Why industrial automation is beginning to rethink engineering compromise.
One of the pleasures of walking a trade show is listening to the language companies choose to describe their innovations.
Some focus on technical specifications. Others talk about applications, customer success stories, or the problems they are trying to solve. After a while, patterns begin to emerge; not because companies are saying the same thing, but because they are often trying to answer the same questions.
At Automate 2026, one particular pattern became increasingly difficult to ignore: almost every conversation appeared to challenge a trade-off the industry had accepted for years.
The conversations themselves varied enormously. Some centred on sensing, others on AI deployment, robotics or industrial computing. Yet despite the different technologies, many seemed to circle back to remarkably similar challenges. How can manufacturers achieve more without making automation harder to deploy, integrate or operate?
That question was rarely asked directly, yet it seemed to underpin many of the discussions taking place across the exhibition floor. The conversation often extended beyond adding new capabilities. Instead, it focused on how those capabilities could be delivered with less complexity, less integration effort and fewer barriers to adoption.
Rethinking compromise
These are the kinds of decisions engineers have always made. Designing an automation system has never been about chasing perfection; it has always been about balancing performance, cost, complexity and practicality to deliver the best overall outcome. Every project requires engineers to weigh capability against cost, speed against precision, flexibility against simplicity and performance against practicality. Improving one aspect of a system often affects another, and those relationships are not signs of poor design but fundamental realities of engineering.
For decades, those trade-offs were accepted as an unavoidable part of industrial automation. Higher resolutions could reduce frame rates. More sophisticated AI often required months of collecting and labelling training data. Greater capability frequently came with longer implementation times, larger budgets or more specialist expertise. Innovation was measured by how well those compromises could be managed.
Manufacturing has rarely demanded perfection. It has demanded confidence. An inspection system doesn’t have to identify every conceivable defect if it can reliably detect the defects that matter most. A robot doesn’t have to solve every manipulation task if it can consistently perform the one required on a production line. The challenge has always been delivering sufficient performance without introducing unnecessary cost, complexity or implementation effort.
Walking around Automate, however, it felt as though more companies were beginning to ask a different question:
“Does this compromise still need to exist? “
Looking back after the show, that question seemed to connect conversations that had initially appeared to have very little in common.
At first, the conversations seemed unrelated. Some focused on making AI easier to deploy. Others explored richer sensing, greater interoperability or reducing the complexity of integrating automation into existing production lines. Yet the more discussions we had, the more they appeared to be answering the same underlying question.
Innovation is increasingly being judged not by what it adds, but by what it removes.
None of this suggests that the laws of physics have somehow changed. Engineering will always involve compromise, and every automation project will continue to balance competing priorities. What appears to be changing is not the existence of trade-offs themselves, but which ones manufacturers are prepared to accept. Increasingly, the expectation is that software, AI, sensing and computing should remove unnecessary friction rather than simply add new capability.
The emphasis is shifting away from simply delivering more capability and towards removing the friction that has traditionally accompanied it. Across the exhibition, conversations repeatedly returned to making automation easier to deploy, simpler to integrate with existing systems and more accessible to manufacturers who may not have specialist AI expertise. Capability still mattered, but increasingly it was being valued alongside usability.
That may sound like a subtle distinction, but it represents a significant change in how innovation is being perceived. For years, success was measured by what a new technology enabled manufacturers to do. Increasingly, it is also being measured by how much time, effort and complexity it removes from deploying those capabilities in the first place.
Seen individually, many of these developments appear incremental. Viewed together, they suggest something more significant. MemryX’s work on efficient AI acceleration illustrates how advanced machine learning is becoming increasingly practical within the constraints of industrial hardware, while SILC Technologies demonstrated how richer sensing can simplify system architecture by consolidating capabilities that once required multiple devices. Different technologies. Different applications. Yet remarkably similar objectives. Rather than pushing technical boundaries for their own sake, both are helping remove the barriers that have traditionally made advanced automation more difficult to deploy.
Engineering will always involve compromise. There will never be a technology capable of delivering every benefit without some form of constraint. Yet Automate suggested that the industry’s ambitions are beginning to change. Increasingly, companies are no longer asking manufacturers to accept complexity as the inevitable price of greater capability. Instead, they are attempting to remove the friction that has traditionally accompanied innovation.
That may ultimately prove to be one of the exhibition’s most significant themes. Not the arrival of a single breakthrough technology, but the growing expectation that progress should demand fewer compromises than it once did.
