Realme 5G down to $145 ** 80M 5G subscribers in Q2, 100M August prove 5G is real.** 5G Worldwide: Saudi first, USA last** 5G Phones $199-260
China is > 70% of 5G as subs pass 80M ** Primary: Cost per bit of wireless is falling at a ferocious rate
5X Advance for Africa with Massive MIMO ** Spectrum Marvell: 5 nm 20-40% better ** $400 TCL REVLL 5G at T-Mobile: Here comes 5G in the USA
$Hundreds of millions in extra chip costs prove Open-RAN not quite ready
What's most important to me to explore as I go to five events, Berlin to Hangzhou?
What technologies to connect the next four million?
Massive MIMO may be a key tool, but of course, anything that brings down the price is crucial. Interesting choices with Wi-Fi.
How many will switch from landlines to wireless as wireless speeds pass 100 megabits?
Capacity is critical, today only 5 gigabytes to 25 gigabytes most places. Free in France, selling 100 gigabytes for $24 points to the future, but that's less than half today's average consumption. Telstra is seeing a modest trend and 5-15% is common. Will that accelerate with better wireless?
When 5G mmWave becomes important, will the technology and high costs of the millions of cells drive competition out of the market?
Some still wonder if one mmWave network is affordable and two will probably be a stretch. The four to seven usually needed for strong competition may be out of reach. Everything might change for the business. Policymakers may have to choose between strong network sharing/wholesale separation or strong regulation.
Will autonomous cars require 1 ms latency or are they fine with 5-10 ms?
It looks to be brutally expensive to deliver 1 ms driving intelligence to the edge. Most carriers, including Verizon and China Mobile, are using money-saving Centralized Ran (C-RAN) systems that control many cells. The router hops mean latency is 5-10 ms, not 1 ms. Virtual and augmented reality and telehealth are fine with ten milliseconds, experts tell me. Many think connected cars will not need 1 ms, especially with DSRC car to car connections. Gerhard Fettweis is an articulate advocate of the 1 ms tactile Internet.
How much capacity do we lose by having three or four networks rather than one efficient one?
One 100 MHz LTE network delivers far more capacity than five 20 MHz nets. Guard bands becoming unnecessary. When the customers of one company demand more, the bandwidth can come from unused spectrum that in four networks would have been locked to another carrier. This is especially important in countries with few landlines, including most of Africa, is they want a video quality Internet.
Will cablecos install remote Phys for the faster full duplex upstream?
Full duplex will allow a gigabit or more of upstream bandwidth by about 2020. They require a new network component, a "remote phy." Will the operators make the investment or will they simply upgrade 3.1?
What fiber speed is necessary?
Most fiber home is GPON, shared 2.4 gigabits down and 1.2 up. 10GPON is now available at a reasonable. Will customers need upgrades? Verizon believes 40 gigabit NGPON will be needed for backhauling 5G small cells but others think much cheaper 10 gigabit fiber is enough.
What proportion of 5G will be millimeter wave and how much Massive MIMO?
Both can deliver speeds well into the gigabits.
Hundreds of smaller questions are open but these are the big trends.