Silver-halide film out-filed every digital technology into the mid-2000s while improving at 23%/yr against digital's 30–90%. And inside digital, the same measurement watched CCD's rate decay while CMOS held, flagging the second disruption years before Sony ended CCD production.
Rate readings are platform output computed from patents public in each year. Note the second story inside the first: the technology that beat film was itself being out-improved.
Engineer Steven Sasson builds the first digital camera, toaster-sized, 3.6 kg, 0.01 megapixels. Executives shelve it to protect film margins.
Sony announces an electronic camera; Kodak commissions a landmark internal assessment. It concludes digital will replace film in roughly a decade. The finding changes nothing.
The Apple QuickTake 100, one of the first consumer digital cameras, is manufactured by Kodak itself. The capability was in-house; the commitment was not.
Market value tops out near $31 billion in February. Film filings run at 2,611 families that year, more than every digital technology combined. The filing chart says fortress; the rate chart says otherwise.
CMOS's improvement rate passes CCD's, years before CMOS is taken seriously on image quality. The first camera phones ship the same year. Round two is already decided.
The filing crossover arrives five years after the rate crossover, the pattern of this entire library, repeating inside the digital camp.
Smartphone volume locks in the active-pixel sensor. Film demand is now falling 20–30% per year.
On 19 January, Kodak files for bankruptcy protection: 37 years after building the first digital camera. Film filings that year: 56 families.
Sony, whose Mavica startled Kodak in 1981, announces the end of CCD sensor production. The disruptor is disrupted, exactly as the rate had been saying since 2000.
Kodak's leadership was not blind. Each of these held at board level, and none of them was the variable that decided the outcome.
True for decades, early digital was a toy next to a good film print. The rate does not measure who is ahead; it measures who is closing.
Rational finance: why trade a razor-blade business for commodity hardware? The rate priced the trend; the P&L priced the present.
Also true, into the mid-2000s CCD noise and uniformity were superior, and the CCD camp said so. CMOS was closing at roughly twice the speed, and its rate crossover came in 2000.
Kiosks, Photo CD, the Advantix film system, an early digital SLR line. Adjacent bets that protected film's role rather than following the fastest-improving technology.
“It would have been great if we'd had something like GetFocus to show management how serious the situation was.”
We forecast the improvement rate from two citation metrics, cycle time and knowledge flow. Both were visible the whole time.
CCD's cycle time stretched from under four years to almost seven, each generation took longer to arrive, while its knowledge flow halved. CMOS held a knowledge flow near 1.1 for two decades: later work kept building on it. Film never moved on either metric. The rate was reading the machinery underneath, not the market above it.
Provenance. Every patent carries a filing date, so the record can be replayed: every improvement rate on this page uses only the patents that were public at the time. No hindsight enters the series. Chart data: GetFocus platform exports, August 2026. The historical facts below come from the public record.
Bring us one live technology decision. Within 48 hours you get a free preliminary Decision Brief: a first-pass radar, improvement rates and suggested triggers.