In Samsung's case, the defects could be directly responsible for damages to human lives. That's a far reach from software vulnerabilities. Not to mention that software bugs can be incredibly difficult to identify, and sometimes impossible until new exploitation methods are discovered, often times well after the creation of the software, and sometimes only due to new software and hardware tools that only later become available.

If you ever wanted an example of how to stifle innovation, read the comment I'm replying to.

My admittedly ambiguous threshold is "significant impact to quality of life"

Battery fires can kill so even a handful is significant, but a security vulnerability that impacts thousands of routers has lower but wider impact. Some companies will be targeted for DDoS or using the routers to probe and infect the company infrastructure... some consumers will end up paying ransomware, or having their finances hacked, or personal info leakes, or bandwidth siphoned.