Scope. Three AltStreet series over a short, fully disclosed window. Premiums: daily reference marks vs the Chainlink underlying for nine curated symbols, seventeen fixings, August 1–17, 2026. Cross-venue spreads: the same-token gap between the two quoting venues (Solana xStocks and a Robinhood Chain Uniswap v4 variant), sampled ~every 15 minutes, eleven days, August 8–18. Volume and breadth: the CoinGecko-sourced Backed xStocks universe on Solana, live-collection days only — the same eleven days. Spreads are quoted gaps before fees, gas, and depth, not executable round-trips. The window is seventeen days; nothing here is a “record,” a trend forecast, or a statement about any longer history. For what the instruments legally are and how the marks are computed, see the guide and the feed methodology; this report does not restate them.
Section 01 · Premium structure
The premium is persistent in this sample: SPY never priced below the ETF
If tokenized-equity premiums were symmetric tracking error, each symbol would spend about half its days rich and half cheap. That is not what seventeen fixings show. Tokenized SPY priced above the underlying ETF on all seventeen — a median of +58 bps, never lower than +28, once as high as +101. Tokenized AMZN was positive on sixteen of seventeen. At the other end of the table, tokenized META priced below its underlying on twelve of seventeen. The cross-section has persistent rich and cheap names within this sample, which is consistent with one-sided flow interacting with wrapper-specific frictions — creation and redemption are not free, so a name retail keeps buying can stay rich — rather than with a market arbitraged flat.
| Symbol | Fixings above underlying | Below | Median premium (bps) |
|---|---|---|---|
| SPY | 17 of 17 | 0 | +58.3 |
| AMZN | 16 of 17 | 1 | +43.8 |
| GOOGL | 15 of 17 | 2 | +32.5 |
| COIN | 12 of 17 | 5 | +15.1 |
| AAPL | 11 of 17 | 6 | +15.3 |
| NVDA | 11 of 17 | 6 | +7.6 |
| QQQ | 9 of 16 | 7 | +6.8 |
| TSLA | 8 of 17 | 9 | -0.3 |
| META | 5 of 17 | 12 | -19.3 |
Source: AltStreet tokenized-equity mark series (the feed behind /data/marks). Premium is the daily reference mark vs the Chainlink underlying, in bps; positive = token above underlying. QQQ entered the marked set Aug 2 (16 fixings). Symbols without a configured underlying feed publish a mark but no premium and are excluded rather than counted as zero.
Persistence does not mean stability. Tokenized AAPL swung from +91 bps to −61 bps between the August 3 and August 4 fixings — a 152-basis-point reversal in one day — and QQQ flipped from +43 to −31 between August 16 and 17. A holder marking a tokenized position to the token's own market is carrying basis volatility on that scale on top ofthe underlying stock's moves.
Section 02 · The Sunday effect
Sundays: dispersion doubles and the skew turns positive
The median absolute premium across the marked set ran 31 bps on the median weekday and 29 bps on Saturdays — but 77 bps on Sundays. The two largest daily readings in the entire series, 86.5 and 77.2 bps, are both Sundays, and the third Sunday's 42.8 bps is exceeded by only one weekday reading. Direction skews with it: on the three Sundays observed, nine of nine, eight of nine, and seven of nine marked symbols priced above the dated Friday close.
| Fixing | Day | Median |premium| (bps) | |
|---|---|---|---|
| Aug 1 | Sat | 27.2 | |
| Aug 2 | Sun | 86.5 | |
| Aug 3 | Mon | 31.4 | |
| Aug 4 | Tue | 29.8 | |
| Aug 5 | Wed | 20.4 | |
| Aug 6 | Thu | 31.0 | |
| Aug 7 | Fri | 9.5 | |
| Aug 8 | Sat | 30.1 | |
| Aug 9 | Sun | 77.2 | |
| Aug 10 | Mon | 30.4 | |
| Aug 11 | Tue | 35.5 | |
| Aug 12 | Wed | 40.9 | |
| Aug 13 | Thu | 46.1 | |
| Aug 14 | Fri | 15.9 | |
| Aug 15 | Sat | 29.2 | |
| Aug 16 | Sun | 42.8 | |
| Aug 17 | Mon | 31.2 |
Source: AltStreet tokenized-equity mark series. Cross-sectional median of |premium| across the marked symbols each day. Weekend fixings are computed against the dated Friday close of the underlying, per the feed's labeling rule. On Aug 9 and Aug 16, two of nine symbol marks were carried forward unchanged from the prior day (quotes not updating), so part of each Sunday cross-section is stale rather than freshly priced.
Two readings are compatible with a positive Sunday skew: participants paying up for exposure they expect to be validated at Monday's open, or thin weekend pools drifting on one-sided flow with no underlying print to anchor them. Seventeen days cannot separate the two — what they do establish is that the weekend session, the product category's headline feature, is when its prices say the least about the asset they reference. The stale-mark caveat in the exhibit note cuts the same way: when a fifth of the Sunday cross-section is carried forward rather than traded, a Sunday “price” is partly an echo.
Section 03 · The frozen quote
Six days of tokenized AAPL: pinned, jumped, pinned again
The clearest single artifact in the dataset is what tokenized AAPL's cross-venue spread did over six days in mid-August. On Thursday the 13th, all 96 samples returned exactly 8.84 bps — minimum, median, and maximum identical for a full day. Friday the spread jumped, touching 103.35. Then Saturday and Sunday pinned at exactly 103.35 bps on every sample, unfreezing only Monday. In a two-venue market, this pattern is consistent with at least one venue ceasing to update its quote: the published spread doesn't widen, it stops moving at whatever gap existed when quoting ceased.
| Date | Day | Median | Min | Max | Samples | State |
|---|---|---|---|---|---|---|
| Aug 13 | Thu | 8.84 | 8.84 | 8.84 | 96 | frozen — all samples identical |
| Aug 14 | Fri | 8.84 | 8.79 | 103.35 | 96 | live |
| Aug 15 | Sat | 103.35 | 103.35 | 103.35 | 84 | frozen — all samples identical |
| Aug 16 | Sun | 103.35 | 103.35 | 103.35 | 84 | frozen — all samples identical |
| Aug 17 | Mon | 91.84 | 52.06 | 103.35 | 84 | live |
| Aug 18 | Tue | 52.06 | 0.29 | 57.44 | 84 | live |
Source: AltStreet pricefeed cross-venue spread history, ~15-minute sampling, Solana xStocks vs Robinhood Chain (Uniswap v4). 'Frozen' marks days where every sample returned an identical value. Per-day sample counts changed from 96 to 84 on Aug 15 with the feed's sampling schedule; both are full-day collections.
The freeze also explains an otherwise paradoxical aggregate: quoted spreads narrowon Sundays. The cross-sectional daily median was 15.2 bps on the first Sunday and 16.9 on the second — the two lowest readings of all eleven days — against 19–31 bps on weekdays. A naive read would call the weekend market tight. The symbol-level data points the other way: several symbols' samples stop changing for long stretches, some pinning at near-zero residual gaps, and the narrow aggregate spread can therefore reflect stale quotes rather than genuinely tighter executable liquidity.
| Date | Day | Median spread (bps) | |
|---|---|---|---|
| Aug 8 | Sat | 29.3 | |
| Aug 9 | Sun | 15.2 | |
| Aug 10 | Mon | 24.5 | |
| Aug 11 | Tue | 24.5 | |
| Aug 12 | Wed | 19.3 | |
| Aug 13 | Thu | 31.0 | |
| Aug 14 | Fri | 21.8 | |
| Aug 15 | Sat | 26.6 | |
| Aug 16 | Sun | 16.9 | |
| Aug 17 | Mon | 24.5 | |
| Aug 18 | Tue | 29.7 |
Source: AltStreet pricefeed cross-venue spread history. Median across the nine marked symbols of each symbol's daily median spread. Sundays produce the two lowest aggregate readings; the symbol-level freeze episodes (Exhibit 3) show why narrow weekend spread statistics can be misleading.
Section 04 · Intraday dispersion
Daily medians of 13–65 bps hide intraday ranges of 100+
Daily medians are deliberately conservative — one bad print cannot move them. The cost of that robustness is that they understate what a participant actually experiences intraday. On August 18, tokenized COIN's spread ranged from 0.59 to 104 bps within a single day against a daily median of 65; tokenized META touched 147 bps intraday against a median of 19, and had hit 169 bps the day before — the widest single observation in the eleven-day window.
| Symbol | Daily median | Intraday min | Intraday max |
|---|---|---|---|
| COIN | 65.2 | 0.6 | 104.0 |
| AAPL | 52.1 | 0.3 | 57.4 |
| AMZN | 47.5 | 1.3 | 84.0 |
| SPY | 31.3 | 15.1 | 44.5 |
| GOOGL | 29.7 | 1.1 | 92.9 |
| META | 18.8 | 0.1 | 147.3 |
| TSLA | 14.5 | 0.0 | 133.4 |
| QQQ | 13.6 | 0.6 | 36.7 |
| NVDA | 12.9 | 0.3 | 32.0 |
Source: AltStreet pricefeed cross-venue spread history, 84 samples per symbol on Aug 18. Quoted gaps before fees, gas, and depth. A wide intraday max is a moment's quote, not a durable executable opportunity.
Section 05 · Where the volume lives
SpaceX and Circle out-trade every tokenized mega-cap
Across the eleven live days, the tracked xStocks universe turned over between $17.4M and $80.1M daily across 39–62 actively traded symbols — roughly $383M for the calendar week of August 10–16. The leaderboard is not what an index investor would predict: SPCXx (tokenized SpaceX, Nasdaq-listed since June 12 — see the guide's pre-IPO section for what those tokens were before the listing) led on seven of eleven days, and CRCLx (tokenized Circle) led on three more. Between them, a two-month-old listing and a crypto-economy name topped the table on ten of eleven days; the exception was the quietest Sunday, when TSLAx led at $1.9M. Tokenized Apple, index ETFs, and the rest of the mega-cap complex consistently trail. On this evidence the tokenized-equity buyer skews hard toward crypto-adjacent and recent-IPO names — the profile of a crypto-native audience buying the names it knows, not of index money moving on-chain.
| Date | Day | Volume ($M) | Active symbols | Top symbol | |
|---|---|---|---|---|---|
| Aug 8 | Sat | 45.2 | 50 | SPCXx | |
| Aug 9 | Sun | 21.7 | 39 | SPCXx | |
| Aug 10 | Mon | 80.1 | 56 | SPCXx | |
| Aug 11 | Tue | 68.8 | 59 | SPCXx | |
| Aug 12 | Wed | 64.1 | 62 | SPCXx | |
| Aug 13 | Thu | 59.0 | 62 | CRCLx | |
| Aug 14 | Fri | 59.1 | 58 | SPCXx | |
| Aug 15 | Sat | 34.9 | 57 | SPCXx | |
| Aug 16 | Sun | 17.4 | 44 | TSLAx | |
| Aug 17 | Mon | 66.8 | 58 | CRCLx | |
| Aug 18 | Tue | 51.4 | 59 | CRCLx |
Source: AltStreet tracked tokenized-equity universe (Backed xStocks on Solana, CoinGecko-sourced), live-collection prints only; backfilled history excluded. Herfindahl concentration ran 0.078–0.104 across these days (roughly a dozen effective names). For scale, Polymarket's daily prediction-market volume ran $35–42M over the same stretch — comparable to this entire universe.
The weekend collapse is symmetric across both observed weekends: Sunday volume was 27% of the following Monday on August 9–10 and 26% on August 16–17, with breadth dropping to 39 and 44 active symbols. Our Robinhood Chain market-structure report found the same shape in a different asset class — memecoins doing one-seventh of their business on the third of days that fall on weekends. Tokenized equities extend that finding to instruments that reference regulated securities: the always-open market keeps its underlying's hours.
Section 06 · What an allocator should take from this
Reading tokenized-equity prices without being fooled by them
- Expect a basis, priced in tens of bps. A persistent +58 bps median on tokenized SPY is a recurring basis investors must account for when entering or marking tokenized index exposure — before fees, and before any question of what the wrapper legally owes you.
- Treat weekend prices with additional caution. Sunday marks carry double the dispersion, a positive skew against a stale benchmark, and a meaningful share of carried-forward quotes. Neither a weekend premium nor a narrow weekend spread is evidence of anything except that the underlying market is closed.
- A quoted spread is not a fill. Intraday maxima exceed daily medians by 2–8x, the two legs live in different wrappers with separate redemption channels, and the feed's depth-adjusted executable measure is a different (and currently unpopulated) statistic.
- Size the market honestly. $383M a week across the whole universe is a promising experiment, not a parallel stock market — and its volume concentrates in crypto-economy names, which says as much about who is buying as about what is on offer.
Frequently asked questions
Do tokenized stocks track their underlying share price?+
Imperfectly, and with persistent structure within the observed sample. Across the seventeen daily fixings in AltStreet's mark series (August 1–17, 2026), the tokenized S&P 500 exposure (SPY) priced above the underlying ETF on every single fixing — a median premium of 58 basis points, ranging from +28 to +101 bps — and tokenized AMZN was positive on 16 of 17. At the other end, tokenized META priced below its underlying on 12 of 17 fixings. Some names are persistently rich and others persistently cheap in this window, which is consistent with wrapper-specific frictions and one-sided retail flow rather than symmetric tracking error. Day-to-day moves can be violent: tokenized AAPL swung from +91 bps to −61 bps between consecutive fixings, a 152-basis-point one-day reversal.
What happens to tokenized equity prices on weekends?+
Dispersion roughly doubles and the direction skews positive against the dated Friday close. The median absolute premium across AltStreet's marked symbols was 31 bps on the median weekday and 29 bps on Saturdays, but 77 bps on Sundays — the two largest daily readings in the series (86.5 and 77.2 bps) both landed on Sundays, and on all three Sundays observed the cross-section skewed positive (9 of 9, 8 of 9, and 7 of 9 symbols above the Friday close). A weekend premium can carry information about where participants expect Monday's open, but it can also reflect pool imbalance and thin liquidity — and roughly two of nine Sunday marks in this series were carried forward unchanged from Saturday because quotes stopped updating, so part of the Sunday cross-section is stale rather than freshly priced.
Why did tokenized AAPL's cross-venue spread freeze at one value?+
The observed pattern is consistent with quotes on at least one of the two venues ceasing to update for entire days. AltStreet samples each symbol's cross-venue spread roughly every 15 minutes; on Thursday August 13, every one of tokenized AAPL's 96 samples returned exactly 8.84 bps — minimum, median, and maximum identical. On Friday the spread jumped, and through Saturday and Sunday (August 15–16) every sample returned exactly 103.35 bps, unfreezing only on Monday. In a market made by two venues, a venue going quiet doesn't widen the published spread — it freezes it at whatever gap existed when quoting stopped. That is why a narrow weekend spread cannot be read as a healthy weekend market.
Which tokenized stocks actually trade the most?+
Recent listings and crypto-adjacent names, not the mega-caps. Across the eleven live collection days (August 8–18, 2026), SPCXx — tokenized SpaceX, which listed on Nasdaq on June 12, 2026 — was the highest-volume symbol on seven days, and CRCLx (tokenized Circle) led on three more; between them they topped the table on ten of eleven days. The exception was the quietest Sunday, when TSLAx led at $1.9 million. Tokenized exposure to Apple, Microsoft-style mega-caps, or index ETFs consistently trails tokenized SpaceX, Circle, Coinbase, and Robinhood — the tokenized-equity buyer, on this evidence, skews strongly toward crypto-economy and recent-IPO names.
How big is tokenized equity trading volume?+
The entire tracked xStocks universe on Solana — roughly 44 to 62 actively traded symbols per day — turned over between $17.4M and $80.1M per day across the eleven live days, about $383M for the full calendar week of August 10–16. For scale, that whole universe's daily volume is comparable to a single day of Polymarket prediction-market volume ($35–42M/day over the same stretch) and is a rounding error against the underlying equities: the real SPY alone routinely trades tens of billions of dollars daily. Concentration is moderate — a Herfindahl index around 0.08, or roughly a dozen effective names — so the volume is not one token's story, but the market is early-stage small.
Are these cross-venue spreads an arbitrage opportunity?+
Not as published. The spreads are quoted gaps between venue prices, not executable round-trips: they are computed before fees, gas, and depth, and closing a cross-wrapper gap can require separate issuance or redemption channels, jurisdiction-specific accounts, and prefunded capital on both legs. Intraday ranges are also far wider than daily medians — on August 18, tokenized COIN's spread ranged from 0.59 to 104 bps within one day against a daily median of 65 — so a single wide print overstates what was durably available. AltStreet's feed reserves a separate depth-adjusted 'executable' measure for this reason, and these figures are not it.
How does AltStreet compute these figures?+
Three collectors against AltStreet's own pricefeed and tracked-universe data. Premiums come from daily reference marks — the median of 15-minute TWAPs inside a fixed daily fixing window, liquidity-gated, compared against the Chainlink underlying feed and labeled against the dated close when the underlying market is shut; nine symbols in the curated set currently publish premiums, and symbols without a configured underlying feed publish a mark but no premium and are excluded rather than counted as zero. Cross-venue spreads sample each symbol roughly every 15 minutes across the two quoting venues (Solana xStocks and a Robinhood Chain Uniswap v4 variant), take the day's median per symbol, then the cross-sectional median. Volume and breadth cover the CoinGecko-sourced Backed xStocks universe on Solana, with every print stamped live-collection versus backfill; only live-collection days are used here.
Methodology & limitations
- Window discipline. Seventeen premium fixings and eleven venue/volume days is a short window disclosed as such. No figure here is described as a record, an extreme, or a trend; day-of-week comparisons rest on three Sundays and are labeled as observations, not seasonality estimates.
- Premiums. Daily reference marks are medians of 15-minute TWAPs inside a fixed fixing window, liquidity-gated, against the Chainlink underlying; weekend fixings are labeled against the dated Friday close. Nine curated symbols publish premiums; marks without a configured underlying feed are excluded rather than counted as zero. Carried-forward marks (identical values on consecutive days) are called out where they affect a claim.
- Spreads are quoted, not executable. Cross-venue gaps are sampled ~every 15 minutes across exactly two venues per symbol and summarized as daily medians. They exclude fees, gas, depth, and the separate issuance/redemption channels of each wrapper. The feed's depth-adjusted executable measure is reserved separately and is not used in this report.
- Volume provenance. The liquidity series stamps each print live-collection vs retrospective-backfill; this report uses live-collection days only (Aug 8–18) and draws no conclusions from the backfilled history. Universe is Backed xStocks on Solana as tracked by CoinGecko — one program on one chain, not the whole tokenized-equity market.
- Sampling change.Per-day spread sample counts moved from 96 to 84 on Aug 15 with the feed's sampling schedule. Daily medians are robust to the change; sample counts are printed in Exhibit 3 so readers can see it.
Researchers and journalists may cite any figure as “Source: AltStreet tokenized-equity mark and venue series.” The live marks publish free daily at /data/marks.
Sources
- Instrument structures, mark methodology, and the SEC's tokenized-securities taxonomy: Tokenized Equities in 2026: Valuation Marks, Venue Spreads, and What Tokenized Stocks Are Really Worth.
- Daily reference marks and feed integrity rules: AltStreet Tokenized Stock Reference Marks and feed methodology.
- Weekend activity patterns in an adjacent on-chain asset class: Inside Robinhood Chain's First Memecoin Boom.
- Product-level comparison of the xStocks program's structure and custody: Backed vs Ondo.
