I have seen this behavior in Chrome before. If one tab is executing a request to a given URL the other tabs will wait for the first tab to finish and then immediately return.
Here is a proof of concept which I am able to replicate in my Chrome browser consistently.
foo.cfm
<cfoutput>#now()#</cfoutput>
<cfscript>
sleep(5000);
</cfscript>
<cfoutput>#now()#</cfoutput>
Open two tabs in your browser to foo.cfm.
Test1: Do this next sequence as fast you can without errors. Click tab1, select the URL bar, hit the enter key. Click tab2, select the URL bar, hit the enter key.
When doing that you'll notice that the time stamps in the first page are 5 seconds apart, and you'll notice the timestamp on the second page did not even begin until the first had finished.
Test2: Do this next sequence as fast you can without errors. Click tab1, hit ctrl+f5, click tab2, hit ctrl+f5.
Now, notice that tab two will finish nearly right after tab1 and the timestamps indicate that processing took place equal to the duration it took you to switch tabs.
The same behavior occurs if you use f5 as well. So basically it's a micro-optimization of the Chrome browser (and likely others) to not make additional http requests for resources which other tabs are already working on. My supposition, which I have not tested yet, is if the resource were to return proper browser caching HTTP headers, then the second request would use the first requests result.