Night session surged, reposting an analysis: Some analysts also asked about Apple's AI investment capex. The CFO replied "In the past, our capital expenditures have been a hybrid model, investing ourselves + third parties, and AI is no exception." I think it can be understood that in addition to their own servers, they will heavily rely on cloud services in the early stages. If they indeed partner with Gemini or OpenAI first, the computing power will naturally come from GCP or Azure cloud support, as this is currently the most time-saving option. Everything is for speed (after all, for Apple, AI is not the goal; ultimately driving device upgrades and firmly holding onto the consumer end is the purpose). This also partly explains the recent capex increases by the three major cloud providers and their repeated emphasis on future inference demand. Just imagine, what level of computing power would be needed to serve the inference services for Apple's 1 billion-user ecosystem? And if the billions of Android users follow suit immediately, how much computing power would that require?$Apple(AAPL.US)

Cloud benefits $NVIDIA(NVDA.US), Apple's performance benefits $Taiwan Semiconductor(TSM.US)?