluctuations of crops and livestock.

Electricians are also a good fit for this model because training takes years, so wages and the supply of electricians stay out of step, and high wages pull more people into training.

Electrician Demand is Derived from the AI Market, not the Electricians' Trade.

The demand for electricians in today's market is an excellent case of derived demand, as consumers of electricity do not want electricians for their own sake, but rather to provide the electrical power necessary to operate their computers, cooling systems, and so forth. The demand for construction and labourers needed to build hyperscale data centres is overwhelming the available workforce, and the premium for a data-centre electrician is much higher than for a regular electrician. Some data centre electricians in Northern Virginia and Texas can earn over $260,000 per year, approximately one-third higher than in other construction trades, and there will be an estimated shortfall of 130,000 electricians in the USA by 2030, according to McKinsey.

The lagging training cycles mean that the available supply of electricians cannot respond to this significant wage signal for several years into the future.

This is the impact of the cobweb phenomenon. The licensing period for electricians consists of four to five years of apprenticeship. Therefore, the supply elasticity for electricians in the short term is nearly zero. Even if wages doubled, there will be no qualified electricians available for several years into the future — the early 2030s. According to the Bureau of Labor Statistics in the USA, approximately 81,000 new electrician openings each year through 2034, with about 20,000 a year replacing electricians retiring. Data center staffing companies exist to deal with this issue of moving a limited number of certified electricians around to the highest bidder. However, the movement of the certified electricians between the data centres and the highest bidders eases the immediate shortage of electricians on one specific site from the data centre. But staffing companies move only a limited number of available electricians, so the overall national shortage of electricians and growing wage pressure continues until new electricians are trained.

High wages in a booming economy create the seeds of future over-production of electricians

The sting of the cobweb phenomenon occurs near the end of the cycle. The current wages of electricians are high enough to attract large numbers of applicants into the electrician apprenticeship programme, as well as larger training programmes for electricians. Therefore, the large number of new recruits in the electricians' programme will commit years into the future to develop their human capital based on the current wage differential between electricians and others. However, they will all enter the programme about the same time, and they will all graduate and become qualified electricians around the same time in the early 2030s. This coinciding graduation of many new electricians has the potential to create an excess of electricians just as data-centre demand growth slows, thereby resulting in lower wages in the data centre sector, once again becoming unattractive to potential new recruits. The cobweb phenomenon can be traced back through historical data in housing markets and hog markets and how the inelastic short-term supply has created cycles of adverse effects in these industries.

The width and duration of the cobweb cycles will be based on how responsive supply is to demand. In short, the greater the responsiveness of the supply to the demand, the smaller the width of the cycles. The current data-centre boom is an almost textbook example of a cobweb; an exceptionally large verifiable demand for electricians and an insufficiently responsive supply of trained electricians to eliminate the oscillation in the costs of hiring electricians.