Over 107 days of 20-level order-book snapshots, xyz:NVDA quoted a median spread of 0.51 bps in US regular hours with $1.7M resting within 25 bps of mid. A $25k market order paid about 0.83 bps in regular hours and 1.53 bps on weekends, when depth within 25 bps fell to $533k. Taker fee is 0.90 bps with growth mode on; funding averaged +5.56% annualised over the last 30 days.
spread · rth
0.51 bps
median · p90 1.51 bps
depth ±25bps · rth
$1.7M
median, both sides
$25k slip · rth
0.83 bps
fills 100%
$25k slip · wknd
1.53 bps
fills 100%
taker fee
0.90 bps
fee scale 1.0 · growth mode on
funding apr · 30d
+5.56%
longs pay when positive · 7d +7.11%
vol 24h · 30d avg
$49.2M
notional
open interest
$157.8M
both sides · 2026-10-02
by sessionNew York time · medians over 107 days
session
snaps
spread
p90
top
depth ±10
depth ±25
depth ±50
$5k
$25k
$100k
$100k fills
$250k fills
US regular 9:30–16:00
4,897
0.51
1.51
$24k
$1.6M
$1.7M
$1.7M
0.56
0.83
1.39
100%
100%
US extended 4–9:30, 16–20
7,185
0.49
1.04
$9k
$1.0M
$1.1M
$1.1M
0.63
1.12
2.03
100%
94%
Overnight 20–4
4,962
0.50
1.34
$7k
$823k
$844k
$844k
0.66
1.24
2.26
100%
87%
Weekend / holiday
9,153
0.48
1.33
$5k
$507k
$533k
$533k
0.74
1.53
2.91
93%
36%
bps unless stated · RTH 9:30–16:00 · extended 4:00–9:30 and 16:00–20:00 · overnight 20:00–4:00 on weekdays · weekend runs Friday 20:00 to Monday 4:00 and includes NYSE holidays · slip = mean of the buy and sell side of a market order walking the visible 20 levels · fills = share of snapshots the size fits inside those levels · "—" = not fillable in the median snapshot
trade flowtaker side · 24 h to 2026-10-02 00:41 UTC · 30 d = last 30 complete UTC days · all markets
alert melarge orders in NVDA, $50k or more→ sign infree · by email · when the data is rebuilt
taker buy · 24h
$41.7M
30d $706.6M
taker sell · 24h
$49.0M
30d $713.4M
imbalance · 24h
-8.0%
of $90.7M · 30d -0.5% of $1.42B
trades · 24h
38,536
192 large orders · 30d 843,960
takers · 24h
1,117
291 makers · 1,228 addresses
takers · 30d
12,710
13,131 addresses incl. makers
top-10 share · 30d
56%
of $1.42B by 12,710 takers · 7 make half
concentration · 30d
0.112
Herfindahl over takers · 1 = one taker
daily net taker flow
USD per UTC day · taker buy − taker sell · up = net buying, down = net selling · 108 complete days · hover a day for its value · today (2026-10-02) is still open and not drawn: +$229k so far
by session · 30 d
session
taker buy
taker sell
net
imbalance
trades
share of notional
US regular 9:30–16:00
$331.2M
$330.0M
+$1.1M
+0.2%
304,682
47%
US extended 4–9:30, 16–20
$214.8M
$208.8M
+$6.0M
+1.4%
246,013
30%
Overnight 20–4
$93.6M
$106.6M
−$13.0M
-6.5%
139,826
14%
Weekend / holiday
$67.0M
$67.9M
−$991k
-0.7%
153,439
10%
sessions in New York time as in the liquidity table above · share = the session's taker notional over all four, $1.42B in 30 days
hour of week · taker notional
New York time · 30 days summed per hour · brighter = more volume · log scale between the 5th and 95th percentile · 168 of 168 hours with trades · hover a cell for USD and trades
large orders: fills grouped by transaction hash, taker and side, $50k or more per order; fills without a hash are not grouped · share = the address's taker notional over all $1.42B of 30-day taker notional (843,960 trades, 12,710 takers) · addresses are public on-chain data and link to the Hyperliquid explorer; an address can be a subaccount, vault, market maker or bot and is not labelled · taker notional is volume, not a change in open interest · rebuilt nightly and at 09:05 ET, not a live tape · raw: /api/xyz/flow/NVDA.json · definitions in methodology
alert meNVDA funding above or below a level I set→ sign infree · by email · when the data is rebuilt
last settlement
+5.48%
+0.0625 bps/h · 10-01 21:00 ET
next · estimate
+12.14%
+0.1385 bps/h · as of 10-01 22:15 ET
apr · 24h
+9.76%
24 of 24 hours observed
apr · 7d
+7.11%
157 of 168 hours observed
apr · 30d
+5.56%
665 of 720 hours · stdev 10.28%
paid by a long · 30d
42.20 bps
of notional, over 665 observed hours
hours above neutral · 30d
20%
at 70% · below 10% · negative 6%
range · 30d
+64.64%
highest hour · lowest -91.46%
daily funding apr
mean of the observed hourly rates per UTC day, annualised · up = longs paid, down = shorts paid · dashed = neutral 5.48% · 110 complete days · hover a day · today (2026-10-02) is still open and not drawn: +6.44% over 2 h so far
hour of week · mean funding apr
New York time · whole archive · the hour a settlement covers · 168 of 168 hours with data · hover a cell for its mean and n
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
Mon
Tue
Wed
Thu
Fri
Sat
Sun
shorts paylongs pay · full shade = ±13.75% APR or more · empty = no settlements
funding by session
session
apr · 30d
hours · 30d
above neutral · 30d
apr · all
hours · all
above neutral · all
US regular 9:30–16:00
+6.04%
117
15%
+9.50%
432
37%
US extended 4–9:30, 16–20
+7.56%
201
26%
+10.37%
727
41%
Overnight 20–4
+8.41%
134
33%
+10.66%
476
42%
Weekend / holiday
+1.61%
213
8%
+5.15%
862
19%
mean of the hourly rates settled in each session, annualised · 30d = the 720 hours ending at the last settlement, all = since 2026-06-14 · above neutral = share of those hours settled above 5.48% APR
largest settlements · 30 d
settled · ET
session
rate bps/h
APR
09-20 00:0009-20 04:00 UTC
weekend / holiday
-1.044
-91.46%
09-13 06:0009-13 10:00 UTC
weekend / holiday
-0.9935
-87.03%
09-27 19:0009-27 23:00 UTC
weekend / holiday
-0.7981
-69.91%
09-20 19:0009-20 23:00 UTC
weekend / holiday
+0.7379
+64.64%
09-07 19:0009-07 23:00 UTC
weekend / holiday
-0.5796
-50.77%
09-13 05:0009-13 09:00 UTC
weekend / holiday
-0.5342
-46.80%
09-04 21:0009-05 01:00 UTC
weekend / holiday
+0.5106
+44.73%
09-13 04:0009-13 08:00 UTC
weekend / holiday
-0.4995
-43.76%
09-13 07:0009-13 11:00 UTC
weekend / holiday
-0.4702
-41.19%
09-12 17:0009-12 21:00 UTC
weekend / holiday
-0.4503
-39.44%
positive = longs pay shorts · the neutral rate is the deployer's carry (base interest 0.01% per 8 h × funding multiplier 0.5), not crowd positioning · every mean is over observed hours; collector gaps are not filled · rebuilt nightly and at 09:05 ET, not live · raw: /api/xyz/funding/NVDA.json · definitions in methodology
slippage curve
median bps over mid vs order size · a line ends where the size stops filling in the median snapshot
hour of week
New York time · median per cell over 107 days · brighter = better · the breathing cell is now
weekly · spread
median bps, all sessions
weekly · depth ±25 bps
median USD both sides, all sessions
cost calculator
per round trip = 2 × median slippage + 2 × taker fee · funding uses the 30-day mean hourly rate, annualised, charged for a position held all day · slippage between tested sizes is interpolated on a log scale · ignores maker rebates, impact that persists after your fill, and fee tiers