Artificial intelligence is expanding into all areas of our lives. At the moment Apple has not shown any platform or tool that resembles those available on the market. However, we know that in Cupertino they work daily on technology related to AI: a search engine, an agent similar to ChatGPT, improvements to Siri and implementation of these tools in their applications. The latest report published by analyst Ming Chi-Kuo assures that Apple will invest more than $4 billion throughout 2024 to maintain the servers necessary to launch an entire complex artificial intelligence system.

Apple will spend more than $4 billion on servers for its artificial intelligence

A few days ago we talked to you about the unknowns around artificial intelligence and Apple’s possible next moves based on Mark Gurman’s predictions. The analyst assured that iOS 18 will be the beginning of the integration of an AI system in Apple applications such as Pages or Numbers, in addition to improvements in Siri, integration of a generative AI system and much more. In addition, among these new features would also be the integration of this technology in the development software through Xcode.

Siri

Related article:

AI will arrive powering iOS 18, Siri and much more

Ming Chi-Kuo, a well-known analyst in the world of Apple, follows in Gurman’s footsteps and launches to ensure that they foresee investments of at least $620 million in 2023 and more than $4.7 billion in 2024 for servers intended for this artificial intelligence technology. A million-dollar investment that would be accompanied by a large-scale deployment of technology.

iPhone 15 Pro Max

With this money we would be around the acquisition of between 2000 and 3000 servers in 2023 and up to 20,000 new servers in 2024. Kuo points out that Cupertino would be purchasing systems equipped with Nvidia’s HGX H100 GPU from the Hopper generation (about $250,000 per server). These servers will allow Apple train generative AI.

Finally, the analyst presents the idea that they could be thinking about designing their own chips for the servers, which would save a few million dollars. However, there are no signs that this is happening and perhaps the rush to launch the technology as soon as possible will leave the design of chips for its own servers for later.