xStocks Liquidity Tax.

1. Executive summary

Tokenized equity tokens (xStocks) trading on Solana's Raydium CLMM pools incur a swap-count-weighted median execution spread of 14 basis points against NBBO fair value — roughly 3× the spread achievable by a competitive dealer-RFQ venue like Qualiq.

Key findings across 4 pairs, 17,482 swaps, $8,276,871 sampled volume (all pairs: 60-day window, sample_rate = 5; true volume ≈ $41,384,355).

2. Headline findings

3. Per-pair analysis

Each pair's section shows the full distribution against NBBO, broken out by session (market hours vs. after-hours), with five companion charts: spread histogram, staleness vs. spread, intraday heatmap, Qualiq counterfactual, and cumulative savings.

3.1 TSLAx/USDC

8,409 swaps · $3,752,854 volume · 2026-03-16 → 2026-05-04

Metric	Value
Market-hours p50	13.6 bps
Market-hours p90	53.36 bps
After-hours p50	67.2 bps
After-hours p90	220.23 bps
Market-hours swaps	5,072 (60%)
After-hours swaps	3,337 (40%)

3.2 NVDAx/USDC

6,426 swaps · $3,592,252 volume · 2026-03-16 → 2026-05-04

Metric	Value
Market-hours p50	11.4 bps
Market-hours p90	44.06 bps
After-hours p50	67.6 bps
After-hours p90	227.01 bps
Market-hours swaps	3,330 (52%)
After-hours swaps	3,096 (48%)

3.3 AAPLx/USDC

1,337 swaps · $492,769 volume · 2026-03-16 → 2026-05-04

Metric	Value
Market-hours p50	22.1 bps
Market-hours p90	65.92 bps
After-hours p50	49.7 bps
After-hours p90	182.77 bps
Market-hours swaps	683 (51%)
After-hours swaps	654 (49%)

3.4 GLDx/USDC

1,310 swaps · $438,996 volume · 2026-03-16 → 2026-05-04

Metric	Value
Market-hours p50	21.2 bps
Market-hours p90	101.45 bps
After-hours p50	80.5 bps
After-hours p90	265.39 bps
Market-hours swaps	429 (33%)
After-hours swaps	881 (67%)

4. SOL/USDC — crypto-native baseline & methodology validation

This section replicates the Jump Crypto (March 2026) propAMM spread analysis using the same on-chain data methodology, establishing that Qualiq's analytical approach is sound and that sub-2-bps execution is demonstrably achievable on Solana for crypto-native pairs.

Interpretation

Raydium CLMM SOL/USDC shows somewhat wider spreads than Jump Crypto's propAMM finding because CLMM is a passive AMM — it prices via inventory risk, not competitive quoting. PropAMMs (programmable AMMs with intent-based quoting) achieve tighter spreads by attracting competitive market-maker flow. Qualiq's MDRF model applies the propAMM principle to tokenized equities: dealers compete via Request-for-Quote, targeting 4.0 bps market-hours against the current 14 bps Raydium AMM baseline.

A. Appendix — data sources & methodology

A.1 Data sources

A.2 Swap discovery methodology

1. Pool-based, not mint-based

   Token mint accounts receive 5,000–15,000 signatures/day (mints, burns, transfers, all pools). Pool state accounts receive only that pool's swap/LP transactions — 10–100× fewer, targeted to the specific liquidity we care about.

2. Signature pagination

   getSignaturesForAddress(pool_id, limit=1000) pages backward through history, filtered to the date range.

3. Batch transaction fetch

   JSON-RPC batches of 10 getTransaction calls per HTTP request.

4. Swap detection

   Fee-payer's preTokenBalances and postTokenBalances must show equal and opposite changes in base- and quote-mint balances.

A.3 Fair-value methodology

• Equity pairs (NBBO proxy) — ASOF backward join to the most recent 5-min bar. swap_market_open is computed from the swap's own UTC timestamp using NYSE trading hours (09:30–16:00 ET), not from the yfinance bar (which only contains market-hours data).
• Crypto pairs (CEX mid) — ASOF backward join to the most recent 1-min Binance bar. Always-open (no market-hours concept).

A.4 Spread formula

A.5 Limitations

• Data coverage. All 4 pairs use the full 60-day window (2026-03-05 → 2026-05-04, sample_rate = 5, 17,482 swaps total). TSLAx and NVDAx are the liquid headline pairs; AAPLx and GLDx have materially lower volume and should be interpreted as directional rather than statistically robust.
• Sampling. Swap ingestor fetches 1-in-5 transactions for TSLAx/NVDAx. Sampled volume represents ~20% of true on-chain volume; all dollar figures above should be multiplied by 5 to estimate true magnitude.
• NBBO proxy. yfinance provides close price as mid, not true NBBO bid/ask. True NBBO requires Polygon.io Developer tier (~$199/mo). Close price slightly underestimates spreads in thin markets.
• After-hours fair value. With no 5-min bar update after 16:00 ET, the mid used for after-hours swaps is the last NYSE close — staleness can be 8–16+ hours for overnight swaps, making after-hours spread estimates noisier (directional, not precise).