21 Comments

Summary:

The U.S. may have the most developed LTE infrastructure of any country in the world, but it’s networks are no means the fastest. Seven other countries scored higher in an OpenSignal study of 4G speeds.

Sweden was the first country to launch an LTE network, and it retains plenty of bragging rights. According to a study by U.K. network-testing firm OpenSignal, Sweden has the fastest 4G networks in the world, averaging download speeds of 22.1 Mbps.

The U.S. was the second country to deliver commercial LTE networks on the world stage, but it ranks far lower in terms of 4G bandwidth delivered. OpenSignal found that networks in Hong Kong, Denmark, Canada, Australia, South Korea and Germany all performed better. The U.S. placed eighth, averaging downlink speeds of 9.6 Mbps.

OpenSignal global LTE speeds

Why the low scores? It probably has to do with the configuration of U.S. carriers’ networks. While most operators around the world secured 40 MHz of spectrum with which to launch their new 4G networks, U.S. carriers have been working with smaller swatches of airwaves. Verizon and AT&T are using 20 MHz for their initial rollouts, while Sprint and MetroPCS are dealing with as little as 10 MHz. If you’re working with half the spectrum, your connections will sport half the bandwidth.

Based in both London and Laguna Hills, Calif., OpenSignal collects its data through crowdsourcing, aggregating measurements recorded by millions of smartphones users who have downloaded its free Android app. There are a lot of similarities between the OpenSignal and Seattle’s RootMetrics. Root supplements its smartphone data with professional testing (see our video on one such drive test in Chicago), while OpenSignal relies entirely on crowdsourcing, but both have started generating very detailed maps of cellular network coverage and performance in different areas of the world. OpenSignal recently expanded its scope to encompass Wi-Fi.

What’s particularly noteworthy about OpenSignal’s latest report is just how far LTE has penetrated around the world in the last two years. OpenSignal tracked LTE signals in 62 countries, including multiple African countries and in the central Asian nations of Uzbekistan and Kyrgyzstan.

LTE image courtesy of Shutterstock user Inq

  1. The numbers from Sweden is something i can agree with to some extent.
    Alot of geographica questions “where” u can obtain 22mb+ 4G still remains..
    There are still alot of places where u can´t even get std. “GSM” coverage.

    Share
  2. Gee, Kevin.. Why does it feel like you first comment that I don’t know what I am talking about then the very next day, rip off my blog….

    Share
    1. I think you an overly high opinion of your influence, David.

      And on the White Spaces story, I never implied you didn’t know what you were talking about. I was objecting to the fact you were lumping my post in with the hype surrounding Super Wi-Fi.

      Share
      1. Kevin, it is not about my influence. It is about knowing what one is talking about. Having influence without real knowledge and understanding misleads your readers. I would categorize your post on Super Wifi with the Kang WP article and the 2011 Economist article on White Space which lacked any understanding of the underlying political and technical facts. Your Sweden post was first tweeted by Professor Susan P. Crawford who definitely knows what she is talking about: telecom monopoly.

        Share
  3. Many of us measured more than 22 Mbps on LTE on Verizon’s network shortly after launch. Clearly they are both throttling and have a much larger LTE subscriber base which will naturally reduce the average data rate through crowd sourced data per sector.

    Share
    1. I don’t think it’s throttling, just the effect of more and more users on the same tower.
      My t-mobile HSPA+ actually outpreforms Verizon LTE at this point.

      Share
  4. Sweden = 22Mbps at $2/GB. USA = 9Mbps at $6/GB. Enough said.

    Share
    1. Where do you come up with these numbers? I’m pretty sure I’m paying $10/GB at 9MBps in the US. :)

      Share
  5. OpenSignal only works on android hence the sample taken is always very limited.

    Share
    1. Stay tuned for iOS application… coming very, very soon!

      It also worth mentioning our sample sizes are usually very large as our app has been downloaded over 3 million times.

      Share
    2. Hi Romk1N,

      I see what you’re getting at, but I think in this case — measuring 4G speeds — that kind of device diversity doesn’t apply. Only the iPhone 5 supports LTE, so any measurements from an iPhone would go back only six months. Also the iPhone 5 bands only supports about half of the networks OpenSignal tested. I admit that the iPhone app would definitely have added a lot more data points to the 3G metrics, but it seems that for 4G measurements it would have only produced more results in a few countries — and a limited number at that.

      Share
      1. Oh my, another Apple Head. The iPhone 5 was LATE to the LTE party! There are a TON of phones on the market that have LTE ability.

        Share
  6. I get around 20MBs all the time on my AT&T LTE phone…

    Share
  7. Reblogged this on mayo615 and commented:
    Further evidence that Yale Law Professor Susan P. Crawford is right about a telecom monopoly in North America that is throttling the Internet and endangering our economic competitiveness. Read Ms. Crawford’s book, “Captive Audience: The Telecom Industry and Monopoly Power in the New Gilded Age.” Professor Crawford is also being promoted to succeed Julius Genachowski, as FCC Chairman. I wholeheartedly endorse her.

    Share
    1. Richard Bennett Thursday, February 14, 2013

      The footnotes in Crawford’s book mostly refer to blog posts and newspaper articles. Curiously, her claim that the U. S. ranks 22nd in download speed is supported by a blog post from 2010 that referred to a 2009 Akamai report.

      The current Akamai report (3q 2012) says the US is 8th in that same category, but Crawford hasn’t seen that in a blog yet.

      She’s a nice, sincere, charming person but poorly informed.

      Share
      1. 22nd or 8th seems like mice nuts to me. The point remains that it is a significant competitive and political problem in North America as a whole. Obama’s SOTU point about decaying infrastructure is alive and well here.

        Having worked with the telco mentality for years, the point about a new Gilded Age of Monopoly is at the heart of the problem.

        http://mayo615.com/2013/02/12/why-net-neutrality-is-so-important-the-telecom-industry-and-monopoly-power/

        or

        http://mayo615.com/2013/02/08/urban-legend-of-free-wi-fi-for-the-masses-devil-in-the-details/

        BTW, the Crawford book was likely already in galley proofs by the time the most recent Akamai report was released.

        Share
      2. Richard Bennett Thursday, February 14, 2013

        Crawford’s book wasn’t in galley proofs before the 2010, 2011 and early 2012 Akamai data came out David.

        The story she wanted to tell conflicts with the facts, so deal with it. Check her Chapter 12, footnote 32:

        “…And the Economist pointed out: “Having led the world in internet access, America has slipped over the past decade to 22nd (behind Latvia and the Czech Republic) with an average download speed of a mere 3.8 megabits per second (Mbps) compared with South Korea’s average of 14.6Mbps…” (“ The Difference Engine: Politics and the Web,” December 24, 2010).

        That blog post uses Akamai’s 4Q 2009 report. My point is that American broadband speeds have risen in the rankings every year since then, so Crawford’s use of primary source data is utterly dishonest.

        Share
      3. where in the Akamai Q3 2012 report does it say the U.S. ranks 8th? I cannot find it.

        Share
  8. Richard Bennett Thursday, February 14, 2013

    Kevin’s analysis the reasons for the range of speeds available is mostly right: U. S. carriers are spectrum-constrained because the total pool of sub-3GHz spectrum is dominated by military uses. By most estimates, 60% of this spectrum is in government hands. With bits/sec/hertz roughly equal for all LTE installations and fiber backhaul in place, the speed of one network vs. another comes down to spectrum and tower density.

    If PCAST had done its job, we’d be freeing some of this military spectrum for civilian use, but alas, all we’re going to get is some weak for of on and off sharing.

    Share
    1. Well, at least I’m “mostly right”. :)

      Thanks for commenting, Richard

      Share
  9. Richard Bennett Thursday, February 14, 2013

    Personally, I don’t set much stock by crowd-sourced apps that purport to measure network speed. There are too many variables involved in this approach for it to be anywhere close to accurate, mainly on the handset end. A user with the latest quad-core spiffy Android phone is going to record higher speeds on the same network as somebody with an older model, so there’s clearly a lot more going on here than network quality. In the Akamai mobile broadband measurements, some networks are pushing 40 Mbps, but that’s not a national average, just one provider.

    There’s anecdotal evidence – some press reports – of cord-cutting in Japan from 100 Mbps FTTB to all-in for LTE, despite the allegedly low speed of LTE in that country.

    Most analysts are still missing the point of LTE in any case. It’s value proposition is low latency IP over a RAN that can knit channels together out of disaggregated spectrum; it’s not just a faster radio, it’s a better system for IP that leads to disengagement from circuit switched voice.

    What we need are some Akamai and SamKnows-style measurements that separate fast networks from slow handsets, and this batch doesn’t come close to the granularity that’s meaningful for comparative study.

    Share

Comments have been disabled for this post