NVDA
A little delay, not a disaster
Nvidia’s next-gen Vera Rubin rollout may be running a bit behind schedule, thanks to thermal issues and some hiccups around SK Hynix’s HBM4 memory qualification. Translation: the AI supercomputer parade is still marching, just not quite at the speed Nvidia probably hoped for.
The Street shrugged and raised the target
KeyBanc’s John Vinh basically said, “Relax.” The firm lifted its price target on Nvidia to $330 from $310 and reiterated an Overweight rating, arguing the delay shouldn’t do much damage to estimates. Why? Because Nvidia can lean harder on B300 GPU shipments instead of the R200 chips, which helps fill the gap while Rubin gets its act together.
Why investors are still smiling
The bigger story here is demand. KeyBanc pointed to stronger-than-expected appetite and raised its 2027 CoWoS supply forecast by 69% to 1.1 million interposers — a nerdy number, sure, but one that basically says the supply chain is still gearing up for more Nvidia hardware, not less.
- Rubin Ultra is still expected to start contributing in late 2027
- Nvidia says Rubin has already entered full production
- The company continues to frame demand as outrunning even Blackwell
Big picture: a delay in next-gen hardware is annoying, but if the market thinks the train is still barreling forward, the stock can keep riding like it’s in the first row of a Taylor Swift ticket queue.