I personally would exclude OpenAI from my list of likely to endure. OpenAI partners are largely switching to Anthropic, their models and tools are generally less well regarded. The only time I had someone advocate for them it was due to some scenario where it was much cheaper.
Problem is that based on what I've read. Altman has way overextended his company. The financial obligations outpace their revenue and likely revenue way too much.
Anthropic has some risk since Microsoft, Google, and Apple insourcing would be a big problem for them.
Well yes, but for example, when someone activated voice interaction with their phone, Google or Apple decide what AI solution is used, and Google already decided to roll their own. When most companies go into AI, just expanding their existing relationship with Microsoft is what their business leaders like. Generally the established companies have preserved the customer engagement even when they resell newcomers.
An example of what you ponder would be like Sears, kmart, Amazon. The thing was that Amazon was greatly valued going direct. People loved the prices and shipping speed. Imagine if instead Amazon decided to partner with Sears and K mart and those companies largely handled the customer engagement with Amazon fulfillment on the backend. Amazon would have been much more exposed to those companies bringing it in house. This latter situation more closely resembles OpenAI and Anthropic right now, at least for the big businesses likely to pay enough after the price hikes.