 |
| A technician works at an Amazon Web Services AI data center in New Carlisle, Indiana, U.S. REUTERS/Noah Berger for AWS |
By Nektarios Michail
The first industrial revolution was the period between 1760 and 1840, which was defined by more efficient and stable manufacturing processes, at a large scale, aiming to reduce labour costs. The transition, led by Great Britain around 1760 and reaching the United States by 1840, ended up improving the standards of living for the Western world. As expected, industry captured most of the total production, with the share of services in GDP for the United States (US) and the United Kingdom (UK) stood at approximately 40 per cent in 1840.
The second industrial revolution (late 19th and early 20th century) was characterized by rapid scientific discovery, standardization and mass production, as well as the wide use of the telegraph and railroads. This led to less human input in traditional industries, while electrification removed the darkness constraint and production shifted also towards the services sector. By 1910, the share of services in GDP for the US and the UK had increased to around 51 per cent.
The third industrial revolution was more of a misnomer: the information age, which started in the mid-20th century was linked to the development of the transistor and, while computing hardware was key to its development, most of the production came from the associated software and the services it created. As a result of the rapid development of such sectors, the share of services in the US rose to almost 70 per cent by 1980. After the internet revolution, the services share of GDP rose further to around 80 per cent for the US and around 75 per cent for the UK by 2016.
The fourth? World Economic Forum founder, Klaus Schwab, suggested in 2016 that one had started, noting that advancements including artificial intelligence, gene editing and advanced robotics, represent a significant shift in industrial capitalism. Others have noted that the fourth industrial revolution marks the beginning of the imagination age.
Interestingly, the fourth industrial revolution bears more similarities to the first rather than the others. While the share of services has increased over time, as capital replacement needs were lower due to mass production, standardization, and software developments, the fourth industrial revolution requires material capital investments to support the services structures that are requested.
An example is the energy requirement of AI, which amounts to 400 terawatt-hours annually, or roughly 1.5 per cent of total world electricity; projections by the IEA suggest that it could more than double by 2030. This could threaten grid stability and increase carbon emissions, as well as reduce hardware lifecycles and increase water consumption for the cooling of computer chips. The only way this can be resolved is via increased spending on infrastructure which, as per recent estimates, could amount to $7 trillion (around 30 per cent of US GDP), for data centres only.
This huge infrastructure spending is unparalleled in recent decades. Barring the reconstruction of Europe after the second world war, nothing of this size has ever taken place since. The large wave of investment in computing power, data centres, electricity implies a material shift in investment patterns. In stark contrast to previous decades, even software companies see higher capital spending; Alphabet, Amazon, Microsoft, and Meta are planning on spending $725 billion on capital expenditures, up 77 per cent from a year ago.
Higher capital spending is also likely to have broader consequences. Unlike services spending, which pretty much boils down to wages, capital goods require machinery, raw materials, other goods, fuel, and so on. As demand for these increases, an economy-wide increase in prices is more likely. The reason is that while services spending goes to individuals who already earn high wages (programmers, developers, etc.) and have little incentive increase spending, such capital expenditure affects a much broader spectrum of the economy, allowing for the end-beneficiary, the worker at an industry that would have otherwise not profited, to reap the earnings from this surge in activity. People in such industries usually have a higher propensity to consume, thus are more likely to increase spending and push prices up.
As the world appears to be moving away from the services-based model that dominated the post-WWII era and into an increased role for capital, it is far more likely that growth will be inflationary. This has wider implications for the overall economy, as bond yields are likely to reprice materially: even at 2 per cent policy rates, an expected inflation of 3 per cent implies at yield of, at best, 5 per cent. Higher yields threat to the viability of public debt, especially in overindebted countries. Ironically, higher private sector capital spending allows governments to proceed with slow and steady spending cuts without hurting growth.
It remains to be seen how permanent this increase in capital expenditure will be and precisely how much of that will be spread out over the coming years. If it continues, the era of low inflation will have passed irretrievably.
Nektarios Michail, Chief Economist, Bank of Cyprus. Views and opinions expressed are personal. The article is republished from the blog of the Cyprus Economic Society
