São Paulo, Brazil → New York, USA RTT
ICMP RTT from São Paulo, Brazil to New York, USA: 122 ms average, 0% loss, stability details, and the route's rank among 19 measured routes.
🇧🇷 São Paulo, Brazil (GRU) → 🇺🇸 New York, USA (NYC)
Measurement round: 2026-08-15T04:03:13Z.
In the 2026-08-15T04:03:13Z ICMP echo round, 50 probes from São Paulo, Brazil to New York, USA returned an average RTT of 122 ms, with a tight spread from 119.57 ms to 129.74 ms. Loss was 0%, jitter was 1.53 ms, and the standard deviation was 1.93 ms, keeping the connection in the good latency tier despite the distance.
The 7,657.6 km great-circle distance has a vacuum floor of 51.09 ms and a fiber floor of 74.99 ms. The measured 122 ms is 1.63 times the fiber floor, leaving a fiber efficiency of 61.5%, so the connection carries noticeable excess delay over the theoretical best case.
This route ranked second among the 19 outbound routes in the round, and its variability, at 1.58% of the average RTT, is well below the route-set median of 3.08%. The combination of zero packet loss and very low jitter makes the São Paulo–New York result notably stable for a route of this length.
Latency Summary
| Metric | Value |
|---|---|
| RTT | 122 ms |
| Jitter | 1.53 ms |
| Packet Loss | 0% |
| Standard Fiber Floor | 74.99 ms |
| Fiber Efficiency | 61.5% |
| Latency Tier | Good |
| Source Region | South America |
| Destination Region | North America |
Ping Measurement Series
The chart below renders the 50-sample ICMP echo measurement round for this route; an accessible table of every sample is included with the chart.
Ping series: GRU → NYC
50 ICMP echo samples at 100 ms intervals · round 2026-08-15T04:03:13Z · avg 122.0 ± 0.27 ms (SE)
Sample distribution: box = interquartile range (120.7–122.5 ms), median 121.7 ms · whiskers = min–max · diamond = round average · each dot is one measured sample.
How to read this chart
The shaded band spans the measured minimum to maximum RTT of all samples collected up to that point, and the thin line traces the running average. The final position therefore matches the round statistics exactly — the band only widens when a new extreme sample arrives.
The dashed line is the round average, and the shaded band at the bottom is the theoretical fiber-optic floor (75.0 ms) for this route.
Full 50-sample dataset
| Seq | Offset (ms) | RTT (ms) |
|---|---|---|
| 1 | 0 | 120.43 |
| 2 | 100 | 120.35 |
| 3 | 200 | 128.32 |
| 4 | 300 | 124.53 |
| 5 | 400 | 122.59 |
| 6 | 500 | 122.27 |
| 7 | 600 | 123.48 |
| 8 | 700 | 121.20 |
| 9 | 800 | 120.70 |
| 10 | 900 | 120.81 |
| 11 | 1000 | 120.93 |
| 12 | 1100 | 122.86 |
| 13 | 1200 | 123.91 |
| 14 | 1300 | 123.91 |
| 15 | 1400 | 121.01 |
| 16 | 1500 | 120.94 |
| 17 | 1600 | 120.21 |
| 18 | 1700 | 120.81 |
| 19 | 1800 | 123.22 |
| 20 | 1900 | 122.41 |
| 21 | 2000 | 121.95 |
| 22 | 2100 | 120.32 |
| 23 | 2200 | 121.13 |
| 24 | 2300 | 120.00 |
| 25 | 2400 | 121.79 |
| 26 | 2500 | 120.45 |
| 27 | 2600 | 121.73 |
| 28 | 2700 | 120.74 |
| 29 | 2800 | 119.89 |
| 30 | 2900 | 119.57 |
| 31 | 3000 | 121.10 |
| 32 | 3100 | 120.62 |
| 33 | 3200 | 122.08 |
| 34 | 3300 | 122.23 |
| 35 | 3400 | 121.23 |
| 36 | 3500 | 120.20 |
| 37 | 3600 | 129.74 |
| 38 | 3700 | 124.50 |
| 39 | 3800 | 122.99 |
| 40 | 3900 | 122.30 |
| 41 | 4000 | 121.69 |
| 42 | 4100 | 121.07 |
| 43 | 4200 | 122.42 |
| 44 | 4300 | 120.53 |
| 45 | 4400 | 121.82 |
| 46 | 4500 | 122.33 |
| 47 | 4600 | 120.49 |
| 48 | 4700 | 122.07 |
| 49 | 4800 | 124.50 |
| 50 | 4900 | 123.63 |
Download This Route's Data
This route's measurement round is published as open data (CC BY 4.0) in CSV, JSON, and YAML. See About This Measurement for citation guidance.
Download this dataset
Per-route ICMP ping series, released under CC BY 4.0. Uncompressed plain text — no decompression step needed.
Theoretical Fiber Latency
The public physical reference uses the GeoNames city centres for São Paulo (-23.54750, -46.63611) and New York City (40.71427, -74.00597). It does not disclose either facility address.
| Physical Reference | Value |
|---|---|
| WGS-84 geodesic distance | 7,657.6 km |
| Vacuum RTT floor | 51.09 ms |
| Standard fiber RTT floor () | 74.99 ms |
| Low-latency fiber material floor | 74.69 ms |
| Engineering floor (5% path allowance) | 78.75 ms |
| Research 1.33× mapped-fiber reference | 99.74 ms |
| Estimated unamplified path loss | 1608.1 dB |
| Transparent optical spans / inline amplifiers | 101 / 100 |
| Published RTT inflation over fiber floor | 1.63× |
Attenuation is used to estimate the number of 80 km optical spans; it does not directly slow light. See the full theoretical fiber latency, GeoNames, attenuation, and amplifier methodology.
Stability & Trend
The current measurement round (2026-08-15T04:03:13Z) collected 50 ICMP echo samples on this route. Distribution statistics from that round:
| Round Statistic | Value |
|---|---|
| Minimum RTT | 119.57 ms |
| Average RTT | 122 ms |
| Maximum RTT | 129.74 ms |
| Standard deviation | 1.93 ms |
| Stdev / average | 1.6% |
RTT trend
Estimated route stability across the selected time window.
- Average
- 121.9 ms
- Minimum
- 120.9 ms
- Maximum
- 123.3 ms
- Average loss
- 0.01%
Route Context
- Departure PoP: São Paulo (GRU)
- Destination PoP: New York (NYC)
- Region overview: South America → North America
- Full global matrix: Backbone Latency Matrix
- Physical methodology: Theoretical Fiber Latency
Related City-Pair Routes
Reverse direction
- New York to São Paulo latency and RTT — 108.3 ms
Fastest routes departing São Paulo (GRU)
- São Paulo to Miami latency and RTT — 74 ms
- São Paulo to Ashburn latency and RTT — 128 ms
- São Paulo to Los Angeles latency and RTT — 130.8 ms
- São Paulo to Seattle latency and RTT — 156 ms
- São Paulo to London latency and RTT — 174.7 ms
Fastest routes arriving at New York (NYC)
- Ashburn to New York latency and RTT — 6 ms
- Miami to New York latency and RTT — 34 ms
- Los Angeles to New York latency and RTT — 59 ms
- Seattle to New York latency and RTT — 59 ms
- London to New York latency and RTT — 63.5 ms
About This Measurement
Published RTT reflects best-case backbone path behavior measured with ICMP echo probes between PoPs. End-user latency also depends on local access networks, congestion, routing policy, and traffic engineering. None of these values is an SLA.
The measurement round behind this page is published as open data under CC BY 4.0 in the Hats Network latency dataset. Per-pair files are available at /opendata/latency/latest/pairs/gru-nyc.pings.{csv,json,yaml}. When citing, include the route, the measurement round timestamp, and the dataset version — for example: "Hats Network latency dataset, São Paulo → New York round 2026-08-15T04:03:13Z, CC BY 4.0".
Data auto-generated on August 15, 2026. Explore the full interactive latency matrix or browse more city-pair routes.