Every Tech Revolution Killed an Industry. None Did It in Under 4 Years.

The Pattern Nobody Talks About

Go to any tech conference, open any financial newsletter, scroll any LinkedIn feed. The assumption is everywhere: AI is going to destroy [industry X] overnight.

Overnight. That's the word people keep using.

But when you actually look at the data — every major industry that was killed by technology, going back over a century — a pattern emerges that nobody seems to mention:

No major industry has ever lost more than 50% of its value in under 4 years from technology disruption alone.

Not Blockbuster. Not Kodak. Not newspapers. Not coal. Not travel agencies. Not typewriters. Not the telegraph. Not horse-drawn carriages.

Every single one of them died. But none of them died fast.

The Evidence

Here's every major industry killed by technology disruption, with actual timelines:

Note: "Time to 50% revenue loss" measures from when the disruptor became commercially viable, not from when it was invented.

How Fast Did They Actually Decline?

Timelines are abstract. Dollars and percentages are not. So here are six of those cases again — this time with the actual numbers, the years they happened, and the share each industry lost. These aren't estimates pulled from a press release. They come from SEC filings, Census Bureau retail data, the EIA, the CDC's household phone survey, Pew, and the Bureau of Labor Statistics.

Sources: Blockbuster peak revenue and 2010 bankruptcy (SEC filings, Wikipedia). Newspaper ad revenue peak/decline (Pew Research, MarketingCharts). Coal's electricity share (U.S. EIA, Wikipedia). Landline household penetration (CDC household phone survey via The National Interest). Travel-agent employment (U.S. Bureau of Labor Statistics). Department-store sales (U.S. Census / Statista; Sears figures from company filings).

Look at the right two columns together and the pattern from the thesis stops being a slogan and becomes arithmetic. The fastest case here — Blockbuster, the textbook example of a company "killed overnight" — still took three years from Netflix's 2007 streaming launch to bankruptcy, and roughly six to seven years from peak revenue. And Blockbuster is the quick one. Newspapers, coal, landlines, and department stores all needed something closer to fifteen years to give up half their ground.

Notice also what "decline" meant for each one. Coal didn't go to zero — it went from running the grid to running a fifth of it, and in 2022 it was still a 19.7% slice of U.S. electricity. Department stores didn't vanish; they shrank from a $232 billion business to a $133 billion one and kept the doors open. Even at the bottom, most of these industries were still very much alive. The thing that died first was their growth and their pricing power — not their existence.

The Investing Implication

This is where the history pays for itself. The disruptor's stock can run for years — tripling, ten-bagging — while the incumbent's revenue erodes only gradually underneath it. Those two facts coexist. They almost always coexist. So the phrase that gets repeated at every conference, "this industry is doomed," is usually true and almost never means "sell tomorrow."

For a portfolio, that splits cleanly into two moves on two different clocks:

• 01 The disruptor moves on the growth clock — fast. If you want to own the winner, the window opens early and the re-rating happens long before the incumbent's numbers look terrible. Netflix's run started while Blockbuster still had thousands of stores. Waiting for the incumbent to actually die means you've missed most of the disruptor's move.
• 02 The incumbent moves on the decline clock — slow. You do not have to dump it the week the threat appears, and panic-selling on the headline is historically the wrong trade. But "slow" is a deadline, not a reprieve. A 15-year, 80% erosion is still an 80% erosion. The right posture is gradual, deliberate rotation that starts early — while the decline still looks survivable — not a single dramatic exit and not the comforting decision to do nothing for another four quarters.

Put the two together and the timing rule for AI disruption is simple: buy the disruptor on the growth clock, trim the incumbent on the decline clock, and never confuse the two. The mistake that history punishes most isn't being early to the disruptor or slow to leave the incumbent — it's assuming both charts move at the same speed. They never have.

Why Decline Is Always Slower Than Growth

New technology grows exponentially. But the industries it replaces decline linearly. There are five structural reasons why:

• 01 Existing contracts don't expire overnight. Power purchase agreements, commercial leases, enterprise software agreements — they run 3-20 years. The money is contractually locked in place even when everyone knows the future has arrived.
• 02 Customers have switching costs. Data migration, staff retraining, psychological inertia. Moving from Blockbuster to Netflix meant learning a new interface, trusting your credit card online, and changing a Friday night habit. Multiply that across millions of households.
• 03 Regulatory moats protect incumbents. Licenses, charters, certifications, zoning laws, professional associations. Taxi medallions didn't disappear when Uber launched. Broadcast licenses didn't evaporate when YouTube went live. Regulation moves at government speed.
• 04 Physical infrastructure persists. Buildings, networks, equipment — they depreciate over 7-20 years. A coal plant doesn't vanish because solar is cheaper. It runs until the economics make it physically impossible to keep the lights on.
• 05 Human behavior changes slowly. Habits take months to years to shift, not days. People kept renting DVDs for years after streaming was clearly superior. They kept buying newspapers after the internet was free. Behavior is the last thing to change.

What This Means for AI

AI is the fastest-deploying technology in human history. Our engine tracks a 1,800:1 speed ratio between AI deployment and physical infrastructure buildout. ChatGPT reached 100 million users in 2 months. The telephone took 75 years to reach the same penetration.

But the supply side isn't the bottleneck. The demand side is.

Contracts, regulations, behavior — none of these have sped up. An enterprise SaaS agreement still runs 3 years. A commercial lease still runs 10. A medical licensing board still meets quarterly. A 55-year-old accountant still needs 6 months to learn new software.

The bottleneck isn't whether AI CAN replace an industry. It's whether the money CAN move fast enough.

The Investor Takeaway

Here's what "slow decline" actually feels like when you're holding:

Blockbuster shareholders watched their stock drop from $30 to $0.06 over 7 years. At every stage, there was a reason to hold — "they'll adapt," "streaming is overhyped," "they still have 9,000 stores." Each quarter the stock dropped another 20-30%, and each quarter there was a new reason to believe the next quarter would be different. Seven years of hope. Then zero.

Kodak shareholders held through a 15-year decline. The company had $14 billion in revenue when digital cameras appeared. They had patents, brand recognition, retail relationships. They even invented the digital camera. None of it mattered. The decline was slow enough to feel survivable at every point — until it wasn't.

The timeline gives you time to ACT. It does not give you time to ignore.

If you're worried AI will destroy your portfolio overnight — the data says you have years, not months. But if you're using that timeline as an excuse to do nothing — the data says every disrupted industry eventually lost 50-80% of its relative value. The slow speed of decline is what makes it so dangerous: it's always comfortable enough to justify holding on for one more quarter.