Why is it not more common to implement anti-cheat on the server instead of the client? Is that not more secure? Couldn’t the server just check what vision a player should have and not provide any other information to prevent wallhacks or maphacks? Or check how fast it is possible to move to prevent speedhacks? Aimbot is a bit harder to detect I guess but what about the other ones?
Not really that expensive. With server rendering the GPUs can be constantly utilized for different clients or different activities. You only pay per minute of usage. GeForce now is I think $5 a month, which it’s not much. If you pay AWS for their ephemeral tear, 5 to 10 cents per hour depending on the GPU you want.
When they’re not being used for gaming, those GPUs could be used for AI training, model generation, a whole host of things. A GPU in your house probably isn’t used 24/7, so that hardware is not being efficiently a moretized over a large scale of activity
Adding at least $5 a month to a game subscription is expensive considering a lot of them are 10-15$/mo. Note that this isn’t just GPU expensive. You also need to calculate game logic on the server too which can also drive up CPU and ram expenses. This is not to mention that a lot of popular online games are freemium and will stand to lose a lot of players if they start charging a subscription.
Another way to word it is that the freemium model is incompatible with server side rendering.