C+

Technology Evolution & Disruption

Technological Revolutions and Financial Capital

Carlota Perez · first published 2002

Perez's claim: every technological revolution runs the same financial script — installation, frenzy, crash, then the golden age. The question that matters for capital is not whether the technology is real, but which phase you are standing in.

The big picture

Perez studied five technological revolutions — from canals and railways to microprocessors — and found one recurring financial choreography. First, financial capital funds the installation of the new infrastructure faster than the real economy can absorb it. Valuations decouple from earnings (the frenzy), a crash re-couples them (the turning point), and only afterwards does production capital take over and monetize the installed base broadly (synergy, then maturity). The core bet: the crash is not the refutation of the technology — it is a scheduled phase between building it and profiting from it.

Why it matters now: this is the sharpest available framework for the 2026 AI question. Data centers, power and chips are classic installation-phase infrastructure. Whether the cycle stands late in the frenzy or already at the turning point is the single most consequential judgment in tech allocation today — and Perez gives it testable criteria instead of vibes.

Perez's phases — every revolution runs the same financial script INSTALLATION FRENZY CRASH SYNERGY MATURITY paper decouples from earnings turning point: crash re-couples production capital profits from USING it financial capital builds the infrastructure ? ? AI 2026: LATE FRENZY OR EARLY SYNERGY? SCORE IT — DON'T FEEL IT the crash is a phase of the revolution, not its refutation
Five revolutions, one script: build → bubble → crash → golden age. The allocation question is never "is it real" — it is "which phase".

The 3 strategic pillars

  1. Installation vs. deployment

    Every revolution has two halves separated by a crash: financial capital builds the infrastructure, production capital later profits from using it.

    The tell is who funds growth and where profits originate — from building the new capacity, or from applying it across the old economy.

  2. The frenzy and the paper-real gap

    Late in installation, asset prices race ahead of any possible earnings path — capital chases the theme itself rather than its cash flows.

    Measurable signatures: valuation dispersion collapses inside the theme (everything AI-adjacent re-rates together), capacity is financed ahead of demonstrated demand, and new issuance floods in.

  3. The turning point sets up the golden age

    The crash transfers the installed base into stronger hands at rational prices — the precondition for the broad, profitable deployment decades.

    Post-crash, regulation tightens, financing discipline returns, and returns migrate from infrastructure builders to the users of the infrastructure across every sector.

What a Closelook reader does with it

The working use is phase diagnosis before allocation style: in installation you own the builders and accept bubble volatility; at the turning point survival and balance sheets dominate; in synergy the winners are the appliers — the companies using the installed base to take margin in ordinary industries. The mistake the book prevents is fighting the wrong war: selling a real revolution because it crashed, or holding builder-phase exposure on deployment-phase assumptions. The phase, not the technology, decides which portfolio is right.

The bridge to the Closelooknet approach

This framework runs under a lot of Closelooknet's architecture already. Rubin is explicitly an installation-phase index — it tracks what builds the AI factory (power, packaging, memory, DC construction), which in Perez's terms is the infrastructure half of the cycle. The Capex Cliff framework asks her turning-point question in modern dress, and Agentic Timeline plus AW40 hold the deployment-side answer: who captures value from USING the installed intelligence. The house view stays probabilistic — the pack's scorecard is built to hold the frenzy case and the synergy case side by side rather than declare the phase settled.

Action-Kit — from theory to practice

Tooling & data

What you needWhere to get itCost
Capex and financing data for the theme's builders The installation-phase diagnostics: who funds growth, how far ahead of demand Company filings; hyperscaler capex disclosures each quarter Capex growth vs. revenue growth of the infrastructure buyers is the single most telling ratio. Free
Issuance and valuation-dispersion data Frenzy signatures: IPO/secondary volume in-theme, dispersion collapse stockanalysis.com IPO lists; any screener's theme-basket valuation spread Free

The formulas

  • Phase read (scored)

    Phase score = Σ criterion weights per phase signature; highest column wins — with the runner-up kept in view
    • 10 observable criteria: funding source, capacity vs demand, valuation dispersion, issuance, regulation posture, profit origin, talent flows, credit conditions, M&A character, retail participation

    Deliberately a scorecard, not a formula — the value is in being forced to score each criterion against evidence.

  • Paper–real gap proxy

    Gap = theme market-cap growth − theme aggregate EBIT growth (trailing 12m)
    • Basket market caps
    • Basket EBIT

    Widening gap = frenzy signature; violent closing = turning point; both narrow and positive = synergy.

Applied Pack · free members

Perez Applied Pack

The phase-mapping scorecard: ten observable criteria scored against the five phase signatures — so your read on the AI cycle is a column of evidence, not a mood.

  • Perez_Phase_Map.xlsx — the 10-criterion × 5-phase matrix with weighted scoring, a phase-read summary that always shows the runner-up phase, and a pre-filled example column you overwrite with current evidence
  • phase_tracker.py — stdlib-only: keeps a dated history of your scorecard reads in a CSV so the phase judgment becomes a time series you can audit later
  • README.txt — criterion definitions, where each data point comes from, and the educational-use disclaimer

Pack security

Macro-free Excel · plain-text Python you can read before you run it · no installers, no network access — the code works only on files you provide. Served only from closelook.net; we never distribute through download portals or email attachments. How to verify in 30 seconds →

SHA-256 57dbe814cb2d25adf09d6138182fa6bebb6a445ddd841e29577775ad6a2e26c8

The pack is a free-membership download — no card, free forever. Membership also delivers the Daily Pulse and Weekly Signal to your inbox.

Get free access →

Closelook publishes a market diary, not investment advice. This condensed read restates the book's ideas in our own words for education — for the author's full argument, go to the source.