ARM
ORCL
The great chip whack-a-mole
Arm CEO Rene Haas basically looked at Washington’s China chip restrictions and said: good luck, have fun. His argument is simple — CPUs are so widely used across so many products that trying to block “AI CPUs” without accidentally blocking a ton of normal computing would be a nightmare.
And unlike Nvidia’s GPUs, which are easier to define with specific performance thresholds and memory-bandwidth rules, CPUs live in a much greyer zone. Translation: if regulators want to draw a neat line here, the line keeps smudging itself like a marker in the rain.
Why investors should care
This matters because Arm’s whole AI story depends on CPUs becoming a bigger piece of the data-center puzzle. The company is pitching its AGI CPU as a next-gen AI building block, and Haas said at Computex that Oracle and ByteDance joined as new customers.
That’s the kind of headline Wall Street likes. More customers means more momentum, and more momentum means the market keeps squinting at Arm like, “Okay, maybe this isn’t just a chip-royalty company with a fancy press tour.”
The catch: geopolitics is still lurking
The U.S. has already tightened the screws on chip shipments to China, including closing loopholes that had let some AI chips slip through overseas channels. So even if CPUs are harder to control than GPUs, this is still a business where policy can change faster than your phone battery percentage when you actually need maps.
- Arm gets a nice validation boost from new customer wins
- Nvidia remains the comparison point for AI-chip export restrictions
- China exposure still means policy risk can show up uninvited
Big picture: Arm’s AI pitch is getting louder, but the export-control fog is getting thicker too. That’s not exactly boring — and for investors, boring is never the vibe here.