Video hosting sites lose money, so it's not like a contender will ever appear, unless the EU pours a ton of money into it, and not even then (just imagine how many EU funds Dailymotion has probably taken)
>Video hosting sites lose money, so it's not like a contender will ever appear,
Only if you assume YouTube's business model is the only feasible one. You could have a video hosting service that operates like a traditional web-host, letting you pay a standard rate or fee to host and list your content rather than relying on ad revenue and algorithmic curation.
Twitch has an ad driven model where the content creators get paid based on subscription counts rather than data harvesting. (Though I wouldn't be surprised if there was a big data harvesting component as well).
Exactly, video hosting can be a viable business. Full disclosure, I am co-founder of Swarmify where this is our business. We provide paid video hosting at affordable prices very similarly to paid web hosting.
I really like what you have to offer. I'm impressed that you offer unlimited video/bandwidth for a reasonable price.
Just a note though, the video example on the homepage comparing YouTube with SmartVideo (SpaceX launch). I can see compression artifacts on the Smart Video. If I'm honest, that really put me off which obviously isn't your intention.
Fair point on that video. We recently changed some internal settings and it looks like that is a good example where we may need to better tune. I appreciate the feedback as it helps us see things we would have missed
Fair Use is an American concept. Even so, YouTube already routinely mutes videos that contain background music.
With the proposed law, the right holders can sue Twitch for hosting that video, unless Twitch ensured they have permission from said right holders to host that video.
They don't say so but people have calculated estimates and it seems there's hardly any way youtube could be _making_ money. Just google it, there's tons of articles with interesting speculations.
The advertising budgets spent on DFP (now the overwhelming majority of budgets) makes it obvious. Youtube makes money. It prints money by information trading on top of that. Subscriptions on top of that. Google covers costs. Maybe not all that money comes back to Youtube as a department, but published revenue numbers are a game for an organization of that size.
A guaranteed adserv on every CDN delivery (and again over a certain number of minutes) is more than the content and serve cost. CPMs for video are net 3$, at least. I don't know who's doing the math, but a link would be interesting with someone demonstrating how they are upside down.
Youtubes infrastructure costs are massive - well into the $1B+ per year category.
Netflix might serve more gigabytes to users, but netflix's content library is tiny in comparison, so edge caches have a near 100% hit rate keeping infrastructure costs a fraction of youtube's costs.
If YouTube has edge caches, which I would be very surprised if they didn't, I'd think their edge cache hit rate would still be pretty high. Not 100%, obviously, but still over 50%–most people watch the same set of popular videos, with a very long tail of unpopular ones that are occasionally watched.
I wouldn’t be surprised if Netflix’s entire library was under 100 terabytes. You could have 100 percent of the content served from edge caches with that.
I haven't tried to estimate the size, but I think one thing that may enlarge it is all of the extra copies of content for various resolutions, different bit rates to support bandwidth issues, and optimized versions for the different mobile devices & apps. That's a lot of different files for 1 piece of content.
Edit--Found this quote from an article[0] written in 2013, so I'm sure it's much larger now. Also, the Open Connect Appliances[1] have about 240TB capacity each, but they just tail the current popular content.
The master Neflix catalog takes up about 3.14 petabytes of cloud storage space, which is converted and compressed down to about 2.75 petabytes, consisting of 100 different versions suitable for watching on more than 1000 different devices.
Only if you assume YouTube's business model is the only feasible one. You could have a video hosting service that operates like a traditional web-host, letting you pay a standard rate or fee to host and list your content rather than relying on ad revenue and algorithmic curation.
Twitch has an ad driven model where the content creators get paid based on subscription counts rather than data harvesting. (Though I wouldn't be surprised if there was a big data harvesting component as well).