What 1925 Already Told Us About 2026

Every few years a piece of journalism shows up that we end up sending around the office and re-reading two or three times. George Anders’s essay in Tuesday’s Wall Street Journal Journal Report on Artificial Intelligence is one of those. The headline is unassuming — “What the 1920s Can Teach Us” — but the argument inside it is one of the most useful frames we have seen for thinking about how to invest, plan, and live through what the next ten years of AI build-out are going to do to America.

Anders’s case is simple: we have been here before. Long before ChatGPT, a different technology swept through American society like wildfire, thrilling young people and eliciting warnings of declining family values. Long before MrBeast, a different kind of entertainment technology created a roster of world-famous celebrities. And long before Elon Musk, a different industry pioneer dragged the entire manufacturing sector into the era of mass production. That was the 1920s — cars, radio, airplanes, full-length movies, in-home electricity — and the shockwave it sent through society looks remarkably like what AI is doing right now.

We think every client should read the original piece. But the points worth pulling forward into the portfolio conversation are these.

1. The shock of profound change is itself a market signal — not a reason to retreat

The pioneering sociologists Helen and Robert Lynd spent a chunk of the 1920s in Muncie, Indiana — which they code-named “Middletown USA” in the book that came out of it — cataloguing the ways fast cars and open roads were annihilating the town’s longtime social norms. So long to the old rule of “a high school boy does not need much spending money.” Goodbye to “rain or shine, I never miss a Sunday morning at church.” America was rapidly becoming more consumerist and more secular. Weekend automobile rambles were replacing church attendance. Courtship moved out of the chaperoned family parlor and into the unsupervised car ride. Traditionalists wrote opinion pieces protesting that “the home is being endangered.” F. Scott Fitzgerald put a reckless-driving incident at the moral heart of The Great Gatsby.

None of that stopped the cars from selling.

That is the first lesson, and it is the one we keep telling clients who are nervous about AI: the discomfort is real, the resistance is real, the cultural anxiety is real — and it is decoupled from the spending curve. Over time, as Anders puts it, “most of the old guard slowly, grudgingly, got used to the new ways.” The corporate capex doesn’t wait for the cultural acceptance. Microsoft, Alphabet, Meta, Amazon, and Oracle will spend around $600 billion on AI infrastructure in 2026 regardless of how a Pew Research focus group feels about ChatGPT.

2. Innovation creates jobs in places no one was looking

This is the part of the Anders piece that we think is the most underappreciated, and the part we want every client to internalize.

In 1920, the profession of “electrician” barely existed. By 1930, more than 80,000 Americans were working as electricians. Census data from 1930 showed 168,000 people working in rubber factories, most of them making tires to accommodate Detroit’s booming production of cars, trucks, and buses. Another 450,000 people were building roads, bridges, and the supporting structures the ever-expanding auto industry needed. Hollywood’s directors and big-name stars like Mary Pickford weren’t the only ones getting rich from the movie-viewing boom — local entrepreneurs opened more than 20,000 movie theaters across America by 1930. Set designers, makeup artists, gossip-minded journalists, all found new work. The car industry didn’t just create assembly-line jobs in Detroit; it created mechanics, filling-station operators, road-atlas publishers, driving instructors, and roadside cafe operators in every county in the country.

This is the framework we are using to build the book. We are deliberately under-weight the consumer-AI application layer (the movie theaters of this cycle — high beta, fashion-driven, easy to over-pay for) and over-weight the spillover trades (the electricians, the tire factories, the road crews). The picks-and-shovels framing isn’t a cliche — it’s a hundred-year-old observation about how every industrial revolution actually mints its millionaires.

3. Electrification went from 35% to 68% in ten years. That is the compounding curve to model.

One number Anders cites that we keep coming back to: in 1920, electricity reached 35% of American homes. By 1930, it was 68%. In one decade, the share of American homes wired for electricity nearly doubled. Refrigerators, toasters, vacuum cleaners, mixers, and electric lights made their consumer debuts in that window. The labor-saving compounding effect on American household productivity was enormous — and the listed equity that captured it (General Electric, Westinghouse, AT&T, the utility holding companies) made multi-generational wealth for the families who bought it early and held it.

Generative AI adoption is moving faster. Per the John Doerr interview in the same Tuesday WSJ section (R2), 50% of Americans say they already use generative AI three years after ChatGPT launched. Doerr — the venture capitalist who backed Amazon at a $100 million valuation that is now $4.67 trillion — calls AI the biggest technology “tsunami” he has ever tracked, bigger than the personal computer, bigger than the browser, bigger than the iPhone. His exact words: “Biggest thing ever. Since everything. It has been underhyped.”

“Just three years after ChatGPT was launched, 50% of Americans say they use generative AI, and the value creation is off the charts.” — John Doerr, Kleiner Perkins, WSJ Tuesday R2

We agree directionally. The honest read is that the adoption curve for generative AI is steeper than the 1920s electrification curve was — because the marginal cost of distribution is essentially zero, and the install base of capable hardware (every smartphone) was already in every pocket before the tsunami started. The investment implication is that the spillover trades will compress in time too. The 1920s electrification build-out ran for roughly a decade. The AI build-out may compress that timeline to four or five years — which is exactly why we are taking exposure now, not waiting for the wave to crest.

4. The safety backlash is coming — it always does — and it is investible

One detail from Anders that landed for us: in the early 1920s, the injury and death rates for both aviation and automobile travel were appalling. Cars back then didn’t have rearview mirrors or four-wheel brakes. Airplane pilots went airborne by themselves — no licenses — after as little as 10 hours of in-flight instruction. The car-related death rate per mile driven in 1920 was twenty times what it is today.

Society did not just absorb this. It responded. Charles Lindbergh spoke out against the perils of low-cost flight-training schools and Congress passed the Air Commerce Act of 1926. Auto safety got better with both industry and government action — better mirrors, better brakes, shatter-proof windshields became standard. Cities like Los Angeles and Detroit installed red-yellow-green traffic lights. New Jersey became the first state to require driver’s licenses in 1924, with the state’s motor-vehicle commissioner declaring it “an absolute necessity… in order to conserve human life.”

Anders’s critique of today’s tech giants is fair: they have so far been more talk than action on AI safety, biomedical implants, and especially social media. Federal legislation keeps failing to come to a vote. But the safety wave is coming — the question for investors is which companies are positioned to ride it. We see three plays:

Cybersecurity. CrowdStrike (CRWD), Palo Alto (PANW), Zscaler (ZS), and SentinelOne (S) are the “shatter-proof windshield” trade for AI. As enterprises deploy AI agents that can read files, write code, and execute transactions, the attack surface explodes. The category leaders should compound through the entire build-out.

Governance / observability. Palantir (PLTR), Datadog (DDOG), and Snowflake (SNOW) sit at the “driver’s license” layer — the platforms enterprises use to track, audit, and govern what their AI systems are doing. As regulation arrives (and it will), these become non-discretionary spend.

Power-grid hardening. Data centers are creating exactly the political backlash Anders alludes to in his discussion of utility rates and local zoning. The names with the deepest political relationships — NextEra (NEE), Constellation (CEG), Dominion (D), GE Vernova (GEV), Vertiv (VRT) — are the ones that will be allowed to keep building. The fragmented merchant generators will not.

5. The radio cynics are always loud — and they are almost always wrong

One of the smaller delights of the Anders piece is the 1927 quote from H. G. Wells, the British author and intellectual, calling radio “inferior” entertainment that should be listened to “only by the sick, the lonely and the suffering.”

David Sarnoff, general manager of the Radio Corporation of America, shot back that he was trying to improve “the happiness of the nation” by delivering popular music to millions of people. Nearly a century later, Anders notes, that same argument still flares — though now it is more likely to involve TikTok, Reddit, or YouTube. “The doubters always have a point,” he writes. “With the passage of time, the innovators usually win out.”

This is the line we want every client to underline. The doubters always have a point. The innovators usually win out. The intelligent investor takes both seriously — respects the doubters enough to manage risk and avoid the bubbliest excesses, and respects the innovators enough to keep core exposure to the trend rather than fleeing it.

6. Where the 2026 parallel breaks down — and what to watch

We do not want to oversell the analogy. Three places where 2026 is not 1925:

Capital intensity. The 1920s automobile build-out was financed largely by retained earnings, regional banks, and a still-young Wall Street. The AI build-out is being financed by the deepest balance sheets in the history of capitalism — the five hyperscalers have a combined ~$320 billion in net cash and trillions in operating cash flow. That changes the financing risk profile. It does not eliminate it — the leverage is showing up in the venture-debt market (a separate WSJ piece in this same edition flagged $68.8 billion in venture debt, with AI taking 64% of dollars) — but the systemic risk is different in character.

Labor displacement velocity. The 1920s auto revolution created many more jobs than it destroyed, but it did so over twenty years. AI is moving faster — possibly a lot faster. The displacement may exceed the spillover creation in compressed windows. This is why our financial-planning work for clients is increasingly emphasizing skill-and-income optionality (multiple income streams, retraining-friendly career bridges, defensive insurance overlays) and why we are talking about long-term-care planning earlier in client conversations than we used to.

Energy. The 1920s ran on coal-fired electricity and gasoline derived from a domestic oil supply that was, at the time, vastly under-exploited. AI runs on electrons that are increasingly scarce. The energy story is the bottleneck and the opportunity in equal measure. We are over-weight the names that solve it — Constellation (CEG), GE Vernova (GEV), NextEra (NEE), Vertiv (VRT), Bloom Energy (BE) — and we expect this to be the single highest-conviction part of the book for at least the next two years.

7. Henry Ford’s answer

Anders closes his piece with a quote from Henry Ford’s late-career memoir, in which Ford rhetorically asked “Are we moving too fast?” and then answered his own question with a resounding No. “It is easier to go along with progress than to try to hold things back. Only old, outworn notions stand in the way of these ideas.”

It is tempting to read that as triumphalism. We read it differently. Ford spent the last decade of his life watching the consequences of the technology he created — the sprawl, the family disruption, the safety toll — and arguing not that those consequences didn’t matter, but that the only sane posture was forward, with eyes open, building the guardrails as you go. That is exactly the posture Anders is recommending, and it is exactly the posture we are taking with the Capital Wealth book.

Anders ends with a quote from Anthropic CEO Dario Amodei: “When put in the darkest circumstances, humanity has a way of gathering, seemingly at the last minute, the strength and wisdom needed to prevail.”

We tend to agree. We also tend to think that humanity gathering its strength and wisdom in a profitable, investible way looks a lot like building 1.4-gigawatt data centers in West Texas, hardening the grid in Virginia, stringing fiber under the Atlantic, and equipping every knowledge worker on earth with an AI co-pilot — and that the public-equity beneficiaries of that build-out are already named, already trading, and already sitting in the client portfolios we manage.