What Companies Will Actually Profit in the AI Era? Lessons from 250 Years of Tech Cycles

1. The Electricity Revolution: Bubbles burst, infrastructure endures, users win

In 1879 Edison lit a 40‑hour lamp, sparking public awe comparable to first‑time GPT use. Capital rushed in, creating over 270 U.S. power companies in the 1880s, but early DC transmission required a plant every mile, leading to high costs and idle generators at night.

Like today’s AI startups and massive compute spending, the era saw speculative frenzy without clear business models. The bubble collapsed in 1893, wiping out hundreds of firms, yet the electric grid survived as a lasting asset.

After AC became viable, from 1900‑1930 the U.S. electricity coverage rose from 8 % to 68 %, and infrastructure such as transformers, wires and meters persisted.

The ultimate winners were not Edison or grid builders but companies that integrated electricity into their business models:

Ford Motor : Founded 1903, introduced electric conveyor belts in 1913, cutting chassis assembly time from 12.5 hours to 93 minutes, enabling affordable cars.

Disney : Founded 1923, used electric sound recording to give Mickey Mouse a voice, globalizing animation.

Coca‑Cola : Launched 1886, leveraged electric refrigeration to distribute chilled soda widely.

These firms merely harnessed electricity; they did not create it.

2. The Internet Bubble: Dark fiber turned into the backbone of Google and Apple

The World Wide Web appeared in 1989, igniting hype; a .com domain could dramatically raise valuation. Netscape’s IPO in 1995 doubled on debut, and a Silicon Valley joke claimed “I don’t know the business, but I can raise money.”

By 1999 the Nasdaq peaked at 5,048 points; the bubble burst in 2000, dropping to 1,114 in 2002, erasing 78 % of market value and eliminating thousands of firms.

After the crash, massive fiber, data‑center and broadband infrastructure remained. Only 5 % of laid fiber was in use, the rest dubbed “dark fiber,” later becoming the foundation for Google (IPO 2004), Facebook (2004), YouTube (2005), and the iPhone (2007) that launched mobile internet.

Carvalho‑Pérez identifies four recurring phases: explosion → speculative frenzy → bubble burst → mature application. Capital first fuels breakthroughs, then speculative excess, after which the over‑built infrastructure is repurposed by survivors.

3. Bubbles as Foundations for Future Eras

Although bubbles destroy short‑term investors, they construct long‑term assets such as power grids, undersea cables, nationwide broadband, and compute clusters. These projects require massive capital that only speculative “hot money” supplies.

Trans‑Atlantic cables once cost $200‑$500 million each, with dozens built simultaneously, totaling over $100 billion in fiber investment. The same pattern repeats now: trillions flow into AI compute clusters, many of which currently serve only low‑margin drawing or text generation, yet they form the “electric grid” for the forthcoming AI era.

4. AI Follows the Same Path as Prior General‑Purpose Technologies

Critics claim AI differs because it replaces cognition rather than merely connecting devices. History shows each general‑purpose technology first displaces existing jobs—oil lamps, manual factories, print media—before becoming ubiquitous infrastructure.

When AI matures, it will be a standard tool across industries, no longer a branding claim.

5. Takeaways for Entrepreneurs and Investors

Do not overestimate AI’s short‑term impact, but also do not underestimate its decade‑long potential to reshape commerce.

The most valuable future players may be modest startups that embed AI into concrete sectors—manufacturing, education, services—rather than those chasing raw model or chip breakthroughs.

Just as Ford, Disney, and Coca‑Cola quietly leveraged electricity, today’s niche AI firms could explode once the underlying compute infrastructure stabilizes.

History repeats its rhythm, and the next wave of world‑changing companies is already building beneath the surface.