@opelsinei I'll add that @BlakeRichardson is correct about HLS being selected and optimized for Meraki MV. HLS delay is normal and results in about a 4 to 5 second of delay because HLS turns the live video stream into a set of 2-second video segment files and does sequential HTTP-based file downloads, and needs the first 2 chunks (about 4 seconds) to begins the streaming process. Meraki has optimized HLS for MV quite a bit, I believe it was originally well over 15 to 20 seconds before being optimized for MV.
The trade-off is that Meraki MV is extremely intuitive and simple to use, natively traverses any firewalls or proxies that allow HTTP in the first place, and requires no special software or browsers or specific Java versions or plugins or ActiveX controls or any of that. It's pretty much any browser on any device on any operating system and you're working with the full power of MV.
Certain quality settings on the "Quality & Retention" page can help slightly, like switching from Enhanced to Standard can shave off a second or two. This can slightly change the way the video is encoded as the stream gets split into files for the HLS stream. This setting can be camera by camera as well, so if the camera is zoomed in for identity of someone being buzzed in through a doorway, the quality may not matter much but the extra 1-2 seconds gives a better experience for the lobby admin buzzing them in.
Unfortunately there will always be some HLS delay (remember the tradeoff: powerful, simple & intuitive with no special software or plugins) and it's already been optimized about as much as it can be. To make sure you have all the up-to-date performance improvements, go to the latest firmware, currently 4.4.
Generally speaking the 4-5 second delay does not cause operational impact. If it were to be optimized even further, it would introduce a lot more overhead, such as more numerous smaller video segments and therefore a higher percentage of protocol overhead, which could also introduce more of a chance for choppy video playback.
As others mentioned, you can also leverage RTSP for the actual real-time video stream, and this can also be configured on a camera by camera basis as needed.