On-chain purchases of US stock tokens see a 30% drop after market close; cross-exchange spreads, net of all fees, yield no profit.

Author: Cooper Duschang, Coin Metrics (Talos)

Compiled by: TechFlow

TechFlow Summary: The total spot market capitalization of tokenized US stocks has surpassed $3 billion, establishing a new on-chain channel for traditional equity allocation. However, recent research from Coin Metrics indicates that compared to the $12.3 billion in equity perpetual contract volume on Hyperliquid over the past week, the spot market remains in its early stages.

Data further reveals that although tokenized stocks support 24/7 quotes, they are highly dependent on the liquidity anchor of the underlying stocks. Once entering after-hours and weekend periods, the 1% order book depth on core platforms shrinks significantly by 19% to 33%, and cross-exchange nominal spreads become unprofitable after accounting for bid-ask spreads and slippage.

Key Takeaways

The total market cap of tokenized stocks has grown to over $3 billion. Investors can directly gain price exposure to US equities on-chain through multiple issuers such as xStocks, Ondo Tokenized Stocks, and bStocks.

During regular US market trading hours, NVDAx traded at a premium 58.6% of the time relative to the NVDA underlying stock, averaging a 3.1 basis point premium; meanwhile, NVDAon traded at a premium 88.9% of the time, averaging a 14.1 basis point premium.

Execution prices for tokenized stocks vary significantly across platforms and order sizes. A tokenized stock might offer the lowest slippage for a $10,000 order on one platform, but not for a $100,000 order, necessitating a combined assessment of slippage depth and asset coverage using cross-platform data.

Introduction

Between October 27 and November 2, various trading platforms recorded over $2 billion in turnover in the tokenized stock spot market. As on-chain allocation demand continues to grow, custodians have been steadily accumulating underlying stocks and issuing more tokens, pushing the total market cap of tokenized stocks above $3 billion.

Chart: Cumulative spot market cap trend for tokenized stocks (xStocks), Source:Talos CM Network Data Pro

Although the market cap of tokenized stocks has steadily climbed to approximately $3 billion, its spot trading volume remains small compared to corresponding equity perpetual contracts. Over the past week, equity perpetual contract volume on Hyperliquid was approximately $12.3 billion, equivalent to six times the total turnover of the tokenized stock spot market ($2 billion).

Chart: Daily average volume comparison for tokenized stocks over 7 days (Hyperliquid HIP-3 equity perpetual contracts vs. stock spot), Source:Talos CM Market Data Pro

In this edition of the "State of the Network" research report, we will deeply analyze the pricing performance of tokenized stocks relative to their underlying equities, intraday spread dynamics during after-hours and pre-market sessions, and the market fragmentation of these assets across different trading platforms.

Comparison of Tokenized Stock and Underlying Stock Performance

Tokenized stocks are typically issued for non-US investors, providing on-chain price exposure to US equities outside of regular market hours. These tokens are issued by specific entities and are generally backed by underlying stocks through licensed custodians.

Because tokenized stocks trade in a 24/7 year-round market, continuous on-chain quoting is possible even when traditional financial markets are closed.

Chart: 24-hour reference price comparison of NVDAx and NVDAon against NVDA underlying stock 5-minute price movement (gray shading indicates weekends or US market holidays/closures), Source:Talos CM Market Data Pro, Yahoo Finance

Currently, only whitelisted specific users qualify to mint new tokens and redeem underlying stocks. This mechanism exposes ordinary token holders without equal redemption rights to potential redemption discount/premium risks.

Between October 27 and November 2, NVDAx issued by xStocks traded at a premium for 58.6% of US market opening hours, averaging a 3.1 basis point premium; meanwhile, NVDAon issued by Ondo Stocks traded at a premium for 88.9% of opening hours, with an average premium reaching 14.1 basis points. During broad market sell-offs, tokenized NVDA also fell below its closing price from the previous Friday.

Chart: Comparison of average premium rates of NVDAx relative to the NVDA underlying stock across trading platforms (US market opening hours vs. after-hours), Source:Talos CM Market Data Pro

The decentralization of trading platforms and fragmented liquidity cause price variations for the same tokenized stock across different markets during both opening and after-hours periods. During the statistical period, NVDAx on Kraken recorded an average slight discount of -0.9 basis points relative to the NVDA underlying stock during US market opening hours.

Through Lead-Lag analysis, we compared the correlation between NVDAx price fluctuations during US market closures and the next day's NVDA actual opening gap over 210 trading days since January 2026, focusing on the following time window calculations:

After-hours session (ET 16:00 to 20:00): Token fluctuations in the evening show extremely weak correlation with the actual opening position of NVDA the next morning, with an R² fit of only 0.24. Shortly after the US market closes, the token itself carries almost no effective forward-looking signal regarding the direction of gains or losses the next day.

Overnight session (ET 20:00 to 04:00 next day): The correlation between overnight fluctuations and the next morning's opening price is similarly weak, with an R² of 0.30, though slightly stronger than the after-hours session, with correlation gradually increasing as it approaches the US pre-market open.

Pre-market session (ET 04:00 to 09:29): NVDAx prices steadily converge toward NVDA's final actual opening price, with the linear regression R² fit climbing further to 0.42.

Full overnight cycle (ET 16:00 to 09:30 next day): Across the entire US market closure period, token price movements are almost perfectly synchronized with the underlying stock, achieving an R² fit of up to 0.98. By the time the US market opens, NVDAx has fully reverted to the true price of the underlying stock as expected.

The tightness between tokenized prices and real stocks at the opening node is extremely high, but this price convergence is primarily completed gradually during the pre-market session. Relying solely on after-hours volatility makes it nearly impossible to effectively predict the next day's actual opening performance.

Chart: Correlation regression distribution of NVDAx during after-hours, overnight, pre-market, and near-open periods against NVDA's actual opening gap, Source:Talos CM Market Data Pro, Yahoo Finance

Tokenized stocks are highly sensitive to trade-induced shocks, a phenomenon particularly evident during US market closures. When the US market is closed, order book liquidity within a 1% mid-price spread range for NVDAx markets on Bybit and MEXC declines by an average of 19%; meanwhile, Hyperliquid sees its order book depth within the same 1% mid-price range shrink by approximately 33%.

Buy and sell orders executed under this sudden liquidity drop further exacerbate pricing mismatches of the same NVDAx token across different markets, which can only be gradually dampened once the underlying US stock resumes normal trading.

Chart: Average buy-side slippage comparison for executing $10,000 and $100,000 sized orders of NVDAx across trading platforms (log scale, gray area indicates after-hours session), Source:Talos CM Market Data Pro

When Do Tokenized Stocks Typically Trade?

Tokenized stocks provide non-US investors with a channel to access US stock prices. By observing trading distribution across different times of the day, we can reveal whether investors continue to participate in trading during after-hours periods with weaker liquidity.

Chart: Heatmap of NVDAx turnover statistics by day of the week and ET hour across 7 spot markets (median $36,000, peak in first opening hour $225,000), Source:Talos CM Market Data Pro

Statistical results show that NVDA xStocks trading activity is primarily concentrated around US market opening and closing times. During the core opening period on weekdays from 09:00 to 10:00 ET, NVDAx achieves an average hourly turnover of $225,000.

There is also significant trading activity during pre- and post-market sessions driven by opening expectations and after-hours breaking news. Although existing data cannot pinpoint the exact geographic location of traders, the offshore exchange landscape supporting these trading pairs indicates that despite dominant participants being international non-US users, their actual trading schedules closely follow domestic US market hours.

Cross-Platform Execution Costs and Performance

Because the price and depth of the same tokenized stock vary across different trading platforms, the actual transaction cost for investors to gain risk exposure largely depends on where the order is executed.

Chart: Distribution of xStock trading platforms with the lowest hourly cost for different order sizes ($1,000, $10,000, $100,000) over the past 7 days (weekend shading indicates US market closures), Source:Talos CM Market Data Pro

For small order sizes, there is a relatively diverse pool of platforms capable of executing trades. Between October 27 and November 2, for $1,000-sized orders, Bybit and Hyperliquid ranked as the platform with the lowest slippage for 58% and 23% of the time respectively; when order sizes increase to $100,000, Bybit or Hyperliquid captures the highest proportion of optimal execution platform time at up to 98.2%.

Chart: Hourly optimal execution costs (bps) for xStock, Ondo, and bStock at a $10,000 order size and comparison of winning issuers, Source:Talos CM Market Data Pro

A horizontal comparison of different tokenized variants reveals that asset execution costs also diverge depending on the issuing entity. Within the statistical period from October 27 to November 2, for $10,000-sized trades, NVDAb on Binance recorded the lowest execution slippage among all NVDA tokenized variants, ranking as the most cost-effective in 96% of observed periods.

Does Spread Arbitrage Space Still Exist After Deducting Trading Costs?

Although quote differences between various trading platforms for different tokenized stock variants appear to create arbitrage opportunities, actual strategy execution must fully account for liquidity constraints and transaction friction costs.

Statistical data from June 2026 to November 2 shows that even with a relatively small order size of $1,000, and before deducting platform trading fees, the sum of average market bid-ask spreads and two-way entry/exit slippage already completely exceeds the observed cross-platform spread magnitude.

Chart: Spread return vs. entry/exit slippage under $1,000 notional principal (Left: Long Hyperliquid NVDAx / Short Binance NVDAb across issuers; Right: Long Hyperliquid NVDAx / Short Kraken NVDAx across exchanges), Source:Talos CM Market Data Pro

Conclusion

Tokenized stocks have built a 24/7 on-chain pricing channel for traditional fixed-hour equity assets for non-US investors. However, at the current stage of development, these tokenized assets remain highly concentrated in trading during traditional US market opening hours, making them highly susceptible to significant liquidity contraction shocks once after-hours begin.

Systematically collecting order book depth, execution slippage, and reference pricing data for tokenized assets helps trading teams more objectively and precisely horizontally evaluate the true execution costs across different stock token variants and trading venues.