Blog | Robotiq

Robotiq releases TSF-85 Tactile Sensor Digital Twin on NVIDIA Isaac Sim

Robotiq Team — Tue, 02 Jun 2026 21:00:06 GMT

Also read NVIDIA's COMPUTEX coverage, where Robotiq appears alongside the latest Isaac GR00T updates.

Robotiq has released the digital twin of its TSF-85 tactile sensor in NVIDIA Isaac Sim, the first industrial-grade tactile sensor digital twin shipping on a commercial collaborative gripper. Tactile sensing promises to accelerate robotics, but its adoption has been limited by the lack of data from industry-ready hardware and accurate simulations. Model builders can now train contact-rich manipulation policies in simulation, then run them on the same physical sensor designed to operate reliably on the factory floor.

Most tactile sensors rely on a deformable contact interface. The very property that gives them sensitivity is also what makes them difficult to simulate. Deformable body simulation is technically demanding, and that is one reason accurate tactile digital twins have lagged behind in the Physical AI stack. The TSF-85 digital twin is built around that constraint. It generates synthetic tactile maps through a custom Isaac Sim UI panel, visualizes them in real time, runs data generation at the simulation refresh rate, and exports to HDF5 for downstream training pipelines.

The TSF-85 digital twin was developed by the CoRo Lab (Laboratoire de commande et de robotique) at École de technologie supérieure (ÉTS) in Montréal, a long-time research partner of Robotiq — a collaboration led by Associate Professor Jean-Philippe Roberge and doctoral researcher Berith Atemoztli De la Cruz Sánchez.

The simulation method behind the digital twin is documented in two peer-reviewed publications cited in the GitHub repo. The first, published in Frontiers in Robotics and AI (2025), draws on a dataset of 53,400 real-world tactile maps to train, validate, and test each simulation pipeline — achieving up to 97% Structural Similarity Index Measure (SSIM) for the hyperelastic model and 90% SSIM for the elastic model on 12 unseen objects. A companion paper at ICCRT 2025 releases an open dataset of 46,200 real and synthetic tactile samples, including data collected using a 2F-85 Robotiq gripper and synthetic samples generated in NVIDIA Isaac Lab.

Simulation accuracy matters, but it only helps if the real sensor remains stable over time. The TSF-85 has been tested through 2.3 million cycles at maximum gripper force, with no significant variation in the output of the tactile signals. This means models trained on its tactile data can continue to rely on consistent signals for edges, shapes, textures, and geometry even after demanding real-world use.

Robotiq has been the go-to components provider for both academic research labs and industrial production for more than a decade. That dual footprint is exactly the bridge Physical AI requires: research-grade flexibility and industrial-grade reliability in the same hardware platform.

The digital twin supports NVIDIA Isaac Sim 5.1 and is available now on GitHub: https://github.com/Lab-CORO/TSF-85. Learn more about Robotiq's physical AI stack at https://robotiq.com/physical-ai

Robotiq Launches IQ to Make Palletizing Automation Faster and More Predictable

Robotiq Team — Tue, 02 Jun 2026 14:30:01 GMT

Most manufacturers who want to automate palletizing face the same problem. Getting a straight answer on whether it fits their operation, what it costs, and how long it takes has always required weeks of back-and-forth, engineering hours, and a site visit before anyone commits to anything.

That is the problem Robotiq built IQ to solve.

Start with a Fit Check, not a site visit

Before any site visit, before any engineering hours, before any quote, IQ asks five minutes of your time.

That is enough to find out whether palletizing fits your floor, what the deployment realistically looks like, and whether the financial return makes sense for your operation, including in 1-shift operations. It is a structured assessment designed to give you a concrete answer at the start of a project rather than at the end of a long discovery process.

If it is a fit, you have a clear path forward. If it is not, you know that too, without having spent weeks finding out.

“Automation does not scale when integration remains manual.” said Samuel Bouchard, CEO of Robotiq.

See your deployment before it happens

A palletizing project has thousands of moving parts. The floor layout, the product mix and the throughput targets, to name just a few. Getting those details wrong late in a project is expensive in time and money. IQ is built to get them right early.

Once the fit is confirmed, IQ powers the full project from the know-how of over 1,000 Robotiq deployments. It captures what your operation actually looks like, connects the right people at the right moment, and generates a validated Workcell design simulated in your factory environment. Cycle time, reach, payload: every spec confirmed before anything is installed.

The gap between what was promised and what gets delivered on day one closes considerably when the work is done upfront.

Built to scale from the first line

When your first Workcell is running and you are ready to add a second line, the same path applies. No custom engineering from scratch. The know-how that built the first is already in the system.

IQ coordinates every stakeholder: your team, your local partner, Robotiq experts, through a structured digital workflow. Your partner stays central. They bring the local expertise, installation capacity, and ongoing support that keeps your lines running. IQ gives them better information and a repeatable process, so every project moves faster than the last.

"For manufacturers, this means a clearer path to automation: fewer surprises, faster decisions, more predictable performance, and better financial justification, including in many 1-shift operations." said Samuel Bouchard, CEO of Robotiq.

First seen at RUC 2026

IQ made its first appearance at the Robotiq User Conference 2026 in Québec City this week, where selected expert partners experienced it firsthand, generating Workcells and seeing what Automatic Integration looks like on real opportunities. The demonstration showed a full project path: from initial input to a running Workcell in as little as 24 hours.

IQ is available now for palletizing applications. New features will continue to be released over time. Robotiq also plans to extend the same Automatic Integration model to additional robotic applications.

“With IQ, we are moving from manually engineering robotic systems one project at a time to automatically generating workcells from real customer inputs, Robotiq components, AI, and proven know-how from thousands of past projects." said Samuel Bouchard, CEO of Robotiq.

Your project starts here

If palletizing has been on your list and you have been waiting for a clearer path to get started, this is it. The Fit Check takes five minutes, requires no site visit, and gives you a concrete answer on whether automation makes sense for your floor and your financials. No commitment, no engineering hours upfront. Just a starting point that is actually useful.

Check the fit. See the deployment. Know the return. Scale it.

Start your Fit Check at robotiq.com/iq-platform

See IQ on a live palletizing project, June 18

Join Robotiq for a first look at IQ: a keynote from CEO Samuel Bouchard and a live walkthrough on a real palletizing deployment.

Common questions about IQ and palletizing automation

What is IQ from Robotiq? IQ is an AI-enabled platform that helps manufacturers start and deploy palletizing automation faster. It begins with a five-minute Fit Check to assess whether palletizing is right for a specific operation, then powers the full project from fit to a validated, deployment-ready Workcell design.

How long does it take to get started with palletizing automation using IQ? The first step takes five minutes. The Fit Check requires no site visit and no engineering hours upfront. It gives manufacturers a clear answer on whether palletizing fits their floor and whether the financial return makes sense, including in 1-shift operations.

Does IQ work for manufacturers with multiple production lines? Yes. Once the first Workcell is deployed, the same path applies to every line after it. No custom engineering from scratch. The know-how from the first project is already in the system.

What does a validated Workcell design mean? A validated Workcell design is a deployment-ready palletizing system where cycle time, reach, payload, and all other specs have been confirmed through simulation in the manufacturer's actual factory environment, before anything is installed.

Is IQ available now? IQ is available now for palletizing applications. New features will continue to be released over time. Robotiq also plans to extend the same Automatic Integration model to additional robotic applications.

Ready to see if palletizing automation fits your operation?

✔️ Five-minute Fit Check, no commitment required
✔️ Validated Workcell design simulated in your factory environment
✔️ Predictable ROI confirmed before deployment

👉 Start your project now with the newly launched IQ platform

Why Do Palletizing Automation Projects Fail? 5 Pitfalls and How to Fix Them

a.lee@robotiq.com (Amanda Lee) — Wed, 27 May 2026 15:16:09 GMT

Palletizing automation is one of the clearest wins in end-of-line operations. The ROI is real, the labor savings are immediate, and the technology is mature. Yet many manufacturers stall out, spending months on projects that should take weeks, or deploying systems that work in the demo but struggle on the production floor.

The good news: most of these failures follow predictable patterns. Here are five pitfalls we see repeatedly, and how to avoid them, illustrated by how Molino Merano, a historic Italian flour producer, turned a tight floor, a staffing problem, and a growing product line into a 14-month payback.

Pitfall #1: Overestimating installation complexity

A lot of manufacturers never start a palletizing project because they're convinced it will mean months of production downtime, deep integration work, and a long commissioning process. That expectation, more than anything else, is what keeps manual palletizing in place long after it stops making sense.

When the solution is pre-engineered and standardized to connect with an existing line, deployment looks very different. Training is part of the package. The conveyor integration is straightforward. The commissioning period shrinks from months to days. The belief that automation is inherently slow to deploy is worth questioning before it shapes your decision. Part of what makes that possible is having the project information well organized from the start: customer requirements, site constraints, throughput targets, and layout realities all in one place, rather than scattered across emails and spreadsheets. When that groundwork is done upfront, the path from decision to running system gets much shorter.

Pitfall #2: Designing for perfect conditions

Real production floors have tight spaces, ceiling limits, layout constraints, and equipment that was installed a decade ago with no thought for what might come next. A solution engineered for a clean, open layout will always struggle when it meets a real factory.

Hardware that adapts to compact footprints and software that handles changing SKUs are not nice-to-haves. They are what determines whether a system still works two years after installation.

Pitfall #3: Not planning for variability

Many manufacturers rarely run one product. They run dozens, and that number tends to grow. A system that handles this year's SKU mix cleanly may struggle badly when a new format gets added or a customer changes their pallet specification.

Building for today's conditions without accounting for tomorrow's variability is a setup for re-engineering costs down the line. Choosing a system with flexible pattern programming, one where operators can make changes on their own, keeps the production line scalable as the business evolves.

Pitfall #4: Starting with the most complex operations

There's a logic to tackling the most complex line first. The biggest bottleneck, the highest labor cost, the most compelling ROI case. But starting with complexity adds complexity. Timelines stretch, scope grows, and the project loses momentum before it ever delivers.

A single, well-scoped project on a line with clear constraints and a realistic payback period does something a complex rollout rarely does: it finishes. This is the foundation of Lean Palletizing — start simple, build operator confidence, create the internal expertise that makes the next deployment faster and easier to approve. Start simple, prove it and then scale.

Pitfall #5: Over-engineering the solution

Customization can feel like thoroughness. The more the system is tailored to your operation, the better it should perform. In practice, highly customized systems take longer to deploy, are harder for operators to understand, and create a long-term dependency on external support for every change.

Standardized automation and proven solutions deliver faster. Operators learn them more quickly, maintain them more confidently, and own them more completely. When someone on the floor can adjust a pallet pattern or troubleshoot a fault without escalating, the system pays back more every single day. The same principle applies to the integration process itself: when the workflow for scoping, validating, and deploying a Workcell is repeatable and structured, partners can move faster and manufacturers face fewer surprises.

How Molino Merano avoided all five

Molino Merano has been producing flour products in the historic town of Merano, in Trentino Alto Adige in northern Italy, since 1985. The company had a floor space problem, a staffing problem, and a product line that kept growing. What they didn't have was time for a 12-month automation project. Here is how they worked through each of these challenges.

What pushed them to act

As the product line expanded, the end-of-line operation started showing the strain. Manual palletizing, where operators lifting and placing every box, shift after shift was slowing throughput and wearing people down. Finding staff for that kind of work was getting harder. And the production floor simply didn't have the space to bring in a traditional palletizer.

What they needed wasn't a large-scale automation project. They needed something that would fit where they had space, go in fast, and work reliably from day one.

What they deployed

Robotiq's cobot palletizing Workcell fit the floor where a conventional system couldn't. No fencing, no area scanners, just a collaborative Workcell that worked safely within the constraints of the existing layout, respecting the actual line rather than requiring the line to change around it. The solution handled multiple SKUs, allowed pallet changes without stopping production, and came with operator training built into the deployment.

What changed

The Workcell was in production within a week of installation.

As the product range had grown, so had the pressure on the team. Automating palletizing meant that pressure didn't have to grow with it. Staffing the end of the line stopped being a recurring problem. Operators moved to other parts of the operation where their time had more value.

But the change that stands out most isn't about throughput or headcount. Before the cobot, the end-of-line team was lifting every box onto every pallet, hundreds of times a day. Back pain was routine and that manual work is gone now. The physical environment at the end of the line is genuinely better, and the team feels it.

Molino Merano even reached a full return on investment in 14 months, across a footprint that fit the floor they actually had.

Questions manufacturers ask before getting started

How long does a palletizing project actually take? Weeks, not months, and the gap is closing. Molino Merano went from installation to live production in under a week. With the right information organized upfront and a structured workflow from scoping to deployment, what used to take months is becoming a matter of days. The timeline depends far more on how well the project is prepared than on the technology itself.

What if our floor doesn't have much space? That's one of the most common constraints, and a good reason to look at cobot solution specifically. They're designed for compact footprints, work without safety fencing, and can be configured around existing equipment rather than requiring the line to move around them.

We run a lot of different products. Can one system handle all of them? Yes, if the system is built for it. The key is flexible pattern programming that operators can manage themselves. If changing a pallet configuration requires a service call, that's a problem at scale.

How quickly will we see a return? It depends on volume, labor costs, and how much downtime the current operation is absorbing. For Molino Merano, with a busy multi-SKU line and real difficulty finding staff, the return came in 14 months.

What happens when something goes wrong? That depends heavily on the system you choose. Standard, pre-engineered solutions are easier to troubleshoot because operators recognize what's happening. Highly customized systems tend to create dependency on vendor support for even basic interventions. Ease of maintenance should be part of the selection criteria from the start.

Get a clear answer instantly:

✔️ Check if your application is compatible
✔️ Estimate ROI based on your inputs
✔️ Get a recommended configuration

👉 Start your evaluation now with the Palletizing Fit Tool

The economics of Physical AI: Why data quality beats scale

nicolas@robotiq.com (Nicolas Lauzier) — Thu, 14 May 2026 13:00:01 GMT

To reach the level of robustness the Physical AI community aspires to, namely generalist policies deployable zero-shot on unfamiliar objects in unfamiliar settings, dataset sizes must grow by several orders of magnitude. To give a sense of scale, extending the logic to LLM-scale data volumes, on the order of 10¹², would require roughly 80 million robots operating continuously for three years. The field is therefore bottlenecked not only by compute or model architecture, but more fundamentally by the rate at which high-quality, real-world manipulation data can be generated.

For a CFO or engineering leader, the implication is direct. The route forward is higher information density per episode rather than more robots running for more hours. A single tactile-augmented trajectory carries more training signals than several vision-only runs, particularly for contact-rich and insertion tasks.

Why scale alone breaks the budget

Physical AI does not have an internet to scrape. The largest open real-robot dataset, Open X-Embodiment, aggregates around 1 million episodes from 34 labs.¹ DROID took 50 operators, 18 robots, and 12 months to assemble 76,000 trajectories.² Physical Intelligence's π0 — arguably the most capable open generalist policy to date — required more than 10,000 hours of teleoperated data before fine-tuning.³ These efforts are formidable, and still modest by several orders of magnitude relative to what genuine generalisation requires.

If volume is the only lever, data collection cost scales linearly with fleet size and operating hours. Multiplied across 10,000 robots, that is a capital expense in the hundreds of millions of dollars before a single model has been trained.

Better sensing multiplies every robot hour

Studies of imitation learning show that robot policies improve as more training environments and objects are added to the dataset.⁴ Vision-language-action models follow the same pattern, but each new data point in robotics produces a smaller performance gain than in language modelling, a consequence of data quality heterogeneity and the scarcity of action-labelled contact-rich interactions.⁵

For a budget owner, this is the core economic insight. A shallower scaling coefficient means brute-force volume buys less performance per episode in physical AI than it does in language. Quality of data therefore matters more. Investing in better sensing hardware early is a multiplier on every hour of robot time that follows.

The Video Tactile Action Model (VTAM) put a concrete number on the multiplier, tactile-augmented policies outperformed vision-only baselines by 80% on contact-rich tasks, from just 10 minutes of teleoperation per task (covered in detail in our previous post).⁶ Well-instrumented end-effectors lead to richer episodes, which means fewer demonstrations needed, which lowers compute per training run, which speeds up iteration, which shortens time to deployment. Each link has a measurable saving.

Additional to tactile sensing, a Robotiq end-effector emits several synchronized data streams per operation cycle — force, torque, position, velocity, and gripper state — each a separate signal the policy can use to disambiguate what is happening at the contact point. Every episode produces more training signals.

What this means for the budget

A well-instrumented end-effector is an investment with a calculable return. Teams that treat instrumentation as the foundation of their data strategy ship sooner and at lower total cost. Teams that defer the investment pay for it twice, once in rebuilt datasets, and once in delayed time to production.

Talk to our technical team about sensor integration for your manipulation pipeline and learn more about how Robotiq can enable your application.

¹ Open X-Embodiment, arXiv:2310.08864 — approximately 1.0 × 10⁶ real-robot episodes spanning 22 embodiments and 500+ skills.

² DROID, arXiv:2403.12945.

³ Physical Intelligence, π0: A Vision-Language-Action Flow Model for General Robot Control.

⁴ Lin et al. (2024), Data Scaling Laws in Imitation Learning for Robotic Manipulation.

⁵ Sartor and Nießner (2024), scaling-law analysis of vision-language-action models and proprioceptive policies. See also Kaplan et al. (2020), Scaling Laws for Neural Language Models, and Hoffmann et al. (2022), Training Compute-Optimal Large Language Models ("Chinchilla").

⁶ Video Tactile Action Model (VTAM), arXiv:2603.23481.

How tactile sensing improves model performance

Jennifer Kwiatkowski — Thu, 07 May 2026 13:00:02 GMT

Vision-language-action models are the current state of the art in robotic manipulation. They still cannot pick up a potato chip without crushing it.

That is the result published earlier this year by the team behind the Video Tactile Action Model (VTAM). On a potato chip pick-and-place task — a task that demands high-fidelity force awareness, where vision alone cannot distinguish a crushing grasp from a holding one — VTAM outperformed the π0.5 baseline by 80%. Across the broader contact-rich benchmark suite, VTAM held a 90% average success rate.¹

The chip is an adversarial example, and that is precisely why it is the right test. At the point of grasp, only contact dynamics carry useful signals. Pressure, vibration, and force/torque tell the policy what is happening, correcting the visual estimation errors that vision-only models cannot detect on their own. A camera, however high its resolution, cannot do that work.

Tactile is not plug-and-play

Tactile sensors do not improve model performance on their own. Most learning pipelines today are built around vision and language; the two modalities with the largest datasets and the most mature architectures behind them. When tactile signals are appended to a vision-first pipeline without intentional design, they tend to get downweighted, drowned out, or lost in training. VTAM works because the architecture forces the model to forecast vision and tactile dynamics together, so the tactile signal directly shapes the learned policy rather than getting absorbed into vision and language. Tactile data only delivers its value when it is intelligently incorporated.

The pattern is now consistent across the literature

The chip is one end of the spectrum, a case where vision fails outright and tactile carries the signal alone. Most real-world tasks sit further along that spectrum, where vision and tactile each contribute and the synergy between them is what drives training efficiency. The pattern is now consistent across the literature.

VTAM is not alone. The ManiSkill-ViTac 2025 benchmark formalises tactile-augmented evaluation across insertion, tool use, and precision assembly tasks. Independent research on tactile sensor configurations and grasp learning efficiency² shows the same lift. Policies that combine vision with tactile feedback consistently outperform vision-only equivalents on contact-rich tasks, and tend to reach the same success threshold from fewer demonstrations.

Failure detection is the second prize

A tactile-conditioned policy registers incipient slip as a vibration signature tens to hundreds of milliseconds before the object actually moves. That window is the difference between re-grasping and a full restart — between 95% and 99% uptime on the same line. Across a fleet, the operational case becomes hard to ignore.

Failure detection is one case of a larger capability: producing accurate, high-resolution labels for what actually happened during the grasp. A binary success/fail label collapses information that the training pipeline could use. Did the grasp succeed cleanly, or did it succeed with internal slippage that the controller recovered from? Did the object settle stably, or did it shift during transport? Tactile sensing can distinguish these cases, and embedded contact perception can label them on-device, turning every episode into a more informative training example, not just the failed ones.

Figure 1. VTAM combines a language model, a predictive vision-tactile world model, and a diffusion-based action policy. From just 10 minutes of teleoperation per task, it learns to predict future actions, states, and forces — enabling contact-rich tasks such as chip pick-and-place, dynamic wiping, and stable peeling. Source: arXiv:2603.23481.

What this means for builders

Tactile sensing has moved from useful addition to defensible requirement for any team aiming at production-grade contact-rich manipulation. The question is no longer whether to instrument. It is whether to instrument now, or pay later in rebuilt datasets and recalibrated models.

VTAM put a real number on the case and other recent work keeps pointing in the same direction. The next generation of foundation models will be built on data that captures contact rather than vision-only.

Ready to take the next step?

Talk to our technical team about tactile integration for your manipulation pipeline and learn more about how Robotiq can enable your application.

¹ Video Tactile Action Model (VTAM), arXiv:2603.23481.

² Representative findings include Tactile Robotics: An Outlook (arXiv) and published work on the impact of tactile sensor configurations on grasp learning efficiency.

What is the best palletizing option for your operation?

Marc Giguère — Tue, 05 May 2026 16:00:00 GMT

The best palletizing solution depends on your production volume, budget, available space, and need for flexibility.

You can go with a fully engineered system, a cobot, or a plug-and-play setup. Each comes with tradeoffs. The key is picking what actually fits your floor, your throughput, and the return you expect.

This quick guide compares the most common palletizing solutions so you can make an informed decision and choose the right technology for your production floor.

High-volume production → centralized palletizers
Durable line automation → robotic palletizers
Flexible operations → cobot palletizers
Fast deployment → in-a-box palletizers
Scalable growth → Robotiq Lean Palletizing
Very low volume → manual palletizing

What are the main types of palletizing systems?

There are six common palletizing solutions used in manufacturing:

Engineered centralized palletizers
Engineered end-of-line robotic palletizers
Engineered end-of-line cobot palletizers
In-a-box palletizers
Lean Palletizing workcells
Manual palletizing

Each differs in throughput, cost, flexibility, and implementation complexity.

How do palletizing solutions compare?

Here is a simplified comparison of palletizing systems based on key decision criteria:

Key differences at a glance

Highest throughput: Centralized palletizers (30–80 cases/min)
Most durable: Robotic palletizers
Most flexible: Cobot and Lean Palletizing
Fastest to deploy: In-a-box palletizers
Shortest ROI: Lean Palletizing (1–2 years)
Lowest upfront cost: Manual palletizing (but highest long-term cost)

Which palletizing solution has the fastest ROI?

Robotiq Lean Palletizing workcells typically offer the fastest ROI, with a payback period of 1 to 2 years.

Centralized systems: 4–7 years
Robotic systems: long-term ROI
Cobot systems: moderate ROI
In-a-box systems: moderate ROI
Manual palletizing: no true ROI due to ongoing labor costs

Which palletizing system is best for flexibility?

The most flexible palletizing options are:

Cobot palletizers → easy redeployment, multiple SKUs
Lean Palletizing → scalable across lines, adaptable workflows

These solutions are ideal for manufacturers with frequent product changes or growth plans.

Which palletizing system is best for high throughput?

For maximum output:

Centralized palletizers: 30–80 cases per minute
Robotic palletizers: 25–60 cases per minute

These systems are best suited for high-volume, stable production environments.

Which palletizing solution requires the least space?

Smallest footprint: Lean Palletizing and in-a-box palletizers
Largest footprint: Centralized and robotic palletizers

Lean Palletizing workcells typically require only 80-150 sq ft, making them ideal for space-constrained facilities.

Palletizing solutions comparison

Criteria	Centralized	Robotic (EOL)	Cobot	In-a-box	Lean Palletizing	Manual
Best for	High-volume, standardized production	High-capacity, durable automation	Flexible, mid-volume production	Simple, standardized operations	Scalable, growing operations	Very low-volume or unpredictable production
Throughput (cases/min)	30–80	25–60	8–15	6–12	8-15	4–6
Cost	$500,000 to $2M+	$400K to $1.2M	$150K-$400K	$120K-$300K	$150K-$350K	Low upfront, high labor cost
Footprint	400–1000+ sq ft	250-600 sq ft	100–200 sq ft	80-150 sq ft	80–150 sq ft	Minimal
Flexibility	Low	Low	High	Low	High	High
Integration complexity	High	High	Moderate	Low	Low to moderate	None
Typical ROI	4–7 years	Long-term	Medium	Medium	1–2 years	Not applicable

Lean Palletizing offers the fastest ROI (1–2 years) and high flexibility, while centralized palletizers deliver the highest throughput but require higher investment and longer payback periods.

Download the comparative chart

How to choose the right palletizing system

To select the best palletizing solution, evaluate:

Required throughput (cases per minute)
Available floor space
Budget and expected ROI
Frequency of product changes
Safety and ergonomics requirements

Final takeaway

There is no single “best” palletizing system.

Choose performance → centralized or robotic
Choose flexibility → cobot or Lean Palletizing
Choose simplicity → in-a-box
Choose entry-level → manual

For most growing manufacturers, Lean Palletizing provides the best balance of ROI, flexibility, and scalability. To see which Lean Palletizing model best fits your operations, try the Palletizing Fit Tool.

Get a clear answer instantly:

✔️ Check if your application is compatible
✔️ Estimate ROI based on your inputs
✔️ Get a recommended configuration

👉 Start your evaluation now with the Palletizing Fit Tool

Why pharmaceutical manufacturers are standardizing robotic palletizing

Linnea Bruce — Thu, 23 Apr 2026 19:26:49 GMT

Pharmaceutical manufacturers are under pressure to increase output, maintain strict compliance, and protect their workforce, all within tightly controlled environments. Yet many facilities still rely on manual palletizing at the end of the line, where variability and risk are hardest to control.

As a result, more pharmaceutical manufacturers are adopting robotic palletizing as a standard part of their operations.

Rather than reacting to labor shortages, these companies are making a strategic shift toward automation to improve consistency, safety, and scalability.

So why do pharmaceutical manufacturers choose robotic palletizing?
Here are the five main reasons driving adoption.

Photo by Roberto Sorin on Unsplash

1. Building new production lines with robotic palletizing from day one

Pharmaceutical manufacturers are increasingly designing new production lines with robotic palletizing built in from the start.

Instead of retrofitting automation later, they:

Plan palletizing alongside upstream processes
Align automation with production ramp-up timelines
Reduce integration risks and delays

This approach ensures that palletizing does not become a bottleneck as production scales.

2. Meeting validation and compliance requirements with consistent palletizing

In pharmaceutical manufacturing, every process must be repeatable and validated. Manual palletizing introduces variability that makes compliance harder to maintain.

Robotic palletizing helps manufacturers:

Standardize pallet patterns and handling
Maintain consistent cycle times
Support validation with repeatable operations

For many teams, robotic palletizing is not just about efficiency—it is about ensuring the process meets regulatory expectations.

3. Improving worker safety with robotic palletizing systems

Manual palletizing is physically demanding and often performed in constrained environments. This increases the risk of injury and operational disruption.

Pharmaceutical manufacturers adopt robotic palletizing to:

Reduce repetitive lifting and strain
Improve safety in controlled environments
Allow operators to focus on higher-value tasks

In many cases, improving safety is a key factor in gaining internal approval for automation projects.

4. Reducing implementation risk with proven robotic palletizing solutions

Even when the ROI is clear, pharmaceutical companies prioritize risk reduction before investing in automation.

To move forward, teams often require:

On-site demonstrations
Proven use cases in similar facilities
Confidence in system reliability and validation

Robotic palletizing solutions that are simple, proven, and easy to deploy are far more likely to be approved and implemented.

5. Scaling robotic palletizing across multiple lines and facilities

Once a pharmaceutical manufacturer successfully deploys robotic palletizing, the next step is standardization.

Teams look to:

Replicate the solution across production lines
Standardize operations between facilities
Simplify training and maintenance

This makes robotic palletizing a long-term strategy for improving productivity and adaptability across the organization.

From manual palletizing to scalable automation

Manual palletizing often becomes a hidden constraint in pharmaceutical manufacturing. It limits throughput, introduces variability, and creates safety risks.

Robotic palletizing transforms this process by enabling:

Consistent, repeatable operations
Safer working conditions
Scalable production as demand increases

How to implement robotic palletizing in pharmaceutical manufacturing

To successfully adopt robotic palletizing, pharmaceutical manufacturers need solutions that are:

Quick to deploy and easy to integrate
Simple for operators to use without robotics expertise
Compact enough to fit existing layouts
Supported by local experts

Robotiq’s palletizing solutions are designed to meet these needs, helping manufacturers automate quickly while maintaining compliance and flexibility.

See how pharmaceutical manufacturers are closing the gap after serialization

Understanding why pharmaceutical manufacturers adopt robotic palletizing is one thing. Solving the challenges that come after serialization is another.

As serialization and aggregation become standard, many manufacturers struggle to maintain traceability and efficiency at the end of the line. This is where palletizing becomes a critical step—not just for handling products, but for preserving data integrity all the way to the pallet.

In our upcoming webinar with OPTEL, we’ll explore how pharmaceutical manufacturers are:

Connecting serialization, aggregation, and palletizing into one continuous process
Maintaining traceability from unit to pallet
Eliminating end-of-line bottlenecks without compromising compliance

You’ll gain practical insights into how to design a palletizing process that supports both operational performance and regulatory requirements.

Save your spot for the webinar with OPTEL.

Why Physical AI isn't scaling yet, and what's holding it back

Linnea Bruce — Tue, 21 Apr 2026 13:00:04 GMT

Physical AI is advancing quickly.

AI models can now recognize objects, plan actions, and adapt to new tasks. But despite this progress, most systems still struggle to scale in real-world environments.

Two core challenges explain why:

Limited real-world dexterity
High cost and complexity of deployment

Until these are solved, Physical AI will remain difficult to scale beyond controlled applications.

What is Physical AI?

Physical AI refers to AI systems that can perceive, decide, and act in the real world through physical interaction.

Unlike digital AI, Physical AI must handle:

Uncertainty in the environment
Variability in objects and materials
Real-time feedback during physical contact

To work reliably, Physical AI systems must combine:

Perception (vision, sensors)
Decision-making (AI models)
Action (robot motion)
Adaptation (force and tactile feedback)

Why isn’t Physical AI scaling today?

Physical AI is not scaling because most systems:

Struggle to handle real-world variability
Require complex and costly integration
Depend on precise conditions to function
Lack real-time adaptability during interaction

In short, they work in demos, but not consistently in production.

The gap between Physical AI demos and real-world deployment

In controlled environments, everything is predictable.

In real-world applications, variability is constant:

Parts are slightly different
Lighting changes
Objects shift during handling
Contact forces are uncertain

This gap between controlled conditions and real environments is where most Physical AI systems fail.

Bottleneck #1: Real-world dexterity in robotics

What is robotic dexterity?

Robotic dexterity is the ability to manipulate objects reliably despite variation in shape, position, and physical properties.

This includes:

Picking different objects
Handling uncertain orientations
Adjusting grip during motion
Managing friction and deformation

Why is dexterity hard to achieve?

Most systems rely on:

Precise positioning
Detailed planning
Limited feedback during contact

This makes them fragile when conditions change.

Common (but limiting) approach: more complexity

To improve dexterity, some systems add:

Multi-fingered robotic hands
Advanced grasp planning algorithms
High-dimensional control

The problem:
More complexity often leads to:

Higher cost
Longer deployment time
Lower robustness in production

A better approach: Simplifying robotic manipulation

Instead of increasing complexity, scalable systems simplify interaction.

Adaptive grippers and compliant designs help by:

Conforming to object shapes
Absorbing positioning errors
Reducing reliance on precise planning

Key idea:
Shift complexity from software to hardware.

This improves reliability without increasing system burden.

Bottleneck #2: Scaling Physical AI across deployments

Even when a system works once, scaling it is difficult.

Why is scaling robotic systems hard?

Because every deployment introduces variation:

New product types
Different layouts
Changing lighting
Operator differences

If each setup requires reprogramming or expert tuning, scaling becomes too expensive.

What makes a Physical AI system scalable?

A scalable system is one that can be deployed repeatedly with minimal effort.

Key characteristics of scalable robotics systems:

Works across variation without major changes
Requires minimal expert intervention
Maintains consistent performance
Has predictable deployment time and cost

Why repeatability matters more than capability

A system that works once is not enough.

The real value comes from systems that:

Work consistently
Can be replicated across sites
Require little customization

Scalability = repeatability at a sustainable cost.

How to make Physical AI systems more scalable

To enable scaling, systems must be designed differently.

Best practices for scalable Physical AI:

Design for variability, not perfect conditions
Use sensing to adapt instead of pre-programming everything
Reduce system complexity wherever possible
Use hardware to absorb uncertainty

The goal is not to eliminate variability, but to handle it effectively.

The role of force and tactile sensing in Physical AI

Why is sensing critical for Physical AI?

Force and tactile sensing allow robots to:

Detect contact in real time
Adjust grip dynamically
Handle uncertainty without reprogramming

This enables systems to adapt during execution—not just before.

How sensing improves scalability

With proper feedback, robots can:

Generalize across different setups
Reduce dependency on precise inputs
Minimize manual adjustments

This is essential for scaling across applications.

From one successful robot cell to many

A scalable Physical AI solution is not defined by a single success.

It’s defined by how easily that success can be repeated.

If each deployment requires starting over, the system doesn’t scale.

The future of Physical AI: Simpler systems that scale

The next phase of Physical AI won’t be driven by more complex AI alone.

It will come from:

Simpler, more robust system design
Better integration of sensing and hardware
Reduced dependency on ideal conditions

The systems that scale will be the ones that:

Handle variability
Deploy quickly
Deliver consistent results

Closing thought: Physical AI must scale to deliver value

Physical AI has the potential to transform robotics.

But impact won’t come from isolated successes.

It will come from systems that scale across real-world environments.

From:
“What can this system do?”

To:
“Can this system scale?”

Because real impact comes from repeatable deployment rather than one-time performance.

Ready to make your robotics application scale?

If you're working on a robotics application and facing challenges with reliability, variability, or deployment at scale, you're not alone.

Talk to a Robotiq expert to explore practical ways to simplify your system, improve robustness, and move from a working concept to a scalable solution.

👉 Get in touch with our team to discuss your application

AI can decide. But can it act? The missing layer in Physical AI

demers@robotiq.com (Louis-Alexis Demers) — Thu, 16 Apr 2026 17:51:04 GMT

Artificial intelligence has made impressive progress.

Models can classify images, generate text, and even plan complex sequences of actions. But when you take AI out of the digital world and place it into a factory, a warehouse, or any physical environment, something breaks.

The AI can decide.

But it can’t reliably act.

This is the gap that defines Physical AI—and it’s where most real-world robotics projects succeed or fail.

The gap between thinking and doing

In simulation, everything is clean and predictable.

Objects are perfectly modeled. Lighting is ideal. Physics behaves exactly as expected.

In the real world, none of that is true.

Parts vary slightly from one batch to another
Surfaces reflect light differently throughout the day
Objects shift, slip, or deform during handling
Contact forces are uncertain

An AI system might correctly identify an object and decide how to pick it. But without the ability to adapt during the interaction, that decision often fails in execution.

This is why many AI-driven robotics demos look impressive—yet struggle when deployed on the factory floor.

Perception isn't enough

Most AI development in robotics has focused on vision.

And vision is important. It helps robots locate objects, understand scenes, and plan actions.

But vision alone doesn’t close the loop.

Humans don’t rely only on sight to manipulate objects. We use touch, force, and feedback continuously:

We adjust our grip when something starts slipping
We feel contact before applying force
We adapt instantly to small variations

Without this feedback, even simple tasks become unreliable.

The same is true for robots.

Physical AI requires a full loop: sense → decide → act → adapt

To operate reliably in the real world, robots need more than intelligence. They need a closed-loop interaction system.

That loop looks like this:

Sense – Vision, force, and tactile inputs
Decide – AI models or control logic determine the action
Act – The robot executes the motion
Adapt – Real-time feedback adjusts the action during execution

Most current systems stop short of this loop.

They sense and decide, but don’t adapt effectively once contact begins.

That missing “adapt” step is where failures happen.

Why manipulation is still the hardest problem

Moving a robot arm from point A to point B is a solved problem.

Interacting with the real world is not.

Grasping, inserting, aligning, or handling objects introduces uncertainty that AI alone cannot resolve.

The challenge isn’t just planning the motion. It’s handling what happens during the motion:

Slight misalignment during insertion
Unexpected resistance when pushing a part
Object slipping during a pick
Variations in material stiffness or friction

Without feedback, the robot either fails or requires extremely tight control of the environment.

And tightly controlled environments don’t scale.

The role of hardware in making AI work

There’s a tendency to treat AI as the primary driver of progress.

But in Physical AI, hardware plays an equally critical role.

Adaptive grippers, force-torque sensors, and compliant mechanisms don’t just execute actions; they make those actions more robust.

They reduce the precision required from AI models by absorbing variability physically.

Instead of needing perfect perception and planning, the system can rely on:

Mechanical compliance
Force feedback
Simpler grasp strategies

This is what enables real-world reliability.

Not perfect AI, but systems designed to handle imperfection.

From demos to deployment

The difference between a demo and a deployed system often comes down to one question:

Can the robot recover from small errors on its own?

In many AI-driven demos, the answer is no.

Everything works because the environment is controlled.

In production, variability is constant. And systems that can’t adapt require:

Frequent human intervention
Complex reprogramming
Tight process constraints

That’s where projects stall.

Physical AI isn’t just about making robots smarter. It’s about making them more resilient to reality.

W hat this means for robotics team s

If you’re building or deploying robotic systems, this shift has practical implications:

Don’t evaluate AI in isolation; evaluate the full interaction loop
Prioritize systems that can adapt during contact, not just before
Use hardware to simplify the problem whenever possible
Design for variability, not perfection

The goal isn’t to eliminate uncertainty.

It’s to handle it effectively.

Closing the gap

AI has reached a point where decision-making is no longer the main limitation.

Interaction is.

Physical AI is about closing that gap: connecting intelligence to the real world through sensing, action, and adaptation.

Because in robotics, the question isn’t just:

“Does it work?”

It’s:

“Does it still work when reality gets messy?”

Ready to take the next step?

If you're working on a robotics application and running into challenges with reliability, variability, or deployment at scale, you're not alone.

Talk to a Robotiq expert to explore practical ways to simplify your system, improve robustness, and move from a working concept to a scalable solution.

How TIDI Products increased palletizing productivity by 30% with automation

Linnea Bruce — Tue, 14 Apr 2026 12:59:18 GMT

TIDI Products, a global manufacturer of infection prevention and patient safety products, transformed its end-of-line operations by automating with Lean Palletizing. The result: measurable gains in productivity, safer working conditions, and more efficient use of labor.

This case shows how robotic palletizing can directly improve manufacturing performance with clear, repeatable results.

Key results at a glance

30% increase in palletizing productivity (first cell)
25% additional productivity gains (subsequent cells)
2× performance improvement by combining work cells
Reduced labor from two operators to one per cell
Significant reduction in strain and injury risk

The problem: manual palletizing limited production capacity

Before automation, palletizing at TIDI Products was fully manual. Operators handled multiple tasks, including:

Building boxes
Filling and sealing them
Transporting and stacking them on pallets

This created two critical bottlenecks:

Labor inefficiency
One operator often had to manage multiple stations, reducing focus and throughput.

Production delays and lost revenue
Backlogs increased, and the company was forced to decline orders.

Not only was manual palletizing slow, it prevented the company from scaling.

The solution: Robotiq Lean Palletizing

TIDI Products deployed Robotiq palletizing solutions to automate repetitive and physically demanding tasks at the end of the line.

The system provided:

Consistent palletizing performance
Flexibility across product lines
Fast operator adoption due to ease of use

This allowed the team to quickly move from manual handling to a scalable, automated process.

Result 1: 30% productivity increase in the first deployment

The first palletizing cell delivered a 30% increase in productivity.

This immediate gain came from:

Eliminating manual handling time
Reducing operator movement between stations
Increasing overall line speed

For manufacturers evaluating ROI, this shows how a single automation project can quickly unlock capacity.

Result 2: Repeatable gains across multiple cells

After the initial success, additional palletizing cells were deployed.

Each new cell delivered:

25% productivity improvements on average

This demonstrates an important principle of automation: once a process is standardized, improvements become repeatable and scalable.

Result 3: Doubling output by redesigning the workflow

Automation enabled TIDI Products to rethink its production layout.

By combining two workstations into one:

Output per cell doubled
Operator requirements dropped from two to one
Uptime increased across operations

This is a key insight: robotic palletizing doesn’t just automate tasks; it enables process redesign for higher throughput.

Result 4: reducing labor constraints while increasing efficiency

Labor challenges were a major driver for automation.

After deployment:

One operator can now manage multiple processes
Workers focus on higher-value tasks like machine operation
Overall efficiency improves without increasing headcount

This helps manufacturers maintain output even in tight labor markets.

Result 5: improving worker safety and ergonomics

Manual palletizing required repetitive lifting of boxes weighing up to 25 lbs (11kg).

This led to:

Strains and sprains
Fatigue and lost workdays

With robotic palletizing:

Injury risks were significantly reduced or eliminated
Physical strain on workers decreased
Jobs became less physically demanding

Improving ergonomics also supports long-term workforce retention.

Result 6: consistent output and improved product quality

Manual palletizing required training and still introduced variability.

Automation delivered:

Consistent pallet patterns
Reduced training time
More reliable output

This leads to better product quality and improved customer satisfaction.

Why this matters for manufacturers considering palletizing automation

This case highlights how robotic palletizing improves:

Production throughput
Labor efficiency
Workplace safety
Operational consistency

Most importantly, the results are measurable and repeatable.

Conclusion: a fast, scalable path to higher performance

TIDI Products achieved:

30% productivity gains immediately
Continued improvements with each new deployment
Safer, more efficient operations

Their advice is simple: Automate sooner to unlock results faster.

For manufacturers dealing with labor shortages, production bottlenecks, or safety concerns, palletizing automation offers a clear path to improvement.

Find out if palletizing automation is right for you

Not sure if your application is a good fit for Lean Palletizing?

Use our Palletizing Fit Tool to quickly evaluate your palletizing needs and see if automation can deliver similar results in your operation.