Two worlds of connection
All these chips must follow shared standards, such as Wi-Fi, Bluetooth and Ethernet, so that devices from different makers can work together.
Why AI made these chips more important. A large AI model is trained on thousands of processors at once. They must pass data to each other all the time. If the links are slow, the processors sit idle, waiting. So in an AI data centre, the chips that move data matter as much as the processors themselves.
Custom chips. The largest cloud companies now design their own AI chips, built for their own work. The industry calls a chip made for one customer a custom chip. Connectivity chip makers design and deliver these chips for them, because they already have the skills to link processors together.
Who buys. The buyers are phone makers, device makers, makers of network equipment, and the large cloud companies.
Every company named here has posted software engineering jobs in India. Famous companies that don't actively hire software engineers in India are left out.
Keep reading, free
One more section is on this page: Chips for data centres and for wireless devices. Sign in to read it here, in full.
Continue with Google- Chips for data centres and for wireless devices
Data-centre networks and custom chips
Wireless chips for phones and devices
Two more companies work in this area:
Inside an AI data centre, the copper links between processors are reaching their limit. So the connectivity chip makers are moving to optics: light, carried on fibre, right to the edge of the chip.
Who these companies hire, and for what, is on What semiconductor and chip design companies hire for.