Why Netflix still using AWS when Amazon is their biggest competitor?

A lot of Netflix's technical "secret sauce" is their own infrastructure via the Open Connect initiative.

You cannot get a silky smooth experience across that many congested networks with just bare AWS.

Why would Netflix want to take on the enormous risks of self-hosting? They don't have the experience and they don't have the skills. They can hire the appropriate people but it's extremely expensive. AWS provides a lot of very valuable tooling and expertise baked into the service price.

No one will know, and those that do will never reveal it. But I can surmise that it's a mutually net beneficial arrangement in the short-run, with Netflix expecting a better solution to present itself either internally or externally at some point. Otherwise, you're right, Netflix wouldn't use AWS and AWS would not supply to Netflix.

In a way, they're allocating money and resources to Amazon rather than developing internally simply because the long-term ROI doesn't warrant it.

And let's not discount the most obvious reason why it's probably not even a concern for Netflix: Amazon Prime Video sucks. Sure, Blockbuster probably said the same thing about Netflix at one point, but it's hindsight. I think most people, at the time, would have never guessed Netflix would be the one to pivot to online streaming either.

It's a very though provoking question and there's no way to distill it into one answer, so I'm just providing my POV.

You're really looking at this from the wrong direction. Your competitor doesn't need to fail in order for your own business to succeed. It isn't a zero-sum game.

It wouldn't surprise me if Netflix had that meeting at some point. What would it cost them to run their own infrastructure? Would that be less than the cost of staying with AWS? Is Amazon exploiting their position as Netflix's hosting provider to unfairly compete with them? Do they fear that they might in the future?

At the end of the day, the answer was very likely a solid "no". Many industries are full of examples of competitors working together to the benefit of both companies. Just look at the aerospace industry.

I would expect Netflix to have reasonably detailed strategic analysis and projections of all of the available options - moving to Google, moving to Azure, moving to Oracle, moving to multi-cloud, colocation, not to mention things we can't even imagine because ... well, we have no real idea.

At different points any one of these may make strategic sense, but as others have highlighted here such moves would consume resources that could be allocated to other areas.

My take would be that increasingly the Netflix business model is really content and production. Disney is the real competitor here. Technology is necessary, but not sufficient.

Also ... Dropbox is a good example of a product that has lost it's way. It went from a very simple and clear product with strong value to something we all now work around. Its more reliable to email or slack files than deal with whatever confusing thing Dropbox is doing.

Amazon Web Services (AWS) and Amazon Prime, while part of the same company, are vastly different services.

While it's fair to call them competitors, it's not really an apples-to-apples comparison.