The People Swept Away by Electricity

We remember the night Edison lit up Manhattan, but we forget the companies, the failed technologies, and the real lives lost on the road to that night. History is written by the victors, but the lessons are hidden in the losses of the losers.

Today, everyone talks about AI and compute, just as everyone talked about electricity in the 1890s. When a general-purpose technology transitions from the lab to the home, its adoption is never a straight line but a battlefield littered with the remains of failed attempts. This article systematically examines the failures, deaths, and missed technological paths in the spread of electricity—not for spectacle, but because we are standing at an uncannily similar historical juncture.

One: The Death of Technology Routes: Winning in the Lab, Losing in the System

1. The Demise of Direct Current: Edison’s Biggest Mistake

In 1882, Edison built the world’s first commercial central power station on Manhattan’s Pearl Street, using direct current (DC). DC had a fatal flaw: it was economically impossible to transform voltages at the time, leading to significant power loss over short distances, with an effective supply radius of about one mile. Following Edison’s approach, electrifying a city would require building a power station every one or two kilometers, which was economically unfeasible.

In 1886, William Stanley demonstrated transformer-based alternating current (AC) transmission in Great Barrington, Massachusetts. In 1888, Tesla’s polyphase AC motor patents were acquired by Westinghouse. AC could transmit high voltage over long distances and then step down to household voltage, making it economically superior to DC.

Edison’s response was not to pivot but to smear. He funded public demonstrations in which stray dogs and horses were electrocuted to dramatize the dangers of AC. He even pushed New York State to build its first electric chair with a Westinghouse AC generator, attempting to make “Westinghoused” synonymous with execution. In August 1890, murderer William Kemmler became the first person sent to the electric chair. The execution went terribly wrong: the first current failed to kill him, and the second lasted well over a minute, producing a gruesome scene. Westinghouse later commented, “An axe would have been more humane.”

The smear campaign failed to save DC. In 1893, the Chicago World’s Fair lighting contract was won by Westinghouse with an AC solution. In 1895, Niagara Falls’ hydroelectric power was transmitted to Buffalo, 32 kilometers away, using AC. The war ended.

The most profound aspect of this case is that Edison lost not in technical capability but in systemic judgment, and he attempted to use public relations instead of a technical pivot. He was so deeply invested in his personal reputation, existing assets (existing DC power stations), and his route that he could not pivot.

2. The Irony Continues: The Century-Old Ghost of Direct Current and Its Return

History added two footnotes to this story:

New York’s DC power supply was not completely terminated until November 2007. The DC network laid by Edison in 1882, with its remaining users (mainly old elevators), kept the DC service alive for 125 years. The “stranded assets” of a failed route can drag a system for over a century.
A century later, DC won back ultra-long-distance transmission. Once modern high-voltage direct current (HVDC) technology matured, it proved superior to AC for undersea and ultra-long-distance transmission. Swiss engineer René Thury built the Moutiers–Lyon high-voltage DC transmission line as early as 1906; it operated for thirty years but was abandoned at the time because the equipment was too complex. Whether a technological path is “right” or “wrong” carries a timestamp: even the right direction can die when the supporting technologies are not yet mature.

3. The Frequency War: A Historical Wound That Continues to Bleed

After AC defeated DC, chaos did not end. Early AC systems used a variety of frequencies: 133⅓ Hz, 60 Hz, 50 Hz, 40 Hz, and 25 Hz were all used.

Niagara Falls’ hydroelectric power station chose 25 Hz (favoring the motor design at the time), but at 25 Hz, lights flickered visibly and were incompatible with the later 60 Hz mainstream. The 25 Hz legacy system in North America took over half a century to phase out.
Southern California Edison initially chose 50 Hz and, around 1948, was forced to convert its entire system to 60 Hz—meaning nearly every electric clock, motor, and appliance in its service area had to be replaced or modified, at astronomical cost.
The most painful case was Japan: in 1895, Tokyo Electric Light bought a 50 Hz generator from Germany’s AEG, and in 1896, Osaka bought a 60 Hz generator from America’s GE. Two independent procurement decisions left Japan with different grid frequencies in its eastern and western halves to this day. After the 2011 Fukushima nuclear accident, eastern Japan faced a power shortage while western Japan had ample electricity but could not provide large-scale support, because the frequency-converter capacity between the two grids was extremely limited. Two nineteenth-century purchasing agents constrained disaster relief in the twenty-first century.

Standards lock-in is the costliest trap in the adoption of general-purpose technologies: choices that seem trivial early on can accumulate into a century of debt.

4. Arc Lamps and “Moonlight Towers”: Winning the First Round, Losing the Whole Game

Before the incandescent light bulb, the pioneers of electrical lighting were arc lamps. Charles Brush’s arc lamp system lit up Cleveland and San Francisco in the late 1870s, and Brush Electric was an industry powerhouse. At the time, a “moonlight tower” scheme was popular: tall towers in cities with super-bright arc lamps simulating moonlight to illuminate entire districts — Detroit, San Jose, and other cities deployed these on a large scale.

But arc lamps were too bright, too glaring, noisy, and required frequent carbon-rod replacements. They suited streets and factories but could not enter the home. Incandescent lamps, though dimmer individually, could enter living rooms everywhere. The arc lamp won the first round by serving the existing demand called “lighting”; the incandescent lamp created the new paradigm of “electricity in every home.” Brush Electric was acquired by Thomson-Houston in 1891, and moonlight towers were dismantled city by city—today only Austin retains more than ten towers dating from 1895, monuments to a path not taken. Tesla’s original two-phase AC design for Niagara was ultimately displaced by three-phase AC as well: even within the winning camp, specific technological paths continued to be eliminated.

Two: The Death of Business: Winning in Technology Does Not Equal Winning in Business

5. Edison and Westinghouse: Both Winners Were Exiled by Capital

The two protagonists of the War of Currents were both pushed out of their own companies by capital.

In 1892, J.P. Morgan orchestrated the merger of Edison’s General Electric with Thomson-Houston, removing the name “Edison” from the company and stripping Edison of control, after which he largely exited the power industry.
Westinghouse won the “War of Currents,” but lost control of the company in the 1907 financial panic due to excessive company debt, and was eventually pushed out of the company he founded, dying in 1914.

The fruits of the battle for general-purpose technology often do not belong to the winning engineer-entrepreneur, but to financial forces that can integrate capital, patents, and scale.

6. The Collapse of the Insull Empire: The Man Who Took the Financialization of Electricity to Its Limit

Samuel Insull, once Edison’s private secretary, later took charge of Chicago’s Commonwealth Edison. He was a true architect of bringing electricity into ordinary homes: he pioneered time-of-use pricing, large-scale grid interconnection, and a business model for extending power networks into rural areas. But he also financed electric expansion through a pyramid of nested holding companies, stacking leverage upon leverage.

The Great Depression struck in 1929, and by 1932 the Insull empire had collapsed, wiping out roughly 600,000 small shareholders and destroying hundreds of millions of dollars in value at the time. Many were ordinary people who had invested their life savings in supposedly indestructible utility shares. Insull fled to Greece and was extradited to the United States for trial. Although ultimately acquitted, he died of a heart attack in a Paris Métro station in 1938 with only a few francs in his pocket. His collapse directly gave rise to the Public Utility Holding Company Act of 1935, reshaping U.S. utility regulation.

The widespread adoption of infrastructure requires massive capital, which inevitably leads to financial innovation. Financial innovation can be a double-edged sword, biting back the entire industry during downturns—the story of “shovels” also has a bubble side.

7. The Lesson from Britain: Over-Regulation Can Suffocate a Nation’s Electrification

In 1882, Britain passed the Electric Lighting Act, giving local governments the right to compulsorily purchase private electricity companies after 21 years at their “residual book value,” with no allowance for goodwill or future earnings. Private capital immediately lost interest in electricity investment. The act had to be revised in 1888, extending the period to 42 years, but the damage was done. Britain went from leader to laggard in electrification: by 1913, London alone had 65 supply authorities operating dozens of different voltage and frequency systems; a family that moved house might have to replace every appliance. Meanwhile, Germany and the United States had already formed large, unified grids.

The timing of regulation is more deadly than its direction: imposing property uncertainty before the technological paradigm has converged can make an entire country fall out of the game for a generation.

8. The Counter-Attack of the Gas Industry: The Last Stand of the Disrupted

A fact often overlooked: after the invention of the incandescent light bulb, gas lighting did not immediately die but experienced a resurgence. In 1885, the Welsbach mantle increased the efficiency of gas lamps by several times, with costs far lower than electric lights, extending the life of gas lighting by two to three decades, only to be completely outmatched by the decline in electric light prices in the 1910s.

Under pressure to be disrupted, in-position technologies often exhibit remarkable improvement capabilities—faster carriages will indeed be made, and they can win for a while. This is crucial for timing predictions: disruption is almost always slower than believers expect but more thorough than skeptics expect.

9. Electric Vehicles in 1900: A Bet a Century Early

Few know that around 1900, electric vehicles made up about a third of the cars on American streets, with steam, electric, and internal-combustion vehicles dividing the market three ways. Electric vehicles were quieter and cleaner, required no hand crank to start, and were seen as the “respectable person’s choice.” Capital responded: the Electric Vehicle Company, known as the “lead-battery taxicab trust,” raised enormous sums to deploy electric taxi fleets in New York and elsewhere, and for a time became America’s largest automobile manufacturer.

The outcome: battery energy density could not support the range, charging networks did not exist, and the discovery of oil made gasoline extremely cheap. EVC went bankrupt in 1907. Edison also failed here—he invested about a decade developing nickel-iron batteries, which were quite successful, but the electric vehicle market he aimed for was already flattened by the Model T. The electric vehicle route slumbered for a full century, only to be revived with the maturation of lithium-ion batteries and another set of infrastructure.

Correct direction + wrong timing = failure. And it’s the most heartbreaking failure: pioneers’ tombstones are inscribed with the business plans of later winners.

Three: The Real Death: The Cost of Widespread Adoption Measured in Lives

10. The Body Suspended Over Manhattan

In the 1880s, New York’s skies were a tangle of wires: telegraph lines, telephone lines, high-voltage arc lights all tangled on the same poles, with insulation deteriorating in the rain and wind, unregulated. The deaths of line workers were commonplace.

On October 11, 1889, John Feeks, a Western Union telegraph line worker, died in Manhattan’s busy district when he touched a live high-voltage line, his body entangled by the wires, burning for nearly an hour in front of thousands of lunchtime passersby, his mouth and nose spewing sparks and smoke. This scene made the front pages of all newspapers, sparking public outrage, directly leading to the mandatory undergrounding of wires in New York and the establishment of systematic electrical safety regulations.

11. Fires, Insurance Companies, and the Birth of “Safety” as an Industry

Early indoor wiring had no standards, and electrical fires were frequent; insurance companies came to treat electrified buildings as high-risk properties. The electrical hazards at the 1893 Chicago World’s Fair prompted the insurance industry to station a young electrical engineer, William Merrill, on site as an inspector. He went on to found Underwriters Laboratories (UL) in 1894. The first edition of the National Electrical Code (NEC) appeared in 1897.

Note the sequence: major accidents first, then safety standards and the industry of certification. Safety is never an inherent attribute of new technologies but a public good “bought” with accidents and lives. Today, the safety assessment, alignment, and certification industries for AI are reenacting the birth of UL.

Four: Bubbles, Frauds, and Fantasies: Every New Technology Comes with a Bubble

12. The Myth of Electrotherapy: Turning Incomprehension into IQ Tax

The public’s lack of understanding of electricity often created opportunities for exploitation. In the late 19th and early 20th centuries, electricity was sold as a panacea, with electric belts claiming to cure impotence and “neurasthenia,” Dr. Scott’s Electric Hair Brush promising to treat baldness and headaches, and electric corsets, chairs, and violet-ray machines flooding department stores. In the 1920s, Los Angeles real-estate developer Gaylord Wilshire—the namesake of Wilshire Boulevard—lost his fortune promoting the I-ON-A-CO electromagnetic collar, which claimed to cure myriad ailments by wrapping an energized coil around the neck, and won over tens of thousands of believers.

Even outright financial frauds abounded: the Electric Sugar Refining Company in the 1880s claimed to have discovered a secret process for refining sugar using electricity, raising nearly a million dollars by financing a “mysterious workshop” that was off-limits to investors. When the fraud was exposed in 1889, the workshop was found to contain only a bunch of fake devices—akin to the “demonstration room fraud” of the era.

Every technology that has been deified has been used as a tool for exploitation. The scale of the fraud is proportional to the public’s reverence for the technology, and inversely proportional to their understanding of it.

13. The Telharmonium: “Cloud Music” Half a Century Ahead of Its Time

Inventor Thaddeus Cahill’s Telharmonium was perhaps one of the electric age’s most tragic fantasies: an electronic synthesizer weighing around 200 tons, installed in the basement of New York’s Telharmonic Hall in 1906. It streamed music in real time over telephone lines to subscribing restaurants and hotels—a complete prototype of subscription streaming, a century before Spotify.

It failed quickly: the machine was enormous and expensive, and its signal bled badly into telephone circuits—people would hear music pouring in while making a call. The telephone company refused to cooperate, the venture went bankrupt within a few years, and all three generations of the machine were sold as scrap without leaving a single recording. The business model was entirely right in its imagination, but the infrastructure needed to carry it—bandwidth, terminals, and amplifiers—would take another fifty years to arrive.

14. Wardenclyffe Tower: Tesla’s End

The technological genius Tesla, who won the “War of Currents,” met his own end in a subsequent gamble. Starting in 1901, he built Wardenclyffe Tower on Long Island, dreaming of global wireless power transmission and free energy. Morgan withdrew his investment halfway through—a widely repeated account has him asking, “If electricity can be transmitted wirelessly and for free, where do I put my meter?” The project was abandoned, and the tower was demolished in 1917 to pay debts. Tesla spent his later years living on assistance in a room at the New Yorker Hotel and died alone and deeply in debt in 1943.

Getting one thing right in a technological revolution does not guarantee survival in the next bet; and grand narratives that do not answer “where do I put my meter?” cannot secure funding, regardless of the beauty of the technology.

Five: The Most Concealed Failure: Buying Electricity, But No Productivity Gains

The final case lacks a dramatic death but may be the most significant.

Economic historian Paul David noted in his famous essay The Dynamo and the Computer that the central power station was built in 1882, yet as late as 1900 electricity still accounted for less than 5 percent of mechanical power in American factories. The significant productivity surge did not arrive until the 1920s—a full forty-year lag.

The reason, as mentioned earlier, is that the first generation of electrified factories simply replaced steam engines with large motors, continuing to use central drive shafts and belts to power all machines (so-called “central drive”). Factory layouts, work processes, and management methods remained unchanged, with electricity merely being more expensive steam. The real benefits came from “individual drive”: after each machine was equipped with an independent motor, factories could be reorganized according to their production processes, factories could be lit and ventilated, and assembly lines became possible—this required rebuilding the factory, rewriting the process, and retraining an entire generation of workers and managers. Before the old factory was fully depreciated, no one was willing to do this.

In other words, for a generation, most companies that paid for electricity were “losers” in this revolution—they bore the cost of the new technology but did not change their organizational form to realize the benefits. This may be the most reflective mirror for every company today that has “purchased AI.”

Six: Epilogue: An Epitaph Index for the Age of Compute

Reviewing the above wreckage, we can derive a list of specific lessons:

Route Failure (Edison’s Direct Current): Those who tied their reputation and existing assets to a specific technology route will fail by using public opinion warfare instead of a shift when the paradigm changes.
Standard Debt (Frequency Wars, Japanese Grid): Early standard choices will incur interest for a century. Before the ecosystem converges, be wary of everything that “locks in.”
Paradigm Mismatch (Arc Light and Moonlight Towers): Winners who excel in the old paradigm often fail to enter the new paradigm because they are too adapted to the old paradigm.
Timing Mismatch (Electric Cars in 1900, Telharmonium): The direction is correct, but the accompanying infrastructure is missing. Pioneers become martyrs. The distinction between “half a step” and “ten steps” lies in the maturity of complementary technologies, not the beauty of the vision.
The Victor Loses (Edison, Westinghouse, Tesla): The spoils of technological warfare go to the capital integrators, not the victors.
Financial Blowback (Insull): Infrastructure booms are inevitably accompanied by leverage booms, and the collapse buries ordinary investors while leaving regulatory frameworks.
Regulation’s Double Edge (Britain’s 1882 Act): Regulation that introduces uncertainty over property rights too early can knock a leading country out of the race for a generation.
Incumbent Counterattack (Welsbach Mantles): “Faster carriages” will indeed appear and extend life by twenty years. Disruption is slower than believers think and more thorough than skeptics imagine.
Safety Lag (Feeks’s Death and UL): Safety standards are products of accidents, not gifts from technology. The cycle of accident—public outrage—regulation—certification industry will inevitably repeat.
Bubble Parasitism (Electric Therapy and Sugar Refining Scams): The market size of fraud equals the public’s reverence minus their understanding.
Organizational Lag (Forty-Year Productivity Paradox): The largest failures are those organizations that buy new technologies but do not restructure their processes—paying the cost of the revolution but missing out on its benefits.

The true thread of the history of electricity’s spread is not “great inventions change the world,” but rather a system that, through countless mistaken bets, real deaths, and capital reduced to ashes, slowly learns how to digest a general-purpose technology. Compute is following the same path. We cannot predict who will be this cycle’s Insull, who will be Charles Brush, or what will become the Telharmonium sold for scrap—but we can be certain that someone already occupies each of those roles today.

History does not repeat itself, but it rhymes. And it rhymes most harshly in the story of the failures.