Fridge hack costing millions and sparking fires stemmed from poor security

We have reached the point in the hardware cycle where the jokes about smart toasters and internet-connected fridges aren't funny anymore. They are expensive, and in some cases, they are literally on fire. A recent security failure involving cooling systems across 700 supermarkets has turned into a nightmare of property damage and financial loss, all because we haven't learned the most basic rule of IoT: if it doesn't need to be on the public internet, don't put it there.

The Anatomy of a Cold-Chain Collapse

The incident wasn't a sophisticated zero-day attack or a state-sponsored mission. It was a failure of security hygiene. Attackers found their way into the industrial cooling units of hundreds of grocery stores by targeting outdated hardware that was left exposed to the open web. Once inside, they didn't just steal data; they manipulated the physical environment. By overriding temperature controls, they caused millions of dollars in food spoilage. Worse, in some locations, the tampering led to hardware malfunctions that sparked actual fires.

For a founder building in the hardware or AI space, this is a sobering reminder that our code has physical consequences. When you're managing 700 locations, you aren't just a software company; you are the steward of a massive piece of critical infrastructure. When that infrastructure fails, you don't just lose users. You lose warehouses.

The Myth of the Managed Perimeter

The problem usually starts with convenience. A technician wants to check a fridge's performance from their house, so they open a port. Or a developer decides that pushing updates via a public IP is easier than setting up a proper VPN or a localized gateway. This is technical debt that carries a high interest rate, paid in liability.

In this specific hack, the devices were running legacy software that had long since passed its expiration date for security patches. This is a classic builder trap. We focus so much on the shipping of version 1.0 that we forget about the maintenance of version 0.8. If your business model involves selling hardware that stays in the field for ten years, but your security roadmap only covers two, you are effectively building a time bomb.

Why Builders Keep Getting This Wrong

Most IoT startups are incentivized to move fast. Investors want to see connectivity, data collection, and 'AI-ready' features. Security is often seen as a friction point that slows down the user experience. But as we see here, the friction of a secure login or a closed network is nothing compared to the friction of a multimillion-dollar lawsuit and a fleet of melted refrigerators.

• Over-connectivity: Does the fridge actually need to talk to the cloud, or just a local controller?
• Legacy bloat: Old code in 'set and forget' devices is the primary entry point for modern botnets.
• The UI Trap: We spend too much time on the dashboard and not enough on the underlying authentication protocols.

We need to move toward a 'Security by Default' mindset. If a device is shipped, it should be locked down. The user should have to intentionally open it up, rather than the builder leaving it open and hoping no one notices. In the case of these 700 supermarkets, the doors were essentially left unlocked in a bad neighborhood.

The Real Cost of Physical Disruption

When we talk about 'disruption' in tech, we usually mean it in the Silicon Valley sense—changing an industry. In this case, the disruption was literal. Supply chains were broken. Insurance premiums will likely skyrocket for any business using similar smart-cooling technology. This creates a chilling effect on the very innovation we are trying to push.

If you are a founder building in this space, you have to realize that your 'smart' feature is a liability until proven otherwise. If you can't guarantee that a remote actor can't turn your device into a fire hazard, you shouldn't be shipping it. The 'move fast and break things' mantra is fine for a social media app, but it is dangerous for industrial hardware.

A Path Forward for Hardware Founders

The solution isn't to stop building smart devices. It's to build them with a healthy dose of skepticism. We need to assume that every device will eventually be targeted. That means implementing hardware-level root of trust, ensuring that firmware updates are signed and encrypted, and most importantly, educating the end-users about the risks of public exposure.

The biggest threat to the IoT industry isn't a lack of features; it's a lack of trust. Every time a fridge sparks a fire, the entire category takes a step backward.

We also need to rethink the lifecycle of these products. If a piece of hardware can no longer be patched, it needs to be sunsetted or disconnected from the network entirely. Leaving 10-year-old unpatchable tech on a public-facing IP is gross negligence in 2026.

The Takeaway

Security isn't a feature you add later; it's the foundation of the product itself. If your hardware interacts with the physical world, your threat model includes fire, flood, and famine. Stop exposing industrial controls to the open internet. If it’s connected, it’s vulnerable. If it’s legacy, it’s a target. Build like your customers' lives and livelihoods depend on it, because as this supermarket hack proves, they actually do.

Read the original at Ars Technica →