There are several other postings from the community about sporadic incorrect power spikes registered on Kasa smart plugs. I’m not sure anybody has gotten to the real root cause, but in most cases the Kasa device was reading the same instantaneous power when the event happened. But you can’t compare Sense historic data to Kasa historic data. Sense history is entirely based on instantaneous samples. History on the Kasa (daily, weekly, monthly) uses a separate, different “call” to the smart plug.
To add to @kevin1’s great comment above, we believe this is an issue with the smart plug. On our end, we didn’t see any usage close to that amount looking at the data from 7/8 on our end. We’ve seen some cases before where a faulty smart plug will lead to crazy high readings like this and have not been able to recreate it on our end.
To add a bit of context. I’ve experienced this issue as well once the KP115s bug out as described in other posts. When working normally I haven’t seen the issue. Additionally, emulated 110s through HA don’t seem to have this issue which makes me believe that the issue resides entirely on the hardware/firmware