Actually, not if you consider compute power.
My Ecobee Suite does a TON of calculations and string manipulations for every cycle (poll) of the Ecobee servers. On SmartThings, this usually takes between 2-4 seconds; on Hubitat it is more like 5-10 seconds. And there's some sort of "leak" on Hubitat, after about 24 hours of running, the ES Manager cycles will take 25-50 seconds, while SmartThings doesn't waver.
So, I suspect that the CPU's my stuff runs on in Amazon's cloud are bigger/better/faster. I did some looped math testing, and the SmartThings CPUs are consistently 2-3x faster than the Hubitat's processor (doing the same task).
Plus, network traffic from Ecobee to SmartThings (and Alexa to SmartThings, and...) all run "local" within the Amazon cloud, where the routing latencies have long been optimized by all the traffic. Ecobee-to-SmartThings transfers happen thousands of times per second (for all the users - not just my code), but Ecobee-to-my_little_Hubitat are a tiny fraction, and involves routing through several IP routing nodes to get here.
A few milliseconds here, a few microseconds there, and things can really start to slow down.
It's because of this that I'm moving most of my cloud-based services (Ecobee, Echo Speaks, Chamberlain, etc.) back to SmartThings - better performance, and if the network is down, they aren't going to work anyway