I was interviewed by a journalist a few days ago about the comparison between the electrical revolution between 1890-1930 with the adoption of digital technologies. This is complicated and I literally could write another book, but here are my reflections on this question.
Fireless Fireworks, Aurora Effect
Hudson-Fulton Celebration, New York City, 1909
for more information, see
David E. Nye, American Technological Sublime, 153-165
The electrical system was deployed by equipment manufacturers beginning in the 1880s. Almost immediately came horizontal integration as many firms merged. By the middle 1890s a duopoly of GE and Westinghouse dominated the market, and they pooled their patents in 1895. Both sold equipment to local utilities, which were gradually linked into regional systems, and later joined large scale power-sharing agreements. To some extent GE and Westinghouse owned stock in these local utilities, but they did not come to dominate that part of the industry. Rather the equipment manufacturers early became international, while the utilities were local and later regional monopolies. There was also a third tier of the industry, consisting of many small companies whose products plugged into the electrical system. There was some vertical integration between these fairly independent layers, but it was regulated by state utility boards and to some degree by the federal government. At the time, it was thought that many utilities, including those that sold gas, electricity, water, and other products delivered by networks, were "natural monopolies." It seemed to make no economic sense to build more than one set of gas lines or electrical transmission networks.
This situation only roughly corresponds with the division today where (1) some corporations produce computer chips and other vital components, (2) those who assemble components into products such as laptops, mobile phones and tablet, (3) those focused on software development as Microsoft did in its formative years, and (4) those making making printers or other peripherals (they are roughly equivalent to those who in 1925 made electrical clocks, radios, appliances, etc.)
How do these two systems differ?
(1) Today, there is less government regulation, which also declined as a part of the electrical system after c. 1980. The old utilities were governed by state boards that oversaw pricing of electricity and had some influence on the structure of competition.
(2) There is less horizontal integration and more vertical integration. Apple is a prime example. It designs its own chips, produces hardware, and the software for its operating system, and also chooses which software products can be purchased through its online store. Google takes a somewhat different approach, as it does not make computers, but it has created a mobile phone standard, etc. Microsoft at first focused on software but later realized it might best sell computers and tablets. There is some interoperability, notably in email, but to a considerable degree each corporation attempts to create a suite of software and services that can capture and hold the attention of users. Facebook is attempting to market a digital world of its own, the Metaverse, which is intended to be a 3D virtual world where users will work, socialize and be entertained. Google has an expanding range of apps designed to serve every need, whether to work, play, or explore. Note that each corporation seeks to become all encompassing, and each one is independent. The consumer can leave one and go to another, but it is not possible to integrate them. The horizonal integration is quite limited.
(3) In this situation, there are no equivalents of local utilities subject to state regulation. Instead, corporations compete with one another to create a suite of useful products that together define a universe or system that supplies an ever-expanding range of services. There was only one local utility for each city or town in 1920, as this was thought to be inherent in the technology - natural monopolies seemed the only sensible option. Today's tech companies all seek to serve every market, and given weak regulation the profits being made are enormous. Look at the largest corporations making up the DOW Jones industrial average in c. 1925, and one finds a list of companies that emerged after 1880 to dominate the economy, many of them energy related, such as GE, Westinghouse, GM, Ford, and many oil companies. In 2020 these have been dethroned from the top of the list. As late as 1990, the top five companies were General Motors, Ford, Exxon, IBM, and General Electric. In 2020 only Exxon remained among the top five, which was made possible by merging with Mobil Oil. Otherwise, the top five were Wal-Mart, Amazon, Apple. and CVS Health.
(4) Electrification interfaced with all aspects of society, as every part of society eventually was included in the network. The electrical system of power generation, transmission, and installation involved wiring up millions of buildings and building the infrastructure needed to supply them with power. It took more more than fifty years to establish. A great deal of standardization was built into the electrical system. This created ideal preconditions for the rapid adoption of digital technologies.
(5) The electrical system developed slowly, reaching various markets in a sequence of expansions that took decades, as described in Electrifying America (MIT Press 1990). The digital transformation began slowly after World War Two, but it is now stampeding ahead, reaching into almost all sectors of the economy simultaneously. The pace of change is far faster than the design of legislation to control it. Governments struggle to keep up. For example, during the development of AI there has been a massive violation of the laws of copyright, and the government and the courts are struggling with a wave of lawsuits. There are many other examples, of course, as each of the separate corporate universes seeks to maximize profits.
(6) At first this digital transformation only took a small amount of the energy being produced, to some extent offset by efficiency improvements in lightbulbs, motors, etc. But now the demands of AI are sucking up vast amounts of energy, and there is an emerging crisis of supply. The electrical system, which grew extremely rapidly between 1900 and the 1990s, after that had begun to reach a stasis, in which per capita demand was falling due to improved efficiencies, roughly counterbalanced by population growth. But in 2025 demand is surging, and the systems are close to being overtaxed. Increasing the generation and transmission systems is a slow business, and therefore a crunch is coming.
(7) One final difference has developed from the confluence of the differences discussed above. The electrical system, given its design, did not lend itself to gathering extensive information about consumers. Each utility had the names and addresses of its customers, and it could keep track of how much electricity they used. But they had no way to track the consumer choices among various electrical products, which they might buy anywhere. Only in the first years did utilities themselves sell light bulbs or appliances, and soon this aspect of the business was handled by independent retailers. The digital system, in stunning contrast, has used its vertical integration to assemble vast, detailed databases on every consumer within its orbit. In effect, the consumer has become a product for sale in a new attention economy, as this database is used to design and pinpoint advertising messages.
If one looks ahead, there are many unknowns. How much will AI grow? Will it prove to be overhyped? How long will solar cells and wind turbines continue to become more efficient? Where is the limit in their development? Steam turbines kept getting more efficient (and larger) for 75 years or more. If alternative energies do the same, then the average homeowner or business building could become a mini-utility, supplying its own needs, charging automobiles, and selling significant amounts of power to the grid. But the problem now is that government is playing catch up, and some inefficiencies and disincentives may be institutionalized before suitable controls are created.
The big question is whether the corporations that now dominate the digital world will be allowed to continue in their present form, or whether like Rockefeller's Standard Oil, like IBM, and like AT&T, they will be reformed and to some degree broken up by legislation. A century ago the Progressives saw the need for such change. The European Union now grasps the need for it. But the current American government under Trump accepts million dollar gifts from the largest digital companies and encourages an emergent Silicon Valley authoritarianism. The once popular idea that digitization and the Internet would lead to greater democracy, mutual understanding, and increased freedom has hardly been realized. For further reading, I recommend the excellent book by Shoshana Zuboff, of the Harvard Business School: The Age of Surveillance Capitalism (2019).
