Why Do Hyperscalers Design Their Own CPUs?
By Sally Ward-Foxton, EETimes | April 10, 2025
SAN JOSE, Calif. – U.S. hyperscalers like AWS, Google and Microsoft have designed their own CPUs for the data center in recent years, investing hundreds of millions of dollars into their own chip designs. All these CPUs are based on Arm. But with other options on the market, what motivates these companies to spend the time and money on building their own silicon?
Arm’s infrastructure chief, Mohamed Awad, told EE Times that while cutting out the middle man and dealing directly with foundries does mean hyperscalers can cut out chipmakers’ margins, this is not the whole picture.
To read the full article, click here
Related Semiconductor IP
- LDO for CPU Cores on TSMC CLN3P-CLN3X
- LDO for CPU Cores on TSMC CLN3E
- LDO for CPU Cores on TSMC CLN2P
- I3C Controller IP – I3C / I2C Slave, Configure User Registers, no CPU Host Required
- I3C Controller IP – I3C / I2C Slave, SCL Clock only, Configure User Registers, no CPU Host Required
Related News
- Why Do Big Companies Make Their Own Sensors?
- SigmaSense Teams Up with Dolphin Design to Deliver Power Efficiency in their Advanced SDC300 Touch Controller
- Dolphin Design unveils the final piece of their platform offer with BAT, an audio solution for high quality AIoT applications
- Nextchip licenses aiMotive's aiWare4 for their Apache6 automotive domain processor