What the Deflation Index actually measures.

Three lines on a chart.

The first line goes down. It's a weighted average of four technology costs that fell, year after year, for as long as anyone has measured them: computing, communications, energy, batteries. We call it the Deflation Index. In 1990 the index was 100. By 2025 it was 0.03. Whatever was in that basket costs about a three-thousandth of what it did when the Cold War ended.

The second line goes up. It's M2 — the broad money supply tracked by the Federal Reserve. Same span. Same starting point. In 1990, M2 was about $3.3 trillion. By 2025 it was $22 trillion. On a 1990=100 scale, M2 climbs from 100 to 676.

The third line also goes up, but more slowly. CPI — the prices ordinary people pay for ordinary things — rose 150% over those thirty-five years, an annual average of about 2.7%.

That's the chart. Three numbers: −99.97, +576, +150. Hold those in your head.

What the index is, and isn't

The Deflation Index is not a counter-CPI. It does not measure the cost of living. It does not include rent, healthcare, child care, college tuition, or a coffee. It deliberately excludes most of what you spend money on.

What it measures is narrower and more specific: the technology input layer — the raw units of capability that everything else is built from. A billion calculations per second. A gigabyte moved across a network. A kilowatt-hour from the sun. A kilowatt-hour stored in a battery. These are the commodities where engineering progress, economies of scale, and a great deal of capital have driven the cost toward zero.

We measure inputs, not products. Not your phone bill — the wholesale cost of moving your bits. Not your electric bill — the cost of generating the electron. The distance between those two numbers is not noise; it's the subject of this entire project.

Where the data comes from

Every series traces to a published source: the Federal Reserve and BLS for money and prices, IRENA for solar, BloombergNEF for batteries, market surveys for compute and connectivity. Each sector begins when defensible measurement begins — batteries in 2010, when the first mass-market packs shipped — rather than stretching back to years nobody measured. We would rather publish a shorter honest series than a longer invented one.

All of it — data, methodology, calculations, and our own corrections — is on GitHub. If a number on this site can't be traced to a source, that's a bug, and we'd like to hear about it.

Why the gap matters

If technology made the things in this basket 99.97% cheaper, the abundance went somewhere. It didn't all show up as lower CPI — CPI rose 150%. It didn't all show up in wages — median real wages grew about 30% over the same period. It didn't disappear.

Most of it was absorbed by sectors not in the index: housing, healthcare, education, services. The places where productivity is hard, supply is constrained, and prices rose faster than the dollar fell. The gap between the falling line and the rising lines is, in a literal sense, where the abundance went.

That's the question this site is built around. Not "is inflation real" — it is. Not "is technology saving us" — partially. The narrower, more honest question: the parts of the economy that did what technology promised — what happened to those gains?