$5M of cameras in the garbage
Planned obsolescence is tolerable on a phone. It is absurd on a security system. Why the intelligence layer, not the hardware, is where the money should go.
"Planned obsolescence is the best part of owning a smartphone!" said no one, ever.
But we accept it anyway. In exchange for a sleek design and a good interface, we agree to replace the hardware every few years. With consumer electronics that feels like a fair trade. So why is it the norm for a $5 million, $7 million, or $10 million security system?
For decades, security, loss prevention, IT and operations departments have bought hardware and software as a closed loop. When the cameras can no longer be updated, they go straight in the trash. Rob Frechette, Spot AI's Chief Partner Evangelist, has spent a career watching it happen, and argues the fix is to invest in the intelligence layer rather than in higher-resolution footage.
Don't rip out millions in hardware to solve a software problem
Many vendors still sell proprietary cameras locked together with their software in a video management system. They promise regular updates, but never mention that the hardware goes obsolete fast. In a few years those cameras struggle with new firmware. If the company wants a modern system, it is rip and replace, and hardware worth upwards of $5 million is suddenly scrap.
The legacy VMS model benefits one party: the vendor. Enterprises lose control over their own camera estate. They are locked into a thousand-plus cameras, updated, or not, on someone else's roadmap, with no way to adapt the system to what their facilities actually need.
If getting a picture of a robber is what matters, then go invest in a premium camera network. But usually, learning about the robbers' behavior is more valuable than high-quality footage.
That is Rob's line, and it is the whole argument in one sentence. The layer on top of the cameras, the agentic AI layer, is what matters. And that layer is camera-agnostic. Whether they cost $200 or $1,500, the cameras you already own are data collectors. Point agents at them and they stop merely capturing footage and start observing, interpreting, and acting. An enterprise can build the agents it needs:
- A loss-prevention agent that alerts staff to suspicious point-of-sale activity
- A security deterrent that reduces loitering, vagrancy and dumpster fires
- A safety agent that scans the warehouse floor for risks
- A line monitor that raises production efficiency
- An operations agent that watches for SOP drift
Vendors want you to think your cameras are the problem. Most of the time, the closed ecosystem they are trapped in is what is holding back outcomes that were previously impossible.
I've walked into situations where enterprises were ready to rip out all their hardware. We pulled in Spot AI's, and those same cameras became AI agents the first day they were installed.
Own your video like any other data
In the next chapter of physical security, enterprises will manage and own their video data alongside all their other business data. You do not have to replace the phone whenever the manufacturer says so. You do not have to replace the cameras either.