I noticed a significant difference in the power usage for a Sonoff S26 R2 with ESPHome and Tasmota:
As you see in the graph, the difference is about 50% in favour of Tasmota.
Are there settings (or other things) for ESPHome to limit the power usage without compromising functionality? I guess sleep or deep sleep are out of the question.
0.4 w is not a big deal perhaps.
But are they reporting the usage at the same rate? I guess that is one thing that could be different if one of them reports the usage more often than the other.
Just to be sure we are comparing the same thing.
One thing to consider here is that the graph of ESP-Home seems a lot more stable with no dips to 0 w as Tasmota seems to have.
See if you can “zoom” in on the data and see if they seem to report equally often using the history tab and a very short time interval
EDIT:
Perhaps this is an accuracy thing?
Connect a coffee brewer, microwave oven or something that actually consumes something and do the test with that connected (and turned on). Or perhaps a LED bulb if that completely removes the accuracy on the other end.
Perhaps the readings is so small that it can’t be sure about the readings.
The ESPHome power usage seems to be fairly stable at 0.9W, as Tasmota fluctuates more around 0.4W (also on the long term for both), both with the relay switched off. A Fibaro wall plug was used for this measurement, and it does not send a new report when the power usage stays the same. I compared the measurements of this Fibaro wall plug with a non networked power meter, and they were the same within 0.1 W.
I agree that for one device it’s not a big deal, but when when you use more an more devices with rising electricity prices, it becomes more interesting to see what is possible. Especially when a simple config change can accomplish this, and because they use electricity 24/7.
Ok. I thought the readings was from ESP-Home / Tasmota on the device.
In that case it’s a fair comparison.
I have always thought ESP-Home was lighter since it only installed what is needed where as Tasamota installs everything.
If I’m not mistaken tasmota uses some power saving features for wifi by default. Esphome on the other hand has wifi power saving functions for the esp82xx disabled by default for better stability/range.
You could redo your tests with light and high power saving enabled and present us your findings 
Yeah that is definitely it. I changed one of my ESPs to power_save_mode: none while debugging a connection issue and this raised the power consumption 0.5W.
Thanks for the tip @orange-assistant !! My findings:
light setting is exactly like the Tasmota graph!high setting does not seem to add anything at first glance, so I will stay with the light setting.I happen to have 2x Sonoff S26 that are both programmed to do everthing exactly the same at the same time. (I.e. switch on/off lights in the livingroom each on one side of the room)
I have notice that the tasmota device is much more stable if it comes down to WIFI connectivity.
For the tasmota I can hardly tell when connection has droppen, but the ESPhome is always booting in ON status (which is for my wife a reason to unplug the lamp, so low WAF ;).
FYI, I swapped brothers devices and problems remains with ESPhome.
What could be the cause?