A passive, long-only, rules-based index on the physical supply chain of the AI buildout. No options, no cash leg, no floor, no manager discretion. Fully invested at all times.
Supersedes RE-FLOOR v1.1 and v2.0, which studied an option-funded downside floor. That work is retained separately: it concluded the floor was buildable but cost roughly two-thirds of the upside to run, and was not implementable below ~$80,000 because sector-ETF LEAPs are quoted rather than traded. This version takes the other path — capture the growth, size the position to survive it.
The constraint on AI is physical, not computational. Data centres need firm power; firm power needs turbines, switchgear, transformers and crews. Chips need fabs; fabs need lithography, etch, metrology and consumables. Everything needs copper, steel, water and cooling. The fuel cycle needs uranium and enrichment. Someone finances the buildout and someone insures it.
The index owns all ten of those layers at equal weight. It deliberately excludes the hyperscalers and model companies — that exposure already sits in every broad index anyone holds. It owns the bottleneck.
| Universe | 61 names, ten sleeves, from the resource-economy research map |
| Sleeve weight | 10.00% each |
| Name weight | Equal within sleeve: 10% ÷ live names in that sleeve |
| Name cap | 5%, excess redistributed pro rata within the sleeve |
| Rebalance | Semi-annual, third Friday of March and September, at the close |
| New listings | Enter at the next scheduled rebalance, never mid-period |
| Cash | None |
| Costs assumed | 45 bp annual fee; 8 bp per unit of one-way turnover; dividends reinvested at a 1.8% blended yield |
Weights currently run from 0.91% (the eleven-name semiconductor sleeve) to 3.33% (the three-name sleeves). Largest position 3.33% — comfortably inside RIC diversification limits.
| Sleeve | n | Members | Standalone CAGR | Vol | Max DD |
|---|---|---|---|---|---|
| Power & Grid | 7 | GEV, ETN, HUBB, PWR, POWL, CEG, VST | 20.1% | 27.2% | −55.1% |
| Semis & WFE | 11 | AMAT, LRCX, ASML, KLAC, ACLS, TER, SNPS, CDNS, CAMT, ONTO, ENTG | 20.9% | 32.9% | −77.8% |
| Defense | 4 | HII, GD, LMT, HWM | 16.2% | 21.7% | −55.0% |
| Metals & Majors | 7 | FCX, SCCO, NUE, BHP, RIO, VALE, GLNCY | 15.8% | 36.2% | −75.2% |
| Precious & Royalty | 7 | FNV, WPM, RGLD, NEM, PAAS, CDE, TFPM | 12.5% | 38.4% | −71.6% |
| Water & Cooling | 3 | VRT, ECL, XYL | 18.3% | 23.3% | −49.5% |
| Fuel Chokepoints | 3 | CCJ, ALB, SQM | 16.4% | 34.3% | −63.6% |
| Optionality | 4 | OKLO, SMR, MP, LEU | 18.6% | 52.2% | −80.6% |
| Finance & Capital | 8 | CME, ICE, TPL, BN, BIPC, BEPC, BX, APO | 20.6% | 25.7% | −69.3% |
| Insurance & Risk Transfer | 7 | CB, WRB, ACGL, KNSL, EG, RYAN, AJG | 13.4% | 19.2% | −41.0% |
The index is less volatile than seven of its own ten sleeves. That is the diversification working.
| Index | CAGR | Vol | Max DD | Worst 12m | Best 12m | Losing years | Sharpe |
|---|---|---|---|---|---|---|---|
| RE-61 | 21.1% | 22.3% | −51.6% | −47.5% | +155.5% | 13.4% | 0.87 |
| QQQ | 14.4% | 19.7% | −51.3% | −46.2% | +84.4% | 14.2% | 0.64 |
| SMH — semiconductors | 16.1% | 27.4% | −65.7% | −51.8% | +154.6% | 27.7% | 0.52 |
| SPY — S&P 500 | 10.3% | 17.5% | −55.1% | −46.7% | +72.4% | 16.4% | 0.49 |
| XLI — industrials | 10.2% | 20.2% | −61.9% | −55.8% | +95.8% | 19.2% | 0.42 |
| IWM — small caps | 5.5% | 24.8% | −71.3% | −47.8% | +126.7% | 32.4% | 0.15 |
RE-61 is net of fees and trading costs; benchmarks are gross total return. Turnover 19%/yr.
QQQ is the honest comparison. It earns 14.4% with a near-identical drawdown. The index has to beat that bar, not the S&P's, to justify existing.
Every rule was chosen by running the alternative over the same 22.7 years with the same costs.
Weighting — sleeve-capping is what cuts the drawdown; equal weight within sleeve edges out cap weight.
| Scheme | CAGR | Max DD | Sharpe | Turnover |
|---|---|---|---|---|
| Equal weight, all 61 | 20.4% | −56.9% | 0.82 | 18% |
| Market-cap weight | 19.9% | −58.7% | 0.78 | 16% |
| Sleeve-capped, cap-weighted within | 20.8% | −52.2% | 0.86 | 17% |
| Sleeve-capped, equal-weighted within | 21.1% | −51.6% | 0.87 | 19% |
| Sleeve-capped, valuation-tilted | 20.3% | −52.2% | 0.83 | 19% |
Sleeve weights — tilting toward the "growth" sleeves or the defensive ones both underperform flat.
| Sleeve weighting | CAGR | Max DD | Sharpe |
|---|---|---|---|
| Equal sleeve — 10 × 10% | 21.1% | −51.6% | 0.87 |
| Balanced tilt | 20.3% | −53.0% | 0.86 |
| Growth-capture tilt | 20.8% | −54.7% | 0.87 |
Rebalance frequency — annual returns marginally more; semi-annual gives up ~40 bp for a two-point shallower drawdown. Monthly buys nothing and doubles turnover.
| Frequency | CAGR | Max DD | Turnover |
|---|---|---|---|
| Monthly | 21.2% | −53.8% | 44% |
| Quarterly | 21.0% | −53.4% | 26% |
| Semi-annual | 21.1% | −51.6% | 19% |
| Annual | 21.5% | −51.9% | 14% |
All 61 names carry a completed intrinsic value: full Damodaran ginzu for operators, commodity NAV on the live spot deck for miners and streamers, utility EV/EBITDA for merchant generation, justified P/BV for insurers, distributable earnings for alternative managers, FFO multiples for the infrastructure vehicles, NAV/sum-of-parts for Brookfield, 2-stage FCFE for the brokers, reverse-DCF for the pre-revenue tail.
Feeding it into the weights made the index worse in testing (20.3% vs 21.1%). It is published on every constituent as context for where new money goes between rebalances, not as an input to the rule. Current fair values and BUY / NEAR / WAIT signals for every constituent live on the Resource Map (last struck 22 October 2026).
This is a concentrated, cyclical, fully invested equity index. It lost 51.6% peak to trough in 2008–09 and 47.5% over its worst rolling twelve months. There is no floor and no hedge.
| Episode | RE-61 | SPY |
|---|---|---|
| 2008–09 financial crisis | −51.6% | −55.1% |
| Worst rolling 12 months | −47.5% | −46.7% |
| Share of rolling years in the red | 13.4% | 16.4% |
| Best rolling 12 months | +155.5% | +72.4% |
Sizing is the only real protection in a long-only product. Held at a weight where a 50% drawdown is survivable, this index has historically paid roughly twice the S&P's return for it. Held at a weight where it is not, none of the above matters.
Full replication, 61 positions, $91–$333 each. Any broker with fractional shares handles it, tracking error is zero, and at this size you pay no fund fee — so the DIY version earns the gross index, roughly 45 bp a year more than every table above.
Each sleeve gets $1,000; within a sleeve every name gets an equal share.
| Sleeve | Names | Per name | Sleeve total |
|---|---|---|---|
| Power & Grid | 7 | $143 | $1,000 |
| Semis & WFE | 11 | $91 | $1,000 |
| Defense | 4 | $250 | $1,000 |
| Metals & Majors | 7 | $143 | $1,000 |
| Precious & Royalty | 7 | $143 | $1,000 |
| Water & Cooling | 3 | $333 | $1,000 |
| Fuel Chokepoints | 3 | $333 | $1,000 |
| Optionality | 4 | $250 | $1,000 |
| Finance & Capital | 8 | $125 | $1,000 |
| Insurance & Risk Transfer | 7 | $143 | $1,000 |
| Total | 61 | — | $10,000 |
Name-level dollars follow directly from the rule: 10% per sleeve, split equally across its live names.
| Version | Names | CAGR | Max DD | Tracking error |
|---|---|---|---|---|
| Full replication | 61 | 21.1% | −51.6% | — |
| Largest 3 per sleeve | 30 | 20.9% | −51.2% | 5.0% |
| Largest 4 per sleeve | 38 | 21.0% | −52.7% | 2.9% |
The 30-name subset holds the largest three names in each sleeve.
Index construction study. Backtested results are hypothetical, computed by applying the rules uniformly across the sample, and are not a record of trading. The $10,000 allocation illustrates the rule at that size; it is not investment advice or a recommendation on any security named.