To catch up with artificial intelligence, Apple developed private cloud computing, a secure computing infrastructure for artificial intelligence services. Apple produces its own artificial intelligence servers at a factory in Houston, Texas. Their photos were posted by social network user X under a pseudonym Xu Qingbo.
Image source: x.com/hsuchingpo
Four images have been released, one of which shows a 2U server designed for rack mounting. It has a modular design: the server is made up of four blocks, each block contains eight boards – each board contains a processor. The plate is located on the outside of the block and the radiator is mounted on the inside. So each server contains four processors, and these are Apple M5 chips. One of the images shows a die with the Apple logo, and it can be assumed that we are talking about the older and most productive version of the M5 Ultra processor.
The main control panel runs along the entire block, with smaller blocks connected to the main control panel. Connecting cables run along the central section of the server – it’s unclear whether they are used for power, data transfer or both. It’s worth noting that software management of server tasks is not performed by the server itself, but by an external Mac Studio computer – and it can be located in a different location.
The design is designed to ensure the most efficient airflow through the unit from front to back. A radiator is located in the middle of each block, and its fins allow air to circulate freely. The sides and bottom of the block are covered with transparent covers that guide air through the channels. One of the images shows a side view of the AI server without the cover.
If you find an error, select it with your mouse and press CTRL+ENTER.
