It's not clear when you dig in a bit. For example, CBRE has "How is Tech Talent quality vs. cost measured?". They answer that there isn't a good way so their proxy is that they picked the top 5 CS schools in Canada as measured by the U.S. News & World Report. So whatever cities have one or more of those institutions will get a big bump and #6 counts for nothing.
They have an AI section because AI is trendy and it shows the SF Bay Area has around 3x the AI employment of Toronto. This might be a bit of a proxy for people working in newer areas.
Fortinet is a good example where they maintain their HQ in the US. It's not a top tier company. Vancouver has only in the last few years gotten relatively small development offices for some of the major tech companies. Some of these were to manage visa issues with the USA; Canadians were expected to move to Washington or California.
It's interesting to discuss but quasi-arbitrary composite metrics are often much lower in quality than other data (real data?).
I agree with you, it's not a perfect rating system, but there isn't any one good system because everyone's rating system is different. I find CBRE is at least consistent. There are some gray areas of course.
Also in terms of numbers any given list is going to have inconsistencies. Some companies list a certain amount of employees in a given city, but many of those are physically somewhere else. Some may actually be listed in more than one city.
What I really want to look at is where the numbers are trending. The Canadian cities, particularly Toronto, Vancouver and Calgary are doing very well.