The Connection Point · Vol. 1 · Issue 15 · June 22, 2026
I'm writing this from Chicago. Automate 2026 opens today at McCormick Place, and if you're walking the floor this week, come find us at Booth #21030. We've got some new things to show you, and I'd rather talk about them in person than over email.
Last week was quieter. I was home, which gave me time to finally sit down with a new 3D printer I'd been putting off setting up. Small thing, but it stuck with me. A few years ago, a desktop printer like this meant calibration headaches and a fair bit of swearing at slicer software. This one was producing usable parts within an hour of opening the box. The whole category quietly matured while most of us weren't paying attention.
That same pattern shows up everywhere in this week's stories: technology that used to require a specialist, now working well enough that the barrier to entry has quietly dropped. More on that below.
GM's new cobots in Detroit are drawing UAW anger, and the timing is the real story.
General Motors has installed roughly 50 collaborative robots at its Factory Zero EV plant in Detroit, working alongside humans to attach body panels on the line. The catch is that this happened a few months after GM laid off more than 1,000 workers at the same facility, and UAW Local 22 has filed grievances over both the job impact and the safety of robots working that close to people.
What strikes me here isn't the robots. It's the timing. Installing fifty cobots a few months after laying off a thousand people is going to read as one thing to the workforce no matter how the press release is worded. GM says it's about safety and ergonomics, and that may genuinely be part of it, but optics matter as much as engineering on a factory floor. Any OEM rolling out automation right now would do well to bring the workforce into that conversation early, not after the robots show up. Read more here.
Coca-Cola Consolidated is putting $35 million into a new glass bottling line in Indianapolis.
The investment will make the facility one of only three Coca-Cola plants in the country bottling in glass, with construction starting late this year. It's expected to create just 15 to 20 new full-time jobs.
That ratio tells you something. Thirty-five million dollars and fifteen to twenty jobs mean the money isn't going to headcount; it's going to equipment and infrastructure. Glass lines run tight tolerances on fill height and cap torque, and this is a brownfield project bolting new capability onto a facility that's been running since 1968. That's exactly the scenario where wiring decisions made on day one save an integrator weeks of commissioning time, or cost them weeks of rework. Read more here.
PAC Machinery, CMES Robotics, and FANUC are showing off an AI pick-to-pack system at Automate this week.
Live at the FANUC America booth, the three companies have built an integrated bagging system that combines AI-driven robotic piece-picking with PAC Machinery's automatic bagger, which runs recyclable paper mailers. It's built for fulfillment and e-commerce operations.
This is a good example of three companies solving three different problems and getting more credit together than any of them would alone. None of it works as a system unless the sensors, the vision camera, and the robot controller can all communicate reliably, which is the unglamorous part nobody puts in the press release. If you're evaluating something like this, ask your integrator how the vision system is wired into the cell before you ask anything about the AI model. The AI is only as good as the data path feeding it. Read more here.
The FIFA World Cup is giving CPG brands a rare mid-summer engagement window.
With the tournament running across 16 North American cities through mid-July, Circana data show CPG brands and retailers have a genuine shot at reaching fans in 48 countries. Nearly a quarter of consumers plan to watch matches with friends and family outside their own household.
The marketing side is straightforward: more eyeballs, more occasions, more shelf space. What's less obvious is how a demand spike like this affects a manufacturer's production schedule. Winning extra shelf placement means actually filling that shelf, which means faster changeovers and less tolerance for downtime during the exact weeks everyone wants product on hand. The marketing team gets the headline. The plant floor has to deliver on it without missing a beat. Read more here.
NVIDIA's latest survey says 9 in 10 retailers and CPG companies are increasing their AI budgets this year.
Half of those companies expect increases of 10 percent or more. Retail and CPG now lead nearly every industry on agentic AI adoption, with 47 percent either using or actively assessing AI agents.
The interesting detail buried in this survey is that the top use case isn't customer-facing chatbots; it's supply chain efficiency and throughput. That tells you where the money is actually going: demand forecasting, inventory rebalancing, operational decisions, not flashy consumer features. If you're running plant operations for a CPG manufacturer, this is the year your finance team starts asking what AI tools you're piloting, whether you've asked for them or not. Read more here.
Editor's pick: deviceWISE is showcasing industrial edge AI at Automate, live through Friday.
At Booth 4233 this week, Telit Cinterion's deviceWISE Intelligence Suite is demonstrating how factory edge devices can make autonomous decisions in real time, connecting machine data directly with enterprise systems. A featured session Wednesday afternoon digs into agentic AI deployed at the edge.
"Industrial automation is moving toward systems that must make decisions locally and act independently," is how Telit Cinterion's president of solutions put it, and that line is worth sitting with. For years, the assumption has been that intelligence lives in the cabinet, the PLC, somewhere central. Edge AI flips that. The decision gets made right where the data is generated, on the machine, on the line. That only works if the physical connection between the sensor and the edge device is solid and fast, which is the same wiring architecture conversation we've been having for years, just with a new reason behind it. If you're at Automate this week, worth a look even if AI isn't usually your department. Read more here.
One deeper thought: the skill that used to take years now ships in a box.
My 3D printer wasn't the only thing this week that quietly absorbed years of expertise into something a normal person can operate. GM's cobots, the edge AI platforms on the Automate floor, the AI piece-picking system PAC Machinery is showing, they're all doing the same basic thing. Taking a skill that used to require a specialist, calibrating a print bed, programming a robot arm, training a vision model, and packaging it into something that works close to right out of the box.
That's good news and complicated news at the same time. Good, because smaller machine builders and smaller CPG plants can now access capability that used to require a six-figure integration budget and a specialist on staff. Complicated, because it changes what counts as a defensible skill. The UAW's concern about cobots isn't really about robots. It's about what happens to the value of human expertise when machines start doing the part of the job that used to require training and judgment.
I see the same dynamic in connectivity. The tribal knowledge problem, the retiring electrician nobody can replace, the controls engineer who's the only one who understands the panel, exists because the wiring method demands it. Connectorized, decentralized architecture doesn't eliminate the need for skill. It moves the skill requirement to a point where it's far easier to find, train, and replace. That's not a smaller conversation than what's happening with AI and robotics. It's the same conversation, just one I've been having for longer.
If you're at Automate this week, come find us at Booth #21030.
Have a great week, everyone.
Colin