At the conclusion of the C-Band auction, I wrote that deployments of millimeter wave technology would largely stop. Even as Verizon reaffirmed its commitment to deploying mmWave, I believe that these plans would be downscaled in due time. This is because the cost for mmWave would strongly suggest this direction. Over the past week, many had questions on the operating costs of mmWave technology. Here, I will address 5G power consumption and its cost. While power is not the main operating cost, people are paying more attention to it. So, to be clear, I am not rationalizing a conclusion based on the cost of power alone. But it serves to illustrate a point. [See here for our analysis on how mmWave compares with FTTH on cost and performance.]
5G Power Consumption Comparison
The power consumption of C-band sites is higher than that of mmWave. C-band radios implement massive MIMO technology with 32 or 64 transmit and receive elements. Many service providers will opt to deploy the 64Tx variant in the dense urban areas which provides more capacity but at higher power draw. For a 100 MHz carrier, the power draw could be between 450 – 550 W for a single sector.
For comparison, a mmWave unit typically operates over a bandwidth up to 800 MHz and typically implement a 2Tx2Rx or 4Tx4rx MIMO antenna. The power draw ranges between 320 – 450 W.
| [per gNB] | C-Band | mmWave | Unit |
|---|---|---|---|
| Carrier BW | 100 (1 CC) | 800 (8 CC) | MHz |
| Antenna mode | 64T64R | 2T2R/4T4R | |
| Tx power | 160 - 200 | 0.5 - 1.5 | W |
| Antenna Gain | 24 - 26 | 23 - 29 | dBi |
| EiRP | 78 - 81 | 55 - 60 | dm |
| Power consumption | 450 - 550 | 320 - 450 | W |
Cost of Power
The annual cost of power for a 3-sectored cell site comes to about $1,300 for C-band and $1,050 for mmWave. This is 25% higher for the C-band site which offers lower capacity on account of smaller bandwidth.
However, consider that for every C-band site, we need 2-4x the number of mmWave sites. On average, the ratio I use is 1:2.4 which is what a Bell Labs/Nokia analysis showed. This leads to mmWave costing around $2,500 in power expense for the same coverage footprint as C-band, or about ~2x the cost.
Other Costs and Benefits
Operating costs related to site lease and antenna mounts, support and maintenance, and software upgrades over time account much more than the cost of power.
The ratio of 2-4x will still be in play: while the C-band sites fit within the existing cellular grid, at least in the urban core. Millimeter wave will require densification, unless one plans a Swiss cheese deployment model.
On the other hand, mmWave does provide an advantage in throughput – somewhere between 2.5 – 3x for an 800 MHz mmWave solution over a 100 MHz C-band carrier. However, mmWave has practically no indoor penetration and will be limited to outdoor, on-street use.
The Fallacy of Cost per Bit
If we compare cost per bit, we find that the cost of power is 4x more expensive in C-band over that for mmWave. However, C-band will cost 50% less in absolute terms. This is why cost-per-bit comparisons have to be considered in the right context.
Industry likes to use cost per bit as a benchmark. But this technique has flaws since it hides the absolute cost that drives the business case and determines profitability.
In fact, I believe some in the industry are using this and similar metrics to make misleading claims. For example, I came across a few publications that are worded to sound as if 5G will reduce total energy consumption of wireless networks. This is simply not true as 5G will add a considerable energy drain. This is one of the topics that no operator likes to discuss! For more on power management in networks see my article on the application of automation in networks.
A Concluding Thought
There are some instances where mmWave will make sense. Perhaps in a cut-throat competitive environment demonstrating a record peak throughput makes for good marketing. But the price for this marketing approach is high. Finally, it may makes sense to operate mmWave in venues like stadiums. In the use case, performance and the ability to provide differentiated and unique services will determine whether it’s worth the cost.