Without more details on the test setup, it's hard to evaluate it. 5G can be deployed with much more spectrum than 4G can be. Deployed with 60-80MHz of spectru

Without more details on the test setup, it's hard to evaluate it. 5G can be deployed with much more spectrum than 4G can be. Deployed with 60-80MHz of spectrum, US carriers have shown off LTE north of 700Mbps. The article said that they were using the 3.5GHz band, but they didn't say how much spectrum they were using which makes the comparison incredibly suspect. Do an image search for "T-Mobile gigabit" or "Sprint gigabit" and you'll see plenty of tests in areas where the companies have a lot of spectrum in a small area doing very high speed. AT&T and Verizon are planning on launching 5G with hundreds of MHz of spectrum in millimeter wave bands. That spectrum won't travel far and will have difficulty going through walls, but will have lots of capacity given how much they have. By contrast, LTE deployments are usually in the range of 10-80MHz. Given the low 50Mbps speed of their 4G test, it seems like they didn't use much spectrum at all for 4G. If they had a reasonable LTE setup with 80MHz of spectrum in 4xCA, 256QAM, and 4x4 MIMO, they should easily beat 500Mbps and probably get closer to 800Mbps. The low LTE speed makes me conclude that they're just looking to promote Elisa (the wireless operator that worked with the newspaper to do the test) and set it up to just make for a good article. What I want is a comparison between LTE and 5G NR that uses the same amount of spectrum in the same band. 5G NR is likely to have new applications with millimeter wave and very high order MIMO, but that won't work at the lower frequencies that are used to create broad coverage. In the 500-2,000MHz range with the same amount of spectrum, how much faster is 5G NR? Now, 5G NR is said to have some great benefits in terms of latency and power use for low-bandwidth devices on top of bandwidth increases, but people aren't showing off real comparisons yet. Also, just to see how bad a test of LTE this was, average speeds at the super bowl were over 50Mbps for three carriers: http://bgr.com/2018/02/05/t-mobile-vs-verizon-fastest-network-super-bowl-2018/ http://bgr.com/2018/02/05/t-mobile-vs-verizon-fastest-networ.... I mean, T-Mobile was able to top 120Mbps in a real-world situation. They couldn't beat 50Mbps on an unloaded LTE network? That just sounds like they gave it very little spectrum. Heck, Ookla notes that average nationwide speeds are 31.58Mbps for T-Mobile, 23.83Mbps for Sprint, 24.39Mbps for AT&T, and 28.21Mbps for Verizon. That's an average that includes places with bad coverage and such. 50Mbps is a very low LTE performance number in a lab which just makes it feel suspect. Now, it could be that 5G NR, like UMTS in its early days, has a lot more room for improvement. UMTS launched at 384kbps which wasn't extremely faster than EDGE and 1xRTT, but did have big latency benefits. A little later, it was pushed further and eventually got as high as 42Mbps (theoretical). Sometimes they kinda exaggerate the early tests because the technology is really better over the long run, but it will take some time to get there.