"Technology on the market can deliver 10X to 25X at reasonable cost" I wrote in 2017. In 2022, I can see another 8X to 15X improvement from technologies already on the market. Traffic growth is all the way down. It was only 15% across 38 OECD countries in 2021.
Verizon CTO Kyle Malady reported that capacity margin was doubling even before putting mid-band spectrum to use and 5G. Putting mid-band to use with Massive MIMO is typically raising the capacity per cell by 3-7X.
Massive over-capacity is why nearly all telcos are offering fixed wireless and seeking to move customers to unlimited plans. Investors focus on sales, particularly postpaid, knowing that more customers can be accommodated
Outside of China and Korea, few telcos have reached even 50% with mid-band spectrum and massive MIMO. 5G carrier aggregation is just coming out of the labs. Telcos will put to use fallow lower frequencies. Moore's Law is continuing to bring down the cost of switches and routers. Automation and cloudification are just starting to pay off.
Here is what I wrote in 2017. It was on target, although I should have emphasized mid-band spectrum as well.
Telcos report costs going down 40% to 60%. That allows T-Mobile U.S. & Sprint to offer "unlimited" with only a few gotchas. Competition and regulation will determine who is actually served.
Technology on the market can deliver 10X to 25X at reasonable cost; Weak competition or weak regulation could hold this back. Speeds over 50 megabits with a cap high enough to watch 100 hours/month of HD TV can be delivered in most of the developed world. The engineers can deliver. Marconi Fellow Paulraj tells me Massive MIMO will bring many of the same benefits to most rural areas, including in emerging nations. Extreme rural areas - the last 1-3% - may not be as fortunate.
Giga LTE offers about a gigabit to the cell site. Beginning this year, phones with a good connection will get 100-300 megabits down. Phones on the cell edge and behind walls are seeing a 4X improvement as well. Gig LTE combines more antennas (4x4 MIMO,) more spectrum (60-80 MHz,) and more bits squeezed in (256 QAM.) Gig LTE is five or six times as much as LTE started at and about three times the capacity of most advanced systems in 2016. The meter in the picture showing about one gig at Huawei's BBF in Tokyo Nov 2016 is Gig LTE. Two gig has been demonstrated by British Telecom. Explained for non-engineers here.
and by China Mobile. Both told me at the Huawei BBW in Tokyo the results were excellent, and they are moving on thousands. There are many refinements to come. 3X-10X
Half a dozen other technologies are starting to have an impact. For example, Telus CTO Ibrahim Gedeon thinks SON & HetNets make his small cell deployments much more efficient. He tells me S&H may make more difference in the next few years than the more publicized advances. There are plenty of claims of the size of coming improvements but no solid data from the field. The effect will not be small.
Most of these tools work together, so multiply the impact of each to get a total. Several are deploying in volume in 2017, but upgrading large telco networks takes time. We don't yet know how to get maximum results or integrate the different pieces. The results are already dramatic. Verizon's David Small estimates his costs are going down 40% per year. I have data from Telefonica that implies a 60% improvement last year.
Most believe it will be four to seven years before we see a large effect in mobile from 5G millimeter wave (mmWave) high frequency. Leading researcher Ted Rappaport expects results sooner. The 20 gigabits at the Huawei demo is exciting, but almost everyone thinks large deployments on mobile will wait until next decade. Giant China Mobile is among the most advanced but only expects 1% of its network upgraded in 2020. Verizon plans to test mmWave fixed to several hundred homes in 2017. mmWave to phones is two to four years away. The advanced technologies I'm discussing here have a bill of materials cost similar to what costs were in 2011. The vendors out with the newest technology hope to get a premium. It may take 2-4 years for the right price to emerge.