Events like Automate offer a handy microcosm of the industry as a whole. It’s not a perfect one-to-one, of course, but based on who attends, exhibits, and the topics they’re most keen to discuss, we extrapolate a handy snapshot of where U.S. automation is as we approach 2025’s midway point.
I filed several stories from the show floor, all of which can still be read over on LinkedIn. A week out from the event’s final day, they can help form a picture of how the industry has changed in the past year and where things are currently trending.
The most frequent topics I encountered at Automate 2025 were, broadly speaking:
• Industrial AI
• Humanoids vs. the world
• Tariff, uncertainty, and decentralizing manufacturing
So let’s start with the last one and work our way backward.
Tariffs
The choice of venue invariably colors how we approach an event like this. Manufacturing is, of course, foundational to any conversation about industrial automation. That this year’s show was held in the heart of Detroit — a stone’s throw from the Canadian border — brought these topics into even sharper focus.
That said, if A3 opted to hold the show on the moon this year (we decided against it – not enough foot traffic), the subject would have still been at the forefront in the shadow of today’s geopolitical landscape.
Reshoring and nearshoring have been the subject of vigorous debates for the better part of a century, but Covid-19 laid bare the fact that the systems we rely on every day are far more brittle than anyone cares to admit. At the center of the issue are decades of concentrating so much essential manufacturing into a single geographical region.
If, say, a global pandemic were to take an early and outsized toll on that area (hypothetically speaking, of course), the supply chain could effectively ground to a halt. Last week’s LinkedIn piece evoked images of 40,000 Ford trucks at various mid-manufacturing stages that populated a NASCAR track back in late-2022. There are plenty of visuals to draw upon from the era, but this one really got to the heart of how reliant even U.S.-manufactured products are on Chinese components. In this case, the bottleneck was the silicon that powers every inch of today’s vehicles.
The conversation has resurfaced with newfound vigor over the past five months, as tariffs threaten to upend global trade in profound new ways. “Uncertainty” was the magic word, with the size and nature of those tariffs seeming to shift on a weekly — or even daily — basis. Multiple Automate attendees noted how such uncertainty could hamper the adoption of industrial automation.
Another short-term effect is, of course, rising prices. Perhaps this will drive a wider adoption of the RaaS (robotics as a service) model, but for now, the upfront costs of purchasing hardware outright will hit small and mid-sized manufacturers the hardest. If, however, the tariffs have their intended effect, automation adoption (in particular, those systems being manufactured locally) could accelerate in the long run. The truth is that any bid to meaningfully revamp U.S. manufacturing is going to lean heavily on automation, tariffs or not.
But, again, uncertainty is the only thing we can count on for certain.
Humanoids
Given the money and attention presently being expended, it’s not especially surprising that humanoids were an essential topic of conversation. It is, however, worth noting how that conversation has evolved in the past 12 months. There are still plenty of interested parties on both sides of the debate, but Automate 2025 found many of us further exploring the nuances of the conversation.
After 9 years at TechCrunch, I can confidently tell you that excitement follows money every bit as much as money follows excitement. This is precisely how hype cycles are built. In the case of humanoids, both elements were firmly focused on locomotion. While I didn’t have the opportunity to directly sit in on those early-stage funding pitches, I suspect many VCs took a “legs or bust” approach to investing.
Excitement around the humanoid form factor almost always involves legs. Is it any surprise, then, that all the big-name industrial humanoid systems are currently bipedal? The past year has caused many parties to rethink the wisdom and necessity of legs. That’s not to say that most humanoid proponents don’t believe there is a place for legs (nearly all do), so much as an acknowledgement that overengineering isn’t the answer to every problem.
AMRs (autonomous mobile robots) have exploded in popularity over the past decade-plus. As such, manufacturers now know how to build them efficiently and reliably at scale. Wheeled bases are fast, relatively safe, and far less expensive than legs. So why are we throwing out the robot baby with the bathwater here?
While humanoids were far from a common sight at Automate 2025, those from smaller startups like Kinisi, Red Rabbit, and Roboligence offered a kind of happy medium between humanoids and AMRs by mounting bi-manual dexterous torsos to the tops of wheeled bases.
It’s worth drawing attention to two other key aspects of the 2025 humanoid conversation before moving on. The first is safety. That news about the humanoid ISO standard dropped the Friday before the event, causing many to revisit this essential aspect. The last is, perhaps, something to look for in 2026. This is the move from humanoid generalists to specialists, as many of the current crop of companies narrow their focus to specific components, rather than the full robot.
Industrial AI
Ahead of my Industrial AI panel with Kence Anderson (Composabl), Christi DeCuir (NVIDIA), Dianne Huiwen Eldridge (Amazon Web Services (AWS), and Bernd Montag (Siemens), one panelist asked that we not spend too much time on generative AI. However hyped humanoids presently still pales in comparison to the fever pitch generated by ChatGPT and its ilk.
At this year’s event, the topic trended toward pragmatism. For much of the general public, generative AI still feels like magic. If an LLM can write a novel based on a few prompts, surely it can get a general purpose robot up and running in the workplace in no time, right? It’s up to the industry’s to set more realistic expectations for what embodied AI can — and can’t — do in the short term. Otherwise, the black box you’re selling might as well be full of magic beans.
Much of the conversation also shifted from generative to agentic. That is, after all, central to AI’s increasingly proactive role in the workplace. As ever, there are lingering questions around how if at all, humans will slot into that conversation. The truth is that lights-out manufacturing is certainly a goal for some, but it’s unlikely to happen any time soon for most.
It's largely agreed that, as robots increasingly take over labor roles, humans will need to be around to train, monitor, and maintain these systems. For this reason, conversations around upskilling and preparing the workforce for this ever-shifting paradigm are becoming increasingly important.
« View All Blogs
Privacy Settings