IMTS 2026 Highlights: Automation, AI and Robotic Machining Trends

IMTS 2026, held at McCormick Place in Chicago from 14 to 19 September 2026, drew 91,295 registrants, including 14,447 students, from 114 countries, with 1,788 exhibitors on 1.17 million square feet. The dominant themes were AI and automation sold as practical tools: robots tending machines, cobots on mobile robots, AI-generated toolpaths and fixtureless part picking. Strong US machine-tool orders set the mood.

Checked 5 October 2026 against AMT's final IMTS report, its opening-day release and exhibitor announcements.

IMTS 2026 in numbers

IMTS, the International Manufacturing Technology Show, is organised every two years by AMT, the Association For Manufacturing Technology. On opening morning AMT counted 80,697 registrants and predicted more than 90,000; the final report put the total at 91,295.

The order data explain the mood. AMT's US Manufacturing Technology Orders report, released the day before the show, put July 2026 orders at US$605.8 million and the year to July at US$4.03 billion, up 37.1% on the same period of 2025. AMT noted it was only the second time since 1998 that monthly orders had stayed above US$500 million for five months in a row. In its closing report AMT also cited a shortage of about 435,000 skilled workers in durable-goods manufacturing.

What were the main IMTS 2026 trends?

Robots tending machines, and robots on wheels

FANUC America's show plan centred on a new R-50iA Compact DC controller that lets its CRX collaborative robots be mounted on autonomous mobile robots, shown serving a new ROBODRILL D74CS machining centre. The idea is one robot moving between several machines rather than one robot bolted to each.

Okuma brought machines with nine of eleven connected to automation, and showed an AI robot cell with Intrinsic, Google's robotics group. Intrinsic's message at the show was that AI with a camera removes the need for part fixtures and for reprogramming the robot for each new part, which is what makes automation viable for high-mix, low-volume work. A week later Intrinsic open-sourced the underlying software; read Intrinsic Core open source: what it means.

AMT's report said Okuma's sales gong rang more than 100 times by Friday morning, including a 24-machine order that broke the gong.

Physical AI and digital twins

FANUC also presented its work with NVIDIA, Google Cloud and AWS on physical AI: robots that perceive, decide and act, trained partly in photorealistic simulations of the factory. Its stand included AI-powered toolpath generation, program generation from user instructions and an autonomous failure recovery demonstration. Our explainer on digital twins in manufacturing covers what a twin is, and is not, in practice.

AI in programming and CAM

Software exhibitors pitched AI as a time saver for CNC programmers. CloudNC, for example, claimed its CAM Assist product cuts programming from days to minutes. Claims like that are vendor figures; the engineering point is that AI output still has to be checked by someone who understands feeds, speeds, workholding and collision risk.

Is robotic machining ready for the mainstream?

Robotic machining, a robot arm holding a spindle to cut, drill or trim, got more floor space than at previous shows. It suits large parts and work where a robot's reach and lower cost beat a large CNC machine. The trade-off is stiffness: a six-axis arm deflects under cutting load far more than a machine tool, so it suits trimming, deburring, drilling and lighter milling, not tight-tolerance finishing. Expect it to grow in aerospace, composites and large fabrication rather than replace machining centres.

What IMTS 2026 means for engineers outside the US

IMTS is a US show, and much of the messaging was about reshoring American manufacturing. The technology is global. The same robot brands, CNC controls and software are sold in Germany, the UK, Japan, Mexico and Brazil, and the same workforce gap shows up in each. A few points travel well:

• Machine tending is the entry point. Most small and mid-sized shops automate loading and unloading first. Engineers who can make a robot and a CNC talk safely are in demand everywhere.
• Data needs a common language. Connected machines only help if their data can be read. MTConnect, an open standard backed by AMT, is common on US machine tools; OPC UA, with its machine-tool companion specifications, is more common in Europe. Learning both is sensible. Our OPC UA explainer is a good start.
• Safety rules differ, the principles do not. A robot cell in the US is typically built to ANSI/RIA standards; in the EU and UK to ISO 10218 and the machinery rules. Guarding, risk assessment and interlocks are the same discipline.

MTConnect and OPC UA in one table

A worked example: why shops automate tending first

The business case most shops at IMTS were weighing is simple arithmetic. Take a machining centre that cuts for 32 hours a week on one shift, because operators are also loading parts, measuring and setting up. Add a robot that loads and unloads from a stocked drawer or pallet system, and let it run unattended for a further three hours after the shift ends, five days a week:

• Spindle hours rise from 32 to 47 a week, an increase of about 47%.
• No extra operator is needed for those hours, but someone must stock the cell, check first-off parts and recover from faults.
• The robot is only productive if parts, tools and workholding are reliable enough to run without a person watching.

The figures are illustrative, not from any exhibitor. The point is the shape of the gain: automation adds hours more than it adds speed, and it moves the work from loading parts to preparing, programming and maintaining the cell. That is technician and engineer work.

What the show did not settle

Trade shows are good at showing what is possible and less good at showing what is reliable. Three questions were left open, and they are worth asking any supplier who quotes IMTS demonstrations:

• How often does the AI get it wrong? AI-generated toolpaths and programs save time only if checking them is faster than writing them. Ask for scrap and crash data from real customers, not demo parts.
• Who supports a mixed cell? A robot from one maker, a machine from another, AI software from a third and a PLC from a fourth makes a capable cell and a difficult support call. Agree who owns faults before you buy.
• How is the connected machine secured? Every machine that reports data or accepts remote programs is a network device. The 2026 attacks on internet-exposed PLCs, covered in our post on the 2026 PLC cyber attacks, are a reminder that remote access needs the same care on a machine tool as on a PLC.

None of these is a reason to wait. They are reasons to pilot one cell, measure it honestly and train the people who will run it before scaling up.

Which skills does the floor actually need?

Every robot cell and connected machine at IMTS needs people to install, program and maintain it. That is the practical message behind the 435,000 figure AMT cited, and it matches what our manufacturing skills gap 2026 analysis found. The skills in order of how quickly they pay off:

1. Robot programming on a real controller's language. FANUC is the most common brand on machine-tool stands. Start with our step-by-step first FANUC teach pendant program.
2. Offline programming and simulation, so you can check reach, cycle time and collisions before anything is bolted down.
3. Machine vision, for fixtureless picking and inspection.
4. Integration: the CNC-robot handshake, safety circuits and the PLC that coordinates the cell.
5. Data and AI literacy, enough to judge AI-generated toolpaths and to get machine data into a dashboard.

Learn it free

The free FANUC Robot Programming course covers the teach pendant and ROBOGUIDE. RoboDK Offline Programming teaches cell simulation across robot brands, Machine Vision and Quality Inspection covers cameras, lighting and AI inspection, and AI for Industrial Automation shows where AI fits in a plant. The next big automation show at McCormick Place is PACK EXPO International 2026. The robotics learning path sets them in order, and our Industry 4.0 guide puts them in context.

The courses are free. An optional EDWartens Certificate of Completion, from US$2.99, can be verified at edwartens.com/verification; it is not a vendor certification and is not accredited. When you are ready, browse machine-building and robotics roles on our job board.

Take the free course

Robotics · Beginner · Free

FANUC Robot Programming with the Teach Pendant and ROBOGUIDE

Robotics · Beginner · Free

RoboDK Offline Programming and Simulation

AI and machine learning · Intermediate · Free

Machine Vision and Quality Inspection

AI and machine learning · Intermediate · Free

AI for Industrial Automation