Measured internet performance in Q2 2026
This research summary compares country aggregates from the immutable 2026-Q2 release. The editorial rankings require at least 500 represented measurements per country, while the downloadable dataset retains every eligible measured record.
What the release can show
The release enables consistent comparison of available country-level monthly aggregates across April–June 2026. Values are weighted by represented monthly sample counts, so months with more observations contribute proportionally more to the quarterly value.
The comparison cohort contains 20 countries with at least 500 represented measurements and covers 95,701 measurements. This editorial threshold limits small-sample ranking noise; it does not turn the underlying sample into a population census.
Download capacity and latency tell different stories
Higher download values indicate more measured throughput; lower latency values indicate a more responsive measured path. Read both with sample size and the limitations below.
Highest download aggregates
#1Singapore
277.55 Mbps · 719 represented measurements
#2Canada
208.50 Mbps · 3,366 represented measurements
#3Hong Kong
181.37 Mbps · 580 represented measurements
#4France
175.10 Mbps · 540 represented measurements
#5United States
173.94 Mbps · 70,784 represented measurements
#6Germany
171.20 Mbps · 621 represented measurements
#7Spain
169.82 Mbps · 624 represented measurements
#8United Kingdom
165.77 Mbps · 3,500 represented measurements
#9Australia
151.63 Mbps · 1,340 represented measurements
#10Mexico
125.02 Mbps · 515 represented measurements
Lowest latency aggregates
#1Singapore
22.84 ms · 719 represented measurements
#2Canada
29.58 ms · 3,366 represented measurements
#3Germany
29.67 ms · 621 represented measurements
#4United Kingdom
30.17 ms · 3,500 represented measurements
#5France
31.84 ms · 540 represented measurements
#6Indonesia
33.43 ms · 2,545 represented measurements
#7Australia
33.83 ms · 1,340 represented measurements
#8Spain
36.06 ms · 624 represented measurements
#9Italy
42.06 ms · 821 represented measurements
#10United States
43.90 ms · 70,784 represented measurements
Calculation and verification
quarterly_metric = Σ(monthly_metric × monthly_sample_count) / Σ(monthly_sample_count)Download the JSON or CSV, filter records where sample_count >= 500, recompute the rankings, and verify the files with the published SHA-256 digest. Dataset bytes are frozen when the release is created, so later report backfills do not silently change this version.
What must not be inferred
- Countries and months have unequal sample coverage.
- A weighted mean of monthly aggregates is not a raw-test median or percentile.
- Device, Wi-Fi, server distance and access technology can affect measurements.
- The comparison threshold reduces ranking noise but does not guarantee representativeness.
- The dataset excludes synthetic estimates but does not claim universal population representativeness.
- FCC provider-reported availability is a different dataset and is not merged into this release.