Which financial package paid for each technology to spread, and how long it took to come. Thirty-one technologies from farming to language models, on the same clock as Tech evolution, testing the waits in Front-wheel drive.
Technologies sit above the line, at the year they arrived. The packages that paid for them run below, each from the first time it appears. An arc lands where a package first paid for a technology, and the gold run continues until it carried the spread at scale. Hover or tap anything for its record.
A technology, coloured by its stem in Tech evolution. Hollow where no financial package carried its spread.
Paid first: a patron, a financier, a ruler or a state, on the row just under the line.
The wait: the arc lands when the package first paid; the gold run goes on to when it carried the spread at scale. A tick marks its legal form.
A package: a financial form, from its first appearance in this record. A branch shows which form grew out of which.
A boom and its crash, on the bottom row, marked at the year it broke. A dashed ring is a boom still running.
Perez's surges: shading runs from each surge's big bang to its crash, then a darker band for the turning point.
What the record shows
In years the wait shrinks. Against the time since, it holds.
Farming waited 6,200 years for written ledgers, and the web seven years for the ad auction. But every interval shrinks as the clock nears the present. Measured against the time since each technology arrived, which is how Tech evolution's clock measures time, the waits show no downward trend. From 1500 to 1950 a package usually reached scale after about a sixth of that time. For the internet, the web and language models it took between three-tenths and four-fifths.
Where the state was the buyer, there was no wait.
Air power, synthetic fertiliser and nuclear deterrence had their package within a year of arriving and at scale within four, because the state paid for them through war and defence budgets. Powered flight had its first army contract within five years. Nuclear weapons never went near a market at all.
Someone paid first, and half the time it was a ruler or a state.
Of the 26 technologies with a known first payer, 13 were paid for first by a court, a crown, an army, a parliament or a public body. Patrons and single financiers paid for most of the rest: Johann Fust for Gutenberg, the Duke of Bridgewater for his canal, Drexel, Morgan for Edison.
Booms come as a package scales.
Front-wheel drive says the gap between a technology and its package breeds bubbles. In this record only two booms came before the technology had a package: the oil boom at Pithole in 1865 and the motor-company promotions of the 1890s. Every other boom came while the package was scaling or after, from the canals in the 1790s to the dot-com bust. The frenzy belongs to the package.
Perez's rhythm holds. AI's place in it is open.
Each of Carlota Perez's five surges shows its frenzy and crash here: the canals (1793), the railways (1847), the Baring crisis (1890), 1929 and 2000. Language models are paid for out of the largest firms' own profits, as Ford paid for mass production, with debt topping it up. Perez places AI inside the information surge, and in that reading production capital should lead, as it does. If AI is a sixth surge, a financial frenzy should follow, and the new debt forms would be its start.
The wait, two ways
Each mark is one technology, at the year it arrived. The dot is the wait until the package first paid for it, the bar runs on to when the package carried the spread at scale, and the ring is the figure Front-wheel drive uses, placed at the deck's own starting year. Switch the measure to see the same waits as a share of the time since each technology arrived.
The dashed lines mark a wait of a tenth, a third and all of the time since a technology arrived. Waits of nothing or less sit on the bottom row.
The deck's six against the record
Front-wheel drive pairs six technologies with their packages and reads the shrinking wait as a pattern. Here is each pairing set against the record.
In Front-wheel drive
Deck
This record
Reading
Language models: a package, and new forms forming
The largest technology firms are paying for AI out of their own profits. Microsoft, Alphabet, Amazon and Meta spent over $251bn on capital in 2024 and about $400bn in 2025, and guided to roughly $720–760bn for 2026, with debt filling more of the gap each year. None of the new forms has yet taken a legal shape of its own, as venture capital did with the 1979 pension ruling.
Whether this is a bubble is argued. US indices corrected by about a tenth in March and again in July 2026, and South Korea's chip-heavy market fell about 40% between June and late July; by October US indices were at records again.
Carlota Perez argues that each technological revolution since 1771 is installed through a financial frenzy and a crash before it spreads through the economy. Her five surges, set against this record:
Surge
Big bang
Frenzy and crash
Package in this record
Holds
The record
Thirty-one technologies: Tech evolution's records at its two highest levels since the first ledgers, plus Front-wheel drive's six. Select a name to show it on the chart.
Technology
Arrived
Paid first
Package that carried its spread
First · legal · scale
Wait
Of the time since
Sources
How this was made
Each technology keeps its date from Tech evolution. For each one, five research passes on 14 October 2026 looked for who paid first and which financial package carried its spread, with three dates for that package: the first time it paid for the technology, when it took legal or institutional form, and when it carried the spread at scale. Where no financial form carried the spread, the record says so and why.
Waits count from the technology's date in Tech evolution. A share of the time since divides the wait by the years between the technology's arrival and now, which is how the chart's clock spaces time: each step to the left is a longer stretch of the past.
Dates before 1500 are approximate, and several pairings are judgement calls; each record gives its reason. The dot-com boom came before the web's ad auction but after its first venture funding (1995), so it counts as a boom of a scaling package.