This is almost possible with 2.6. I would need to add just a couple of new variables to the conditions to make it doable. Specifically, a variable like "nextIsRound" to determine if the next level is a round or break, and a variable like "elapsedSeconds" so an event can be triggered at X number of seconds after a left starts. Part of the overhaul in 2.6 was to make it easier and more consistent to add new variables, so this should actually be pretty easy.
With those, I would think you could create an event that triggers at the start of a round that plays the "The blinds are now" audio file. Then, create an event for each blind level that plays the blind amounts at, say, 3 seconds into the level (however long the "The blinds are now" audio file takes to play), with conditions like "isRound and (smallblind = 400) and (elapsedSeconds = 3)". Then create another event that triggers, say 10 seconds into the round that plays "The running ante is now". And finally, create events for each ante level, to play at say 13 seconds into a round, with conditions like "isRound and (ante= 100) and (elapsedSeconds = 13)".
Yeah, it's a little hacked together. The alternative is to combine the individual audio files into a single audio file. So, a single audio file might be "The blinds are now 400 and 800. The running ante is now 100". And then it would need only a single event with conditions like "isRound and (smallblind = 400) and (ante = 100)". I would recommend this, as it (a) doesn't need the "elapsedSeconds" variable, and (b) wouldn't have a high probability of sounding as robotic as several audio files played back to back.
The "The blinds will increase in 2 minutes" or "There will be a tournament break, commencing in two minutes" announcements would still need the "nextRoundIs" variable.