The smart manufacturing world’s big event of the year, Hannover Messe, saw 130,000 visitors descend on the German city of Hannover last week. Numbers were reported as very similar to last year and still much lower than before the pandemic: Might this be the new normal for trade shows, as casual freebie-trawlers stay at home while more serious commercial prospects still value the opportunity to supplement their online research and webinar-watching with some intense face-to-face conversation?

My immediate impressions were:

  • China remains important to this European event. Chinese exhibitors were everywhere, running the gamut from tiny booths offering cheap industrial components to larger exhibition spaces showing off world-class industrial automation. Chinese visitors began to return in force in 2023, and this year, China rose from the number three source of visitors to number two, behind Germany. As George Lawrie and I discussed in our recent report on Goldilocks manufacturing, globalisation isn’t dead, but it’s definitely shifting.
  • Everyone had an AI story, even if few made much sense. Last year, I commented that ChatGPT was “pretty pervasive [but] more likely to be a throwaway comment or booth eye candy than a carefully considered component of an exhibitor’s broader proposition.” AI replaced ChatGPT as the preferred term for many this year, but most of their stories hadn’t improved. Although the banners said AI, they really still meant generative AI (like ChatGPT), ignoring a large body of good work on AI and machine learning in the manufacturing sector that stretches back a decade or more. As Martha Bennett and I found while researching our recent report on genAI in manufacturing, this industry is still searching for good generative AI use cases for which customers will be prepared to pay.
  • The industrial metaverse: Vendors go big or go home. The metaverse felt as big as ChatGPT at last year’s show, but Forrester predicted that 75% would stop talking about it. There’s nothing wrong with the underlying technologies, which we’re bullish about, but the overall proposition just isn’t particularly coherent. One or two big exhibitors doubled down on their industrial metaverse story, but most appear to agree with us: The metaverse was (thankfully) hard to find at most booths this year.
  • If there was a big new thing, it was well hidden. Last year’s focus on ChatGPT morphed a bit to become this year’s AI. 2022’s sustainability theme was still (rightly) bubbling under everything. But there didn’t seem to be a buzz around the show for something new. With constrained budgets, labour shortages, and a host of other practical challenges, maybe it’s good that the sector appears more interested in pragmatically consolidating existing advances than enthusiastically chasing after some new shiny thing. We should enjoy this: It probably won’t last.

AI: It’s Not All Bad

Behind the thick layer of AI froth, there were some interesting nuggets. AWS made the most of its reasonably open approach to generative AI’s large language models with a useful side-by-side comparison of responses from different models to the same prompt. To really help developers select the best model for their specific use case, I think it would be helpful to pull public pricing information from elsewhere on the AWS site and do the maths for the user: If this response cost €.001 to generate, is the one next to it worth €0.01? Sometimes, the answer will be yes. Over at the Microsoft booth, a collaboration between the cloud provider, Siemens, and Harting did a nice job of taking a natural language description of a machine part, checking the Harting catalogue for a likely match, and — if one wasn’t found — turning it into a (Siemens-powered) CAD model of a new part that Harting could then make and sell.

Perhaps the biggest short-term benefit from all the AI hoopla will be unlocking budget to clean up data. Every CEO seems to want a genAI thingummy to show. A few are even pretty good. But all of them need to be fed a rich diet of clean, semantically mapped, authoritative data. How many tech leaders, I wonder, see their boss’s latest enthusiasm as a way to secure budget for the dull-but-important data cleansing exercise they’ve been unsuccessfully pitching for years?

We Still Can’t Do This Alone

I commented on the importance of partnerships last year, and it’s still true in 2024. Big tech vendors, big consultancies, big systems integrators, and big industrial OEMs were all falling over one another to celebrate partners and partnerships. Companies such as GE (now Vernova) that might once have occupied a large booth of their own were content with a spot inside AWS’s space. AWS, Microsoft, SAP, and others had huge stands, almost entirely given over to the work of their partners.

No one firm has all the answers for hardware, software, domain knowledge, and deployment into customer workflows, forcing them to give up a little control in order to partner effectively to deliver the customer’s outcomes. Open standards play a part (Cybus, Litmus, and others were among those reporting some interest in Unified Namespace), but so does the industry’s growing enthusiasm for data fabrics that loosen the tight connection between a single application and its data: AVEVA Connect, Hexagon’s Nexus, Autodesk Forge, and others were all on show and had made clear progress since last year, although there’s still thinking to do about how these federated pools of data might effectively interoperate with one another. The cloud providers that sit underneath these solutions also had news about the growing sophistication of their enabling technologies.