The most recent average US download speed from Ookla Speedtest for October 2014 is 30.70Mbps. In October 2013 the average speed was just 19.67Mbps. Globally the average download speed is 21.03Mbps up from 15.61Mbps in October 2013. In 2012 the average speed was 13.17Mbps down globally and 15.42Mbps in the US.
Clearly 2014 has seen major improvements to the worlds internet backbone. With just a 2.44Mbps average increase between 2012 and 2013 but a 5.61Mbps increase in 2014. The United States seams to be leading the charge to faster speed with a increase of 11.03Mbps in just one year.
If you really want speed in the United States here are a few places and companies to check out.
The top ten states for internet speeds are:
- New York 39.24Mbps
- Delaware 38.92Mbps
- New Jersey 38.86Mbps
- Maryland 36.92Mbps
- Missouri 36.62Mbps
- California 35.36Mbps
- RoadIsland 35.24Mbps
- Virginia 35.09Mbps
- Massachusetts 34.90Mbps
- Nevada 34.55Mbps
Coming in at the bottom of the list are Kentucky with 15.87Mbps and Maine with 14.86Mbps.
If you want the fastest possible internet in the US the top 5 cities to live in are: (Note: Ookla did not include Google Fiber in this list.)
- Kansas City, MO 68.59Mbps
- Austin, TX 68.27Mbps
- Huntington Beach, CA 50.52Mbps
- Flushing, NY 49.83Mbps
- New York, NY 48.24Mbps
The top 5 major ISPs in the US are:
- Google Fiber 230.69Mbps
- Verizon FIOS 42.27Mbps
- Cox 39.42Mbps
- Comcast 38.70Mbps
- Charter Communications 38.45Mbps
These numbers come from millions of test run on Ookla websites every day. They averaged out the tests from the last 30 days. They also only count tests that take place with in 300 miles of the client and the host. You can view the speed test results here.
More about the numbers in this report from Ookla.
An index is traditionally defined as a numerical scale used to compare variables with one another or with some reference number. For purposes of the NetIndex, Ookla defines an index as a weighted average of data collected over the 30 most recent days.
To calculate an index, Ookla first ensures that distance and infrastructure bottlenecks have a minimal impact on accuracy. To do this, we track the distance between the test location and the Ookla Speedtest server. Thanks to the breadth of our infrastructure, we have a server within 300 miles for the vast majority of the world population.
To determine the averages for broadband download and upload, we first average one hour’s worth of test results for each unique IP to get the IP Averages. Next, we average all of the IP Averages for one hour to determine the Hourly Average. From there, we average all of the Hourly Averages for one day to find the Daily Average. Finally, we average all of the Daily Averages for up to 30 days to get the final value.
With mobile download and upload, the averages are based on one day’s worth of tests from each device to first determine the Device Averages, which is then averaged to determine the Daily Average. We then average the Daily Averages for up to 30 days to determine the final value.
Nightly, we review 24-hour increments until we identify 30 days of data with acceptable parameters. To ensure the index value is current, we do not go back further than six months to find those 30 days of data used to compute the final index value. We ignore days where the average distance is more than 300 miles to ensure events, such as server downtime, do not affect the aggregated number.