September 3, 2026
Process Intelligence, 60 kW Cutting, AI Quality, and Manufacturing at Sea
This week’s stories cover a wide range of laser manufacturing, but they share a common theme.
The laser itself is only one part of a production system.
Scanner control is becoming more closely connected with process monitoring. Cutting systems continue pushing to higher power. AI is being tested as another tool for understanding process quality. And laser additive manufacturing is moving out of the laboratory and into some very unusual production environments.
We also have a special MWL IMTS Watch this week as the show approaches, plus this week’s MWL Automation Note: When Is a Reusable Function Worth the Effort?
We also have two invitations for MWL readers: one for experiences From the Shop Floor, and another for consultants who may be interested in contributing practical technical articles to a future Consultants Corner.
1. Laser Process Control and Process Monitoring Move Under One Roof
Blackbird Robotics and Lessmüller Lasertechnik are merging under a new company called iQLASE.
The combination is interesting because the two companies have concentrated on different—but increasingly connected—parts of laser processing.
Blackbird specializes in remote laser welding, on-the-fly processing, scanner control, and software. Lessmüller specializes in optical process sensing, OCT technology, and in-process quality assurance. The existing Blackbird and Lessmüller names will continue to be used at the product level while the companies combine under iQLASE GmbH. (Blackbird Robotics)
Why It Matters
I think the bigger story is not really the merger.
It is what is being combined.
We have traditionally thought about scanner motion, laser control, seam tracking, process monitoring, and inspection as separate pieces of a system.
But the more these technologies communicate with each other, the more useful they become.
For example, a scanner controller knows where the beam is supposed to be. OCT or another sensor can provide information about where the joint actually is or what happened during the weld. Process-monitoring software can then connect that information with the programmed path and process parameters.
That moves us from simply controlling the laser to understanding the process as it happens.
It also connects closely with our recent MWL discussion about scanner interfaces. As scanners, sensors, and process-control systems become more integrated, the amount and quality of information moving between them becomes increasingly important.
Source: Blackbird Robotics — Combined Expertise in Laser Process Control and Monitoring
2. How Much Laser Power Is Enough? A 60 kW Fiber Laser Goes Into Production
The 3-Dimensional Services Group has installed a 60 kW G.Weike fiber laser at its manufacturing operation.
According to Manufacturing News, the machine has a 16 × 32-foot cutting area and is intended for heavy fabrication, steel structures, thick-plate cutting, and high-volume industrial applications. The article reports capability for cutting material up to 4 inches thick, along with 45-degree bevel cutting. (Manufacturing News Online)
Why It Matters
Sixty kilowatts is an impressive number.
But I think the more useful manufacturing question is:
When does more laser power actually make the part less expensive?
Higher power can increase cutting speed and extend the thickness range of fiber-laser cutting. But the laser is still only one part of the economics.
Assist gas consumption, cutting-head capability, optics, material handling, acceleration, piercing time, edge-quality requirements, downstream processing, maintenance, and machine utilization all matter.
At some point, the question stops being:
“How many kilowatts can we put on the machine?”
and becomes:
“How many good parts per hour can the complete system produce?”
For certain heavy-plate applications, very-high-power fiber lasers are increasingly worth comparing with processes that have traditionally handled heavy fabrication.
That is something worth watching.
Source: Manufacturing News — The 3-Dimensional Services Group Adds 60 kW Fiber Laser to Lineup
3. Can AI Look at a Laser-DED Build and Explain Its Quality Decision?
Researchers from the Colorado School of Mines, GAC R&D Center Silicon Valley, and Georgia Institute of Technology have investigated using vision-language models for quality assessment in wire-laser directed energy deposition.
Rather than training a completely new machine-learning model for the application, the researchers used in-context learning, giving the models a limited number of application-specific examples.
The study evaluated Gemini 2.5 Flash and Gemma 3:27b. According to the researchers, the in-context-learning approach achieved classification accuracy comparable with specially trained machine-learning models while requiring relatively few application-specific samples. The models could also provide human-interpretable explanations for their decisions. (Metal Additive Manufacturing)
Why It Matters
This is the kind of industrial AI work I find interesting.
Not because an AI suddenly replaces the process engineer.
It doesn’t.
The interesting part is that the model can potentially look at process information and say more than simply:
GOOD or BAD.
It may also be able to explain which image features influenced its decision.
That matters because one problem with many machine-learning quality systems is that they can become another black box.
If an AI-assisted system can point toward the image features that influenced its quality decision, a process engineer has somewhere to start.
There is still a large distance between a research study and a validated production inspection system.
But this points toward a practical role for AI in manufacturing:
Helping people interpret process information rather than simply producing another alarm.
Source: Metal AM — Vision-Language Models Investigated for Quality Assessment in Metal AM
4. Laser-Wire DED Goes to Sea
A recently published Naval News report highlights an unusual deployment of laser additive manufacturing.
During RIMPAC 2026, Snowbird Technologies’ SAMM Tech manufacturing system was operated aboard the USS Essex while the ship was underway from San Diego to Hawaii.
The system produced 316L stainless-steel components using laser-wire directed energy deposition with nitrogen shielding, then used integrated CNC machining to achieve the required final geometry and surface finish. The manufacturing capabilities are packaged into a ruggedized, containerized platform. (Naval News)
Why It Matters
This may be one of the clearest examples of where directed energy deposition can make sense.
It isn’t about trying to replace every machined component with a 3D-printed component.
It is about putting manufacturing capability close to where a replacement part is needed.
The system combines additive buildup with conventional machining, which is important.
DED can produce the near-net shape.
Machining can provide the dimensional accuracy and surface finish the finished component requires.
That combination may ultimately be more important than either technology by itself.
For remote factories, ships, military operations, mines, energy facilities, and other locations where waiting days or weeks for a replacement component can be extremely expensive, the value proposition becomes fairly easy to understand:
Move the manufacturing capability closer to the problem.
SPECIAL THIS WEEK: MWL IMTS WATCH
Industrial Laser Conference Comes to IMTS 2026
Instead of Product Watch this week, I thought it made sense to look ahead to IMTS 2026, September 14–19 at McCormick Place in Chicago.
Of particular interest to MWL readers is the Industrial Laser Conference, presented by the Laser Institute of America on Wednesday, September 16, from 9:30 a.m. to 6:00 p.m.
The conference program covers manufacturing applications including laser marking, cutting, welding, heat treating, remote welding, cleaning, and additive manufacturing.
The conference will be held in the West Building, Room W190-B. (IMTS)
Why MWL Is Watching
IMTS covers far more than lasers.
That is actually one reason I think it is valuable for laser manufacturing engineers.
A productive laser machine also depends on automation, motion control, robotics, workholding, inspection, software, material handling, and increasingly AI and digital manufacturing.
Seeing the laser as one part of that larger manufacturing ecosystem is important.
Over the next couple of weeks, MWL will be watching for interesting laser, additive, automation, motion, and manufacturing technologies coming out of IMTS.
Event Information: IMTS — Industrial Laser Conference 2026
MWL AUTOMATION NOTE
When Is a Reusable Function Worth the Effort?
Most controls programmers have probably written something that worked perfectly well—and still knew they didn’t really want to copy it into the next machine.
Maybe it controls a gas system.
Maybe it operates a clamp.
Maybe it performs a homing sequence, controls a laser, runs an inspection cycle, or coordinates several axes.
The first time you need it, you write the logic.
The second time, copying it may still be faster.
But eventually there is a point where building a reusable machine function becomes worth the additional effort.
What Do I Mean by a Reusable, Validated Function?
I’m using validated here in the practical machine-building sense—not as a regulatory claim.
The idea is that the function has a defined job and has been independently tested well enough that we don’t have to reinvent its basic behavior every time we use it.
A good reusable function should own the things necessary to perform that job reliably:
Inputs and outputs. Interlocks. Machine states. Fault handling. Timeouts. Diagnostics. Recovery behavior.
The part sequence shouldn’t have to understand every detail inside the function.
It should be able to say, in effect:
Do this operation with these parameters—and tell me when you are finished or why you couldn’t finish.
When Does the Extra Work Start Paying Back?
I think there are some practical warning signs.
If the same logic is being copied into several machines, if the same commissioning problems keep appearing, or if fixing one bug means remembering every place that code was copied, you are probably getting close.
Another sign is when the machine capability stays basically the same but the part sequence changes.
The machine may still need to clamp, position, turn on gas, enable the laser, perform a motion, inspect something, and release the part.
What changes from product to product is the order, parameters, or number of times those capabilities are used.
That is where separating the machine functions from the higher-level part sequence can become extremely useful.
There Is a Cost
Reusable code does not appear for free.
Someone has to define the interface, think through the failure modes, add diagnostics, test recovery conditions, document it, and resist the temptation to bury one machine’s special case inside what was supposed to be a general function.
And there is another very real problem:
The machine has to ship.
Most experienced programmers already know where they would like to clean things up.
But once the machine works, production wants it, the customer wants it, and another project is waiting.
That is how working code gets copied for years.
Don’t Standardize Everything
There is another side to this.
Not every mechanism needs a reusable software architecture.
A simple, one-off machine with a handful of I/O may be easier to understand with straightforward code.
Creating six layers of abstraction around a cylinder and two proximity switches isn’t automatically better engineering.
The goal isn’t to make the software look sophisticated.
The goal is to make the machine easier to build again, easier to troubleshoot, and easier for the next person to understand.
One More Reason This May Matter
Well-structured reusable functions could also become valuable as AI-assisted maintenance improves.
Imagine an AI maintenance tool looking at:
Function = CuttingGasControl State = WaitingForPressure CommandedPressure = 18 bar ActualPressure = 11.4 bar Timeout = Active
That information already tells a technician quite a bit.
Compare that with hundreds of poorly named bits and an alarm saying:
FAULT 27
Better structure helps the person troubleshooting the machine today.
And it may eventually help the software assisting that person tomorrow.
Working code and maintainable code are not always the same thing.
The question is deciding when the extra work required to get from one to the other finally starts paying you back.
FROM THE SHOP FLOOR
One of the things I’d like to add more often to MWL Weekly is experience from the people actually building, operating, maintaining, and improving these machines.
A clever fixture. A commissioning problem that taught you something. A maintenance lesson. A process improvement that worked—or one that didn’t.
From the Shop Floor can only work if readers contribute.
If you have a short story, lesson, photo, or manufacturing problem that other MWL readers might learn from, send me a LinkedIn message.
It doesn’t need to be polished. We can work on that part.
INTERESTED IN CONTRIBUTING? — CONSULTANTS CORNER
There is another group I would like to hear from: independent consultants and specialists working in manufacturing.
Many consultants have years of very specific experience in areas such as laser processing, welding, cutting, automation, motion control, machine safety, additive manufacturing, fixturing, inspection, maintenance, and process development.
That experience can be valuable to MWL readers—and sharing it can also help consultants introduce their expertise to manufacturers who may need it.
If there is enough interest, I would like to introduce an occasional section called:
Consultants Corner
The idea is simple.
Consultants can submit a short, useful article based on something they know well.
It might explain a common problem, share a lesson learned, point out something manufacturers often overlook, or offer practical advice that could help someone make a better engineering or purchasing decision.
The emphasis should be on teaching, not selling.
We won’t publish advertisements disguised as articles. But contributors would receive a brief author identification, their company or consulting practice name, and a way for interested MWL readers to learn more about their work.
There is no charge to submit an article, and publication will be based on whether the subject is relevant and useful to MWL readers.
If you are a consultant working in laser materials processing, advanced manufacturing, automation, motion control, additive manufacturing, machine safety, inspection, maintenance, or a related field, and you have something worth sharing, send me a LinkedIn message.
If we receive good contributions, Consultants Corner may become a regular part of MWL Weekly.
Closing Thought
This week we have scanner control moving closer to process intelligence, fiber lasers moving toward 60 kW, AI being taught to interpret laser-deposition quality, and a complete laser-DED manufacturing system operating aboard a ship.
Different technologies, but the direction seems similar:
Manufacturing systems are becoming more integrated—and the value increasingly comes from what the complete system can accomplish, not from any one component.
And as MWL grows, I’d also like more of what appears here to come from the people doing the work.
If you have a shop-floor lesson, send it.
If you are a consultant with useful knowledge to share, send that too.
I would be interested to hear what you are seeing.
Manufacturing With Light