Tokenized Commodity & Metals Index Engine
A 100% cash-collateralized futures engine that mints composable, open-transfer commodity and metals index tokens, built to solve the storage, basis-risk, and liquidity problems that have kept everything except gold off-chain.
There is no single price for oil or for "metals." Our edge is packaging that fragmentation into one clean, capital-efficient token.
1. Market context (mid-2026)
On-chain RWAs (ex-stablecoins) sit at roughly $29B, up ~263% YoY from ~$7.9B in 2024; including stablecoins the market is north of $240B. Tokenized Treasuries (~$13.4B) and private credit (~$16.8B) dominate. Tokenized commodities are only ~$7.3B and ~$6B of that is gold, with XAUT and PAXG alone holding ~90–97% of the gold segment.
Read plainly: commodities are the most under-built large RWA category, and everything past gold is effectively greenfield.
2. Why we started here: the metals gap
We began with an index thesis, specifically a Battery / EV-metals index. We hit a wall almost immediately. On-chain, the only asset-backed, holdable metals tokens are gold and (thinly) silver. For platinum, palladium and the industrial/battery metals (lithium, nickel, cobalt, copper, rare earths), the only exposure that exists is leveraged synthetic perps (Ostium lists XPT/XPD; Hyperliquid HIP-3 lists metals), price bets settled in USDC against a liquidity pool, not something you can hold as backed collateral or compose into an index.
So the precise gap is not "no exposure", it is no spot-backed, redeemable, composable token for anything past gold. You cannot build a credible backed metals index out of perp positions. To build the index, we first have to build the primitive. That gap is the opportunity.
3. Two product lines from one engine
The same cash-collateralized brokerage architecture supports both a fast near-term product and the full index vision.
Line A: Direct spot-metal tokens (fast path)
The simplest version of the product: hold the metal at a Tier-1 broker (IBKR), mint a token against the position. For gold/silver this is London unallocated spot; for platinum, palladium, and battery metals it is COMEX/NYMEX/CME futures, which is anyway the only wholesale access point for those metals.
Why this beats the incumbent vaulted-bullion token model (PAXG/XAUT): carry cost. Headline figures (derivations in Appendix B):
IBKR US spot gold storage ≈ $0.15/oz/month (~0.04%/yr at 2026 prices); London unallocated ≈ ~0.1%/yr; spreads passed through from LBMA liquidity providers, not widened.
Retail physical bullion: 2–8% dealer premium on coins/bars + 0.5–1.5%/yr vaulted storage.
Incumbent gold tokens layer storage fees on top of gated physical redemption.
So a spot-metal token on broker rails is structurally cheaper to carry than both physical bullion and existing vaulted-bullion tokens and it is the only practical route to put Pt/Pd/battery metals on-chain as backed assets at all.
Line B: Cash-collateralized rolling-futures index tokens (the full vision)
The differentiated primitive: composite index tokens backed by 100% cash-collateralized rolling futures, with open Web3 composability and managed short exposure. Detailed in §5–§6.
The two lines converge: for the exotic metals, Line A is the futures engine of Line B.
4. Competitive landscape: backed token vs. synthetic perp
The real incumbents for non-gold commodity exposure on-chain are not the vaulted-bullion tokens (PAXG/XAUT); those only cover gold. They are the RWA perp venues:
Ostium: the largest RWA-perps DEX (~$6B monthly volume, ~$213M open interest as of mid-2026, ~97% in non-crypto pairs; gold alone ~$71M OI). Broadest metals coverage: gold, silver, copper, platinum (XPT), palladium (XPD). Oracle-priced (Stork), liquidity-pool counterparty, on Arbitrum/Polygon.
Hyperliquid (HIP-3): dominant perp DEX (~44% of all perp-DEX volume). Permissionless RWA perp listings since Oct 2025; ~$2.65B RWA-perp open interest. USDC-settled, up to 40x leverage, 24/7. Liquidity heavily concentrated in one deployer (Trade.xyz, ~90%+), a real fragility.
Synthetix: synthetic assets via a pooled-debt model (Optimism/Base); a pioneer, but RWA is not its focus and the architecture is heavy.
All three give synthetic price exposure: a leveraged bet settled in USDC against a crypto-collateralized pool, with funding-rate cost and holder-side liquidation. None give you a backed, redeemable asset you can hold without leverage, use as clean collateral, or compose into a curated index.
That is Nex Labs' lane. We are not competing for the 40x speculative trade; we are building the asset-backed, low-leverage, redeemable index token for treasuries, stablecoin collateral, and allocators who want the exposure held, not rented. Honest trade-offs: the perp venues have deep live liquidity today (we have a cold-start to solve) and sit closer to trustless (we carry broker + issuer counterparty risk, see Appendix E). Our offsetting advantages: no funding-rate bleed, no holder liquidation, redeemable backing, index composability, and a MiCAR-compliant wrapper.
5. The cost case vs. retail exposure
Most retail commodity exposure (ETFs like USO/UNG, plus Turbos, Certificates, CFDs) carries hidden, compounding value-leaks. Headline drags below; full math in Appendix A.
Financing drag
8–9%/yr (overnight borrow markup)
0% (100% cash-collateralized, no borrowing)
Bid-ask spread
0.50–1.50% (broker markup)
< 0.02% (direct CME/ICE books)
Management fee
0.45–1.00% (static expense ratio)
Programmatic pass-through
ETF roll friction
up to 10–15%/yr in steep contango
index-managed roll
Physical-backed designs carry their own bleed: crude land storage of $0.30–0.70/bbl/month works out to ~8.5%/yr of principal at $70 oil; grain storage of $0.05–0.08/bushel/month is a ~12%/yr drag on wheat. Cash-collateralized futures convert that physical liability into an interest-earning treasury balance.
The yield offset. Positions are 100% cash-collateralized, and futures need only a thin margin posting, so the large majority of the backing sits as free cash that earns a money-market yield at IBKR (benchmark minus ~0.5%, ≈3–4% on USD in 2026). That yield runs against roll cost: where an ETF holder eats contango with nothing to offset it, our idle collateral is working. In normal or backwardated markets, net carry can be roughly flat-to-positive; in steep contango it goes negative, but materially less so than the ETF because of the yield. (Full net-carry math and the account-segment caveat: Appendix B.)
6. Architecture (condensed)
Composite indexing. Each token baskets multiple benchmarks (e.g. 40% Brent / 40% WTI / 20% Dubai) into one asset, killing single-node basis risk and smoothing localized spikes, a lower-volatility instrument suitable for treasury and stablecoin collateral.
Multi-account segregation. Each pool runs on its own dedicated IBKR sub-account ledger. A Short-Oil-V1 mint funds a segregated short sub-account; capital is fully siloed, so a spike in one market can never drag down another pool's collateral.
Open-transfer liquidation (Epoch State Mutation). A short's loss is theoretically unbounded, so the pool is engineered so 100% cash backing is not exhausted. Two defenses: (1) over-collateralization, short notional is capped well below pool cash, a deliberate buffer; and (2) conservative buy-to-cover, brokerage orders fire at a maintenance threshold set far above zero equity, not at the brink. On-chain, the token does not freeze, its backing simply mutates from live futures to settled cash.
After mutation, token price = settled cash balance ÷ total supply. Holders can still exit on secondary AMMs/OTC at a market-driven discount; arbitrageurs buy in to claim raw cash value post-clearing. New shorts route to a freshly provisioned V2 sub-account with a 100% clean runway.
Residual gap risk (stated honestly). Buffers shrink tail risk; they do not erase it. Futures can lock limit-up or gap over a weekend, WTI went negative in April 2020, and the LME cancelled nickel trades mid-squeeze in March 2022, so an exit may be impossible at the modelled price. Design consequence: the settled-cash token redeems at actual remaining cash, which can sit below the pre-event NAV, and a protocol buffer/insurance fund (not the individual holder) absorbs first-loss slippage. We do not represent the buffer as infinite.
Narrative indices. The index layer composes themed baskets: Strait of Hormuz Oil, Green-Tech Rare Earths, Battery / EV, bridging real-world geopolitical flashpoints with on-chain prediction-market demand.
Compliance. MiCAR Asset-Referenced Token (ART) path for EU passporting, but see the open question in Appendix E: a futures-referencing index may instead be classified as a MiFID financial instrument, a heavier path. Worth a formal legal opinion before the framing is locked.
7. Roadmap phases
Phase 1: Architecture & API staging (current). IBKR API sandbox + multi-account provisioning; audits of the epoch-shifting contracts.
Phase 2: Index composition & oracles. Nex Labs Composite Engine (multi-benchmark feeds); Chainlink NAV oracle tracking across sub-ledgers; reserves-verification design (Appendix E).
Phase 3: Spot-metal tokens & energy indexing. Ship Line A spot gold/silver tokens and the Global Oil Index (long-side rolling futures); first narrative geopolitical index.
Phase 4: Metals index & sharded shorts. Platinum/palladium/battery-metal futures tokens → the original Battery/EV index; automated buy-to-cover + open-transfer epoch mutation with simultaneous V2 deployment.
Appendix A: Cost-drag math (retail & physical)
Crude oil physical storage. Commercial land-based tank storage (Cushing, Rotterdam) runs ~$0.30–0.70/bbl/month under normal conditions, spiking to $1.50+/bbl/month in super-contango. At $70/bbl, $0.50/month = $6.00/yr = ~8.5% of principal annually, before insurance, shrinkage, and vaporization.
Agricultural storage. Grain elevator fees ~$0.05–0.08/bushel/month. At $6.00/bushel wheat, $0.06/month = $0.72/yr = ~12% annual bleed.
Retail financing drag. Overnight Turbo/CFD positions are bought on margin; brokers charge a benchmark (SOFR ~5%) plus a 2.5–4% retail markup → ~8–9% annualized to hold a long open overnight.
Retail spreads. A wholesale NYMEX crude contract may show ~0.01% ($10 on a $70k contract); retail Turbos/CFDs run 0.50–1.50%, starting every trade deep in the red.
ETF roll drag. Commodity ETFs hold futures and roll monthly. Predictable, large rolls get front-run by HFT (sell front-month low, buy next-month high). In steep contango this tracking error costs ~10–15%/yr vs spot.
Appendix B: IBKR metals economics, collateral yield & net carry
Figures from IBKR published rates (2026); verify in Client Portal before relying on them, as IBKR may change rates at its discretion.
Gold / silver (spot, unallocated).
US Spot Gold (USGOLD): min commission $2.00/trade; storage ~$0.15/oz/month. At ~$4,800/oz that annualizes to ~$1.80/oz = ~0.04%/yr.
London unallocated (XAU/XAG USD): storage ≈ notional × 0.1%/yr (worked example: 500 oz × $1,900 × 0.1% ÷ 360 ≈ $2.64/day). Spreads passed through from LBMA liquidity providers without widening.
Physical delivery is possible (FideliTrade, US-only) but expensive: ~$500 processing + shipping/insurance, i.e. only worth it for genuine offtake, not for token backing.
Platinum / palladium. No IBKR spot product. Access is via CME/NYMEX/COMEX futures (PL, PA; 100 troy oz/contract for palladium), with eligible physical delivery via registered warrant → handled by the cash-collateralized futures engine (Line B), not spot.
Battery / industrial metals. Lithium, cobalt, nickel, copper and related contracts trade as CME cash-settled futures → again, the futures engine, not spot vaulting.
Collateral yield & net carry. IBKR pays interest on idle USD in the securities segment, benchmark (Fed Funds) minus ~0.5%, ≈3.1–3.8% in 2026, on balances above $10k for accounts over $100k NAV; the commodities segment, where futures margin is posted, earns nothing. Because margin is a thin slice of notional, IBKR's universal account sweeps the excess cash to the securities segment, so the bulk of a 100%-cash-collateralized pool earns yield while only the posted margin does not. Net carry ≈ collateral yield − roll cost − fees: roughly flat-to-positive in normal/backwardated markets, negative in steep contango (but smaller than an ETF's, which has no yield offset). The segment-sweep behaviour must be confirmed operationally and rates can change at IBKR's discretion.
Takeaway. Gold/silver get the cheap spot/unallocated path; everything more interesting rides the same futures rails we already need for the index. Versus retail bullion (2–8% premium + 0.5–1.5%/yr storage) and vaulted-bullion tokens (storage + gated redemption), broker-rail tokenization is the lower-carry structure across the board.
Appendix C: Why physical-backed commodity tokens fail (the 4 blockers)
Most prior attempts tie a token to a physical barrel/bar in a warehouse, which introduces four severe friction points:
Illiquid premium / mismatched arbitrage. Slow, gated physical redemption lets on-chain prices detach from spot by up to ~5% in volatile periods; arbitrageurs can't cheaply burn a token and take delivery of a barrel or copper cathode.
Storage, decay, operational burn. Grains spoil, oil needs containment, metals need vault security. Passing variable real-world carry onto an ERC-20 produces decaying balances that break DeFi composability.
Geographic basis risk. Anchoring to one custody node (e.g. WTI at Cushing) means a localized disruption spikes the token, failing as a global macro hedge.
SEC vs CFTC turf war (US). A direct claim on a commercial contract draws CFTC scrutiny as an unregulated retail commodity transaction while the SEC eyes the wrapper as a security, choking institutional issuance.
Nex Labs sidesteps all four by moving from physical-custody logistics to programmatic capital efficiency: cash-collateralized rolling futures (no storage/decay/redemption), composite indexing (no single-node basis risk), and a MiCAR path (clear EU legal footing, subject to the classification question in Appendix E).
Appendix D: Demand pools
DeFi treasury diversification. DAOs/stablecoin protocols are overexposed to USD/Treasury/wrapped-BTC collateral; a liquid, macro-resilient, non-fiat commodity index is uncorrelated collateral for lending markets without inheriting crypto volatility.
Anti-inflation alternative to BTC. Institutional allocators want direct exposure to the structural inputs of the economy (energy, food, metals) to hedge real CPI prints, rather than BTC as a proxy store of value.
Narrative-driven speculation. Web3 capital has strong appetite for high-velocity thematic products (cf. prediction markets); themed indices plug geopolitical flashpoints directly into on-chain momentum.
Appendix E: Open questions: reserves transparency, counterparty trust & MiCAR classification
Centralized trust model. Assets live in IBKR sub-accounts, so the operative risk is operational control over that account (a compromised credential or rogue insider), not IBKR's solvency; IBKR is well-capitalized and over-collateralized. Nex Labs is nonetheless a trusted issuer. The Chainlink NAV oracle prices the position; it does not prove the cash exists.
Reserves dashboard: verifiable or worthless. A live dashboard of real-time sub-account balances is worth building only if the underlying data is independently verifiable: read-only broker API feeds an auditor can pull directly, signed/timestamped statements, and periodic third-party proof-of-reserves attestation. A self-reported balance the issuer could edit is not proof of reserves, it is the same unverifiable display that preceded the FTX and Celsius collapses, and it carries direct fraud liability. The feature earns trust only when a holder can verify it without trusting us. Build the verifiable version, or don't claim reserves.
Custody & withdrawal controls (IBKR institutional). Three layers blunt the rogue-actor/compromise risk: (1) multiple authorizers, withdrawals and changes to settlement instructions require N joint approvals, and the required count can only be increased; (2) whitelisted destinations, funds can only leave to pre-registered accounts, and re-pointing the whitelist is itself authorizer-gated; (3) least-privilege roles, traders get trade-only access with no withdrawal rights, auditors read-only. Pair with a reconciliation halt: auto-pause minting if broker balances and NAV-implied reserves diverge past a threshold. Keep on-chain admin-key holders separate from IBKR funds authorizers so no single team controls both sides.
Other open items. 24/7 token vs. closed-futures-market hours (weekend/overnight stale-NAV depeg risk); AMM liquidity cold-start (who seeds depth); revenue model (mgmt fee vs mint/redeem spread); mint/redeem mechanics (permissionless vs whitelisted APs, KYC gateway); settlement chain choice.
Last updated