SPYx is a token on Solana that stands in for a share of SPY, the big fund that tracks the S&P 500. If you swap dollars for SPYx, is the price you get close to the price of SPY itself? We watched it every minute for 29 days. The answer: yes, while US stock markets were open. The typical gap on a $10,000 swap was about 0.03% from SPY's price.
A token like SPYx is only useful as a stand-in if its price follows the real thing. A big, lasting gap would mean paying more, or getting less, than the fund is worth. We wanted to see how big the gap actually is, and whether it grows when the stock market is closed. Stocks sleep at night and on weekends. Solana does not. So SPYx keeps trading when there is no SPY price to compare with.
For scale, a gap of 0.03% on a $760 share is about 23 cents.
This covers one token and one swap size, over 29 days. It says nothing about how SPYx behaves in a sudden market shock.
All rows: SPYx, $10,000 quotes, 2026-09-02 to 2026-09-30 (UTC). Gaps are in basis points (bp). 1 basis point = 0.01%. The "gap" compares the middle of the buy and sell prices with the reference. A median is the middle value: half the minutes had a smaller gap and half a larger one. The 95th percentile is the gap that 95% of minutes stayed within.
| Measure | Value |
|---|---|
| Median gap | 2.9 bp (about 0.03%) |
| 95th percentile gap | 7.7 bp (about 0.08%) |
| Median, with sign | +1.5 bp (token above SPY) |
| Largest gap | 1,445.8 bp, one bad quote |
| Minutes more than 100 bp (1%) away | 5 of 7,614 |
The largest gap is not a market move. In that minute the buy leg was priced at 761.63 and the sell leg at 540.56, while SPY was at 761.14. One route was broken. We kept it in the figures.
The 2.9 bp above is one median over the whole window. Taken day by day, the median moved between 1.6 bp and 4.1 bp (about 0.02% to 0.04%).
The typical gap stayed small every day. It ranged from 1.6 bp (0.016%) on September 28 to 4.1 bp (0.041%) on September 18, across 20 trading days.
| Day (UTC, 2026) | Typical gap on the day (bp) |
|---|---|
| Sep 2 | 2.1 |
| Sep 3 | 2.5 |
| Sep 4 | 2.3 |
| Sep 8 | 3.9 |
| Sep 9 | 3.8 |
| Sep 10 | 3.8 |
| Sep 11 | 3.6 |
| Sep 14 | 4.1 |
| Sep 15 | 3.5 |
| Sep 16 | 3.1 |
| Sep 17 | 3.3 |
| Sep 18 | 4.1 |
| Sep 21 | 3.5 |
| Sep 22 | 3.7 |
| Sep 23 | 2.1 |
| Sep 24 | 2.2 |
| Sep 25 | 3.3 |
| Sep 28 | 1.6 |
| Sep 29 | 1.9 |
| Sep 30 | 2.2 |
| Period | Buy above sell, median |
|---|---|
| Market hours | 5.2 bp |
| Weekday nights | 5.1 bp |
| Weekends and holidays | 5.0 bp |
At $10,000 the buy-sell spread was about the same in market hours, overnight and at weekends.
The buy price sits about 5 bp (0.05%) above the sell price in every period: 5.2 bp in market hours, 5.1 bp on weekday nights and 5.0 bp on weekends and holidays. The gap does not widen when the stock market is shut.
| Period | Typical gap (median) (bp) |
|---|---|
| Market hours | 5.2 |
| Weekday nights | 5.1 |
| Weekends and holidays | 5.0 |
"Futures-implied SPY" is our estimate of what SPY would be worth right now, using SPY's last closing price and how far S&P 500 futures have moved since.
| Period | Median gap | 95th percentile |
|---|---|---|
| Weekday nights | 4.8 bp | 13.3 bp |
| Weekends and holidays | 7.4 bp | 16.9 bp |
The gap is smallest in market hours and a little larger when the stock market is shut. The typical gap is 2.9 bp in market hours, 4.8 bp on weekday nights and 7.4 bp on weekends and holidays. The gap that 95% of minutes stayed within rises from 7.7 to 13.3 to 16.9 bp.
| Period | Typical gap (median) (bp) | 95% of minutes within (bp) |
|---|---|---|
| Market hours | 2.9 | 7.7 |
| Weekday nights | 4.8 | 13.3 |
| Weekends and holidays | 7.4 | 16.9 |
| Period | Median gap | 95th percentile |
|---|---|---|
| Weekday nights | 10.8 bp | 56.8 bp |
| Weekends and holidays | 28.1 bp | 60.1 bp |
These two rows are not a tracking error. SPY did not trade in those minutes, so its last close is hours or days old. The gap is mostly what the market did since. This is why the futures are the better reference off-hours.
Example (made-up prices). Friday SPY closes at 760. Over the weekend the token trades at 763. Against Friday's close the gap is about 0.39%. If futures say the market rose 0.4% by then, SPY would be worth about 763 and the real gap is near zero.
40,770 of the 41,760 minutes in the window (97.6%) have a usable quote pair. The longest gap was 29.3 minutes, on Sunday 2026-09-13.
Once a minute we asked a swap service for two live prices on SPYx: one to buy with $10,000 of USDC (a dollar-linked token) and one to sell back about $10,000 of tokens. These are quotes, not filled orders. We call the average of the two per-share prices the mid.
Every figure above is for the $10,000 quote size. A smaller or larger swap would see a different price.
We compared the mid with:
Only SPYx was measured. We have quotes for a second tokenized stock, but no matching real-price series for it.
Window. 2026-09-02 00:00 UTC to 2026-10-01 00:00 UTC. Quotes start on 2026-09-01. SPY and futures one-minute prices cover the same window. Hourly SPY prices go back to 2024-09, but the quotes do not, so the window is 29 days, not two years.
Price per share. SPYx is a Token-2022 token (a newer Solana token standard) with a scaled-UI multiplier that rises as dividends accrue. We divide each quote by the multiplier recorded with it. Without this the token looks like it trades at a premium.
Reference timing. A quote is taken about 23 seconds into a minute. We compare it with SPY's price at the start of that minute, so the reference can be up to 23 seconds old.
What counts as market hours. A SPY one-minute price exists for that minute. Labor Day (2026-09-07) therefore counts as off-hours.
Off-hours reference. The futures-implied value is SPY's last close times the futures move since the same minute. The futures-to-cash difference drifts, and a dividend between the two dates would add to the gap. We did not measure that error. Minutes when futures are not trading are dropped, which removes most of the weekend from the futures rows.
Dropped quotes. A minute needs both a buy and a sell quote. 499 minutes (about 1.2% of the window) had only one of the two, or a leg with no route, and we dropped them. In all, 2.4% of minutes have no usable pair.
Outliers. Single-leg bad quotes, like the one above, are in every figure. They move the maximum and the count over 100 bp, not the median.
What this does not tell you.
These figures are our own measurement over this window. You cannot rerun them from public data, so treat them as a report of what we saw, not a result you can reproduce.