The Internet of Things (IoT) is a long way from becoming a mature technology. From wearable devices to industrial sensors and consumer conveniences, IoT vendors and users are still trying to figure out what the technology does best as it grows into a $9 trillion market by 2020 (according to some estimates).\nAnd yet, IoT is somehow already faced with a huge and growing problem of obsolescence. The problem, ironically, lies in the \u201cthings\u201d themselves.\nApple Watch: A premature antique\nDon\u2019t believe me? Consider the solid gold Apple Watch Edition, launched in 2015 and sold for $10,000 to as much as $17,000 a pop. A traditional watch at that price point would be expected to last decades, perhaps even generations as it turns into a family heirloom. But with the announcement of Apple Watch OS 5 at the company\u2019s World Wide Developers Conference this week, the original version of these fancy timepieces can no longer keep up. They simply won\u2019t run the latest version of the operating system due out this fall, and they won\u2019t have the features of brand-new Apple Watches that cost a tiny fraction of that amount.\n\nThis is not some surprising new development. Among other shocked observers at the time, even I asked, "Why spend $10,000 on an Apple Watch that will be obsolete in 2 years?\u201d Well, perhaps my post was a bit aggressive \u2026 the Edition didn\u2019t turn into a \u201cpremature antique\u201d for three years, not two!\nSome things are hard to replace\nIf you think about it, this is a relatively simple-to-solve instance of IoT obsolescence. The more common, and much harder to deal with problem, is IoT devices located in places that can\u2019t easily be reached. Last fall, for example, I wrote about the (voluntary) recall of almost half a million St. Jude Medical IoT devices due to a risk of hacking.\nNo big deal, right? Devices get recalled for security fixes all the time. Unfortunately, in this case, the devices involved were pacemakers installed not in some easy-access equipment rack, but in patients\u2019 chests. Swapping them out would be a very big deal (fortunately, as of publication of the post, none of the pacemakers had actually been compromised).\nThe issue goes much further than watches and pacemakers. Smart cars have IoT systems that will become obsolete long before the vehicles in which they\u2019re installed reach the end of their useful lives (sort of like aging 8-track players still riding around in the dashboards of cars from the 1980s).\nAnd it gets worse. For example, many industrial sensors essential to delivering the benefits of IoT are located in hard-to-reach spots where replacement or upgrades would be difficult, expensive, or hazardous. Sure, remote software upgrades can help alleviate the problem in many cases, but sometimes you need to fix or replace the thing itself.\nA modular solution\nThere is a solution for some of these issues. Back in 2015, I suggested that Apple a offer \u201cmodular, replaceable guts\u201d for its Apple Watch Edition. Allowing users to replace the functional parts of the device, which everyone knew would need to be updated periodically, would have allowed them to keep using the super-expensive gold case, engineering an end run around the inevitable obsolescence of this very expensive internet \u201cthing.\u201d\nOf course, that would likely have made the Watch even bulkier and might have affected performance in a variety of ways. On the other hand, they might not be turning obsolete in a few months. That approach won\u2019t solve all of IoT\u2019s obsolescence issues, but it\u2019s a start. More to the point, perhaps, IoT vendors and buyers need to consider the useful life of the devices, as well as what will happen when (not if) they break or can no longer keep up with technology advancements.\nSure, that might add some cost and clunkiness. But it has to be done. IoT is too new and too promising to become prematurely obsolete.