
Morgan Stanley reckons Nvidia’s GB200 racks mint money
Rivals left bleeding red ink in AI factory economics
Number crunchers at Morgan Stanley have been adding up some numbers and divided by their shoe size and reached the conclusion that Nvidia’s monster GB200 NVL72 racks are the only kit worth buying if you want to turn an AI factory into a cash cow.

Cambricon cashes in on anti-Nvidia orders
Beijing's AI chip pipsqueak rakes in the yuan
Chinese chip upstart Cambricon is laughing all the way to the bank after netting a record profit on the back of Beijing's vendetta against Nvidia.

Nvidia pushes into self-driving cars as robot future looms
Chip giant sees its robot dreams on four wheels first
Nvidia has set its sights on self-driving cars as the next major cash cow for its physical AI push, hoping to ride a wave of autonomous vehicles before humanoid robots make it out of its smoke-filled labs.

Infineon links up with Nvidia for physical AI robots
Chipmaker arms Jetson Thor with smarter motors
Infineon has teamed up with Nvidia to push physical AI deeper into humanoid robotics, stuffing smarter motor control into the chip designer's Jetson Thor developer kits.

Nvidia shows off Blackwell GB200 and GB300 racks
Chipmaker talks modular MGX architecture and rack guts at Hot Chips 2025
Nvidia used Hot Chips 2025 to show how its Blackwell GB200 and GB300 systems are bolted together, highlighting its modular MGX platform and sprinkling in a bit of open compute virtue-signalling along the way.

US giants queue up for TSMC's Arizona silicon
TSMC's stateside fabs are in hot demand
The Fruity Cargo Cult Apple, Supermicro, Nvidia, and Broadcom are all shoving each other to the front of the queue for production capacity at TSMC’s Arizona factory.

Nvidia pulls plug on China-only H20 GPU
Nvidia backs away as Chinese market turns sour
Nvidia appears to be winding down its H20 GPU line for China, as the economic and political winds continue to blow straight in its face.

Nvidia tapes out Rubin chips
Huang reckons Rubin will reshape compute from the ground up
Nvidia supreme dalek Jensen Huang has confirmed that his outfit is already elbows-deep in its next big AI architecture, Rubin, which is apparently going to be a full-blown “revolution” for the compute world.

AMD leaks FSR 4 source code
SDK update locks features behind DLLS
AMD had a bit of a senior moment this week when it accidentally dumped the FSR 4 source code on GitHub before yanking it offline when it realised they’d left the safe open.

ARM poaches Amazon chip guru for AI push
Ramps up in-house CPU plans
Chip designer ARM has poached one of Amazon’s top silicon brains, Rami Sinno [pictured], to spearhead its in-house AI chip efforts, in a move that looks suspiciously like a direct shot at Chipzilla, AMD, and Nvidia's datacentre turf.