HSI Methodology v0.1 · working methodology

How household-controlled intelligence becomes a measurable economic ratio.

The Household Sovereignty Index measures the relationship between net autonomous productive income and the household needs that remain externally purchased.

HSI = Net Autonomous Productive Income / Remaining Household Needs

1.00 = sovereignty parity.

Formal notation: HSI = NAPI / RECR, where NAPI (Net Autonomous Productive Income) is divided by RECR (Residual External Cash Requirement — remaining household needs).

Every published value is designed to resolve from headline index → household distribution → calculation → constituent → source observation → timestamp → transformation → methodology version.

HSI=

NAPI

RECR

NAPI (Net Autonomous Productive Income)

Net annual market income generated by qualifying household-controlled autonomous intelligence after direct productive operating costs and annualized productive-capital costs.

RECR (Residual External Cash Requirement — remaining household needs)

The annual cash value of household needs that remain externally purchased after verified household-controlled infrastructure supplies qualifying needs internally.

HSI-C (Household Sovereignty Capacity — modeled)

HSI-C 0.29 means modeled qualified productive capacity equals 29% of the household's selected residual external requirement.

Scope

The ratio is the final compression layer, not the whole model. Many inputs resolve into NAPI, many inputs resolve into RECR, and HSI is what remains when one is divided by the other.

The founding chain

sunlight → electricity → compute → intelligence → autonomous action → market value → residual external cash → household needs

Calculation sequence

Six steps. No hidden leap.

HSI-C measures capacity, not realized income. Every non-observed constituent remains explicitly classified and visible.

01
Define the household
Select household structure, geography, tenure, and a named need standard. Income is not required for the capacity calculation.
02
Resolve external need
Calculate the baseline requirement, then subtract only needs supplied internally with verified equivalent service. The remainder is RECR.
03
Run the sovereignty checks
Test annual and temporal solar support, local execution, and autonomous work. Each check is answered on its own evidence.
04
Price accepted outcomes
Apply defensible output-market prices and a visible paid-utilization assumption. Technical capacity is not silently converted into demand.
05
Calculate net capacity
Subtract direct productive operating costs and equivalent annual capital cost from modeled gross autonomous revenue to obtain NAPI.
06
Locate the crossing
HSI-C equals NAPI divided by RECR. 1.00 means modeled output covers remaining household needs, with the three sovereignty checks answered.

Sovereignty checks

Energy, local intelligence, autonomy.

A household isn't sovereign just because the economics work on paper. Three things must also be true.

01Energy

Can household-controlled solar support the productive compute? Measured as SEI (Sovereign Energy for Intelligence). Methodology v0.1 requires SEI-A (Sovereign Energy for Intelligence · Annual Sufficiency) ≥ 1.00 for provisional qualification, and reports SEI-T (Sovereign Energy for Intelligence · Temporal Self-Supply) separately while the strong-form temporal threshold remains unresolved.

SEI-A ≥ 1.00 · SEI-T reported

02Local intelligence

Can all counted productive AI work run locally without renting a frontier model? Measured as SCI (Sovereign Compute and Intelligence). Every productive AI workload counted toward HSI must be executable on household-controlled compute without a rented frontier-model API.

SCI = 1.00

03Autonomy

Can the system finish economically useful work without continuous human direction? Measured as EA (Economic Autonomy), through the APWB (Autonomous Productive Work Benchmark). The threshold is experimental and is published as such.

Threshold experimental

Where the cheapest qualifying stack sits. The SCF (Sovereign Compute Frontier) tracks the lowest-cost combination of household-controlled hardware and locally executable intelligence that clears the economic-autonomy check. It is expected to move discontinuously when new models, memory systems, accelerators, or quantization methods appear, and that discontinuity is a property of the measurement rather than a defect in it.

The productive side

Four questions decide what a household can earn.

Each question has a formal instrument behind it, and each instrument carries its own evidence state. Parts of this chain are deterministic today; parts are still being measured into existence. Both are shown.

01 · Benchmark infrastructure being developedCan it do the work?
Autonomous Productive Work Benchmark · APWB
02 · Comparable-market pricing today, direct transaction pricing laterWhat is the result worth?
Market Value Oracle · MVO
03 · Scenario today, observed utilization laterCan the work find a buyer?
Paid Utilization · PU
04 · Calculated deterministically from the aboveWhat does the household actually net?
Net Autonomous Productive Income · NAPI

What APWB actually measures. APWB measures the whole system — local model, agent and orchestration, tools, memory, compute and software environment — against a productive intention rather than a single prompt.

The MVO evidence ladder. Strongest to weakest: settled autonomous transaction, comparable settled service, accepted contract, published service price, market quote, comparable service proxy, human compensation proxy, pure scenario. A priced outcome inherits the class of the rung it was priced from.

Paid Utilization scenarios. Paid Utilization measures the share of qualifying available productive capacity that actually finds paying demand. PU10 · PU25 · PU50 · PU75 are v0.1 sensitivity scenarios, not observed market utilization and not forecasts.

PU10
PU25
PU50
PU75

Capacity and reality

Capacity and reality are measured separately.

HSI-R publishes when representative qualified observed data exist. HSI-C is never used as a substitute for HSI-R.

HSI-C (Household Sovereignty Capacity — modeled)

Capacity

What qualifying infrastructure could produce under explicit standardized market and utilization assumptions.

HSI-R (Household Sovereignty Realized — observed)

Realized

What it actually produces using observed transactions, utilization, energy, costs, and residual needs.

Looking forward is a separate series. The HSI-FWD (Household Sovereignty Forward Curve) is a planned probabilistic curve over crossing dates under explicit technology, cost, capability, utilization, and household-need scenarios. It publishes with its scenario assumptions and model version, separately from measured HSI, and never in place of it.

Naming and geography

The series is U.S. HSI-C.

C and R are measurement modes; geography is a separate dimension. The series is U.S. HSI-C today, and the same methodology extends to other countries and regions as their inputs qualify.

National construction

A distribution, not an average household

v0.1 models a weighted distribution of U.S. households, not a single fictional average. The working national headline is the weighted median household HSI-C. The engine also retains the full distribution — P10, P25, median, P75, P90, coverage, and a provisional Sovereign Household Share — for research use. Aggregation methodology remains provisional pending validation before v1.0.

SHS (Sovereign Household Share)

A research statistic, kept in the engine

The share of modeled households with a qualifying HSI at or above 1.00. It is retained in the engine alongside the distribution percentiles and reported as provisional while v0.1 aggregation methodology is under validation. Distribution statistics belong in institutional and research views; they are not needed to understand what HSI measures.

Evidence language

Uncertainty is part of the result.

Every input receives exactly one primary status. The vocabulary prevents a model assumption from masquerading as reality.

Observed

Directly measured by an eligible external source or actual transaction.

Measured

Generated through a documented HSI experiment or benchmark.

Derived

Calculated entirely from observed and/or measured inputs.

Proxy

An observable substitute for a target not yet directly measurable.

Scenario

An explicit hypothetical assumption chosen for analysis.

Unobserved

A required concept with no defensible measurement yet.

Replacement rule

Every non-observed assumption carries a rationale, introduced version, sensitivity, and the condition under which better evidence replaces it. Calculated values inherit the evidentiary weakness of their inputs: Scenario evidence cannot silently become Observed or Derived simply because it passes through a formula.

Scenario → Proxy → Observed

Provenance

Lineage before a value can publish.

Every published constituent retains source, observation period, retrieval date, geography, units, classification, confidence, transformation history, assumptions, methodology version, and lineage. A candidate cannot publish when required lineage, licensing, freshness, reconciliation, or integrity checks fail. The complete registry, licenses, and current HSI status live on the sources page. This table does not imply that a source has already produced a publishable snapshot.

Research-beta HSI source registry preview
Registry ID	Constituent	Canonical source	Class	Cadence / state
`DS-ENERGY-001`	Location-specific PV production	National Laboratory of the Rockies · PVWatts V8	Derived	Model versionModeled; not a published national snapshot
`DS-ENERGY-002`	Solar resource and weather	National Laboratory of the Rockies · National Solar Radiation Database	Observed	Dataset releaseResource data supporting PVWatts; not a published snapshot
`DS-ENERGY-010`	Residential electricity retail sales	U.S. Energy Information Administration · EIA Residential Electricity Retail Sales	Observed	MonthlyConnector exists; not a published national extract
`DS-CAPITAL-001`	Representative residential PV installed cost	U.S. Department of Energy · DOE PV System Cost Model	Derived	AnnualModel ingest not in a published snapshot
`DS-CAPITAL-002`	Distributed solar and storage installed prices	Lawrence Berkeley National Laboratory · U.S. Distributed Solar and Storage	Observed	AnnualProject data not in a published snapshot
`DS-HH-001`	Household archetype weights	U.S. Census Bureau · ACS PUMS	Observed	AnnualNational production ingest in progress
`DS-RECR-001`	Household expenditure basket	U.S. Bureau of Labor Statistics · Consumer Expenditure Survey	Observed	AnnualProduction basket in progress
`DS-RECR-002`	Category consumer-price inflation	U.S. Bureau of Labor Statistics · Consumer Price Index	Observed	MonthlyProduction CPI mapping in progress
`DS-RECR-003`	Regional price level	U.S. Bureau of Economic Analysis · Regional Price Parities	Observed	AnnualSame-year spatial rule exists; production ingest in progress
`DS-TASK-001`	Occupation and task taxonomy	National Center for ONET Development · ONET Web Services	Observed	QuarterlyTaxonomy ingest not in a published snapshot
`DS-TASK-002`	Autonomous productive-work benchmark	Household Sovereignty Index · HSI Autonomous Productive Work Benchmark	Measured	Benchmark releaseHSI-certified measurement not yet released. External capability evidence is tracked while the benchmark matures.
`DS-COMPUTE-001`	Full-system inference power reference	MLCommons · MLPerf Inference Power	Measured	Benchmark releaseExternal capability-frontier reference; not an HSI-certified result
`DS-COMPUTE-003`	Wall energy per accepted APWB outcome	Household Sovereignty Index · HSI Compute Harness	Measured	Certified runHSI-certified measurement not yet released. External benchmark evidence is tracked while the HSI Compute Harness matures.
`DS-MODEL-001`	Locally executable model identity	Multiple · Model Publisher and Hugging Face Registry	Observed	Release eventRelease, licensing and memory evidence; not a capability claim
`DS-HW-002`	Household hardware acquisition price	Multiple · Household Hardware Acquisition Price	Proxy	WeeklySampling methodology still research
`DS-MARKET-001`	Realized market-clearing output prices	Future qualified transaction systems · Autonomous Output Market-Clearing Price	Observed	TransactionNot yet available
`DS-MARKET-002`	Comparable-output price research	Household Sovereignty Index · Comparable Output Price Research	Proxy	MonthlyComparable-output research; not realized clearing prices
`DS-UTIL-001`	Paid Utilization	Future qualified transaction systems · Paid Utilization	Scenario	TransactionPU10 / PU25 / PU50 / PU75 sensitivity scenarios
`DS-REV-001`	Realized qualifying household machine revenue	Future opt-in household transaction network · Realized Qualifying Household Machine Revenue	Observed	TransactionNot yet available. HSI-R stays unavailable.
`DS-FIN-002`	Treasury yield-curve reference rates	U.S. Department of the Treasury · U.S. Treasury Daily Treasury Yield Curve Rates	Observed	Business dayRate series not in a published snapshot

Publication governance

The last valid publication stands until a better candidate passes.

Publications are produced from frozen inputs and immutable snapshots. A candidate must pass validation before a public channel pointer can advance. Failed candidates do not replace the last valid publication. Source revisions create new observations and, where appropriate, separately labeled restated series. Previously published observations are preserved rather than silently rewritten.

01
Publications are produced from frozen inputs and immutable snapshots.
02
A candidate must pass validation before a public channel pointer can advance.
03
Failed candidates do not replace the last valid publication.
04
Source revisions create new observations and, where appropriate, separately labeled restated series.
05
Previously published observations are preserved rather than silently rewritten.

Language

Definitions & terminology.

This is the glossary for the whole site. Terms coined by HSI are labeled as such and link back to the section that defines them; borrowed industry units are labeled separately so a reader can always tell which is which. Calculator and index labels link here.

HSI terms

HSI term

HSI-C

Household Sovereignty Capacity — modeled

How household-controlled intelligence becomes a measurable economic ratio.

Six steps. No hidden leap.

Define the household

Resolve external need

Run the sovereignty checks

Price accepted outcomes

Calculate net capacity

Locate the crossing

Energy, local intelligence, autonomy.

Four questions decide what a household can earn.

Capacity and reality are measured separately.

Capacity

Realized

The series is U.S. HSI-C.

A distribution, not an average household

A research statistic, kept in the engine

Uncertainty is part of the result.

Lineage before a value can publish.

The last valid publication stands until a better candidate passes.

Definitions & terminology.

HSI-C

HSI-R

NAPI

RECR

SEI

SCI

SEI-A

SEI-T

EA

APWB

MVO

PU

SCF

SHS

HSI-FWD

kWh

PVWatts

Boundaries that stay in force.

Read the U.S. series, then test a household.