How AI, Robots and Data Can Transform Manufacturing Amid Economic Uncertainty
Key Highlights
- Connected systems combining CMMS, SCADA, ERP and monitoring tools can be used to support trustworthy AI predictions and ROI-focused investments.
- Agentic AI is evolving beyond chatbots into collaborative partners that ask clarifying questions and initiate workflows, potentially replacing traditional enterprise software with outcome-based pricing models.
- Humanoid robots transition from science fiction to factory floors at companies like GXO Logistics and Schaeffler to address skilled labor shortages.
As the manufacturing industries enter 2026, a paradox confronts the industry in a way that hasn’t been encountered recently. One component of this paradox is something industry has become accustomed to: technology continues to advance at breakneck speed. The other component is one industry is not so well accustomed to: the current economic headwinds, with a significant contributing factor being the tariffs implemented by the Trump administration.
The last time tariffs were implemented to the degree we’re seeing now was in 1930 with the Smoot-Hawley tariffs.
In addition to the tariffs are the realities of flat growth and large degrees of economic uncertainty. The argument can be made that these other conditions have been exacerbated, if not largely created, by the tariffs.
Against this challenging backdrop, transformative technologies such as agentic AI and humanoid robots appear poised to reshape what's needed and what’s possible on the factory floor.
Of course, I’m not suggesting that manufacturers chase after these new technologies blindly to secure a stronger foothold. But it has certainly become clear that success will increasingly belong to manufacturers who build solid data foundations first, then strategically deploy emerging technologies to solve specific business problems. With that in mind, let’s take a look at what several industry experts expect for 2026.
The rise of physical AI and intelligent robotics
Perhaps the most significant shift in 2026 will be AI's migration from the digital realm into physical spaces. Referencing this “physical AI”, Anthony Vetro, president and CEO of Mitsubishi Electric Research Labs, said this type of artificial intelligence is one that must contend with the friction, inertia, heat and the unpredictable realities of the physical world.
Success requires treating hardware, AI, data infrastructure and human workflows as inseparable elements of a unified strategy.
He predicts that industrial companies should expect to see physical AI “moving beyond the digital world” starting slowly in 2026. Therefore, it’s good time to start planning early for how this could impact your operations.
Such movement may already visible in the robotics sector. Global installed industrial robot capacity could reach 5.5 million units by 2026, according to Deloitte's analysis. However, annual new robot sales have plateaued at just over half a million units since 2021, suggesting the market is at an inflection point rather than experiencing explosive growth.
This potentially indicates that a breakthrough may come from humanoid robots, which transitioned from the realm of science fiction to factory floors in 2025, with companies such as GXO Logistics and Schaeffler. As the skilled labor shortage continues to burden manufacturers, Vetro noted that these robots are being deployed for production monitoring, accessing difficult areas and alerting teams to malfunctions.
As Automation World recently reported, Adrian Stoch, CEO Americas at Hai Robotics and former chief automation officer at GXO Logistics, believes humanoid robots represent the most transformative technology to hit manufacturing and warehouse operations in decades.
Effective buy-in and training will be the difference between checking a technology box and achieving sustained, impactful improvement across operations.
Stoch stressed that what makes humanoids so revolutionary isn't their human-like appearance, it's their versatility. Unlike traditional automation that excels at single, repetitive tasks, humanoids represent what Stoch called "the only category of technology that can have the same units performing different tasks throughout the day."
And he’s concerned that the U.S. may fall behind on this front, not for technological reasons, but cultural and economic reasons.
"We're years away from seeing humanoids in wide deployment in the U.S., but not in China,” he said, “because there is big demand side push in China for humanoids, which will help the tech mature faster there.”
Outside the realm of humanoid robots, Dustin Snell, senior vice president, agentic solutions at Automation Anywhere sees intelligent robots bringing software-level productivity gains into physical industries to perform real-world tasks with AI-level precision. “Early on, companies may use these efficiencies to boost profits rather than cut costs, but competitive pressure will eventually drive prices down,” he said. “As AI powered robots learn to build, maintain and operate real infrastructure, the speed of progress in the physical world will start to mirror the acceleration we have already seen in software.”
The key challenge here is that, unlike digital AI, physical AI must obey the laws of physics. No matter how sophisticated the algorithms, if they don't account for real-world constraints, they won't deliver reliable results. This means manufacturers need to approach physical AI implementation with intentionality and realistic expectations.
Beyond chatbots to agentic AI
The way manufacturers interact with AI is undergoing a fundamental transformation. In 2026, expect to see agentic AI evolve beyond static chatbot experiences into real-time, self-generating intelligence software that dynamically creates user interfaces and adapts workflows on the fly.
Manufacturing's unique ability to create controlled experiments in production processes gives it a critical advantage in validating AI investments.
Snell predicts a shift from one-way prompts to two-way collaboration. Rather than passively waiting for commands, AI will act as a collaborative partner that asks clarifying questions, follows up and initiates next steps when appropriate. This will allow manufacturers to move toward instruction-based workflows that formalize what AI should do, when to escalate and how to engage with humans.
“Manufacturing companies will deploy AI agents to make autonomous operational decisions that directly impact efficiency and cost reduction — expediting product lots to meet delivery deadlines, optimizing inventory routing based on real-time demand signals and automatically routing products for quality inspection or determining optimal manufacturing sequences,” said Tim Long, global head of manufacturing at Snowflake. “This business-outcome driven approach will accelerate adoption of agentic AI in areas where automated decision-making delivers clear operational improvements and competitive advantages. Early adopters will gain significant operational benefits as these systems prove their value in controlled production environments.”
This move toward agentic AI use also has immediate implications for software development in manufacturing. For example, the mechanical aspects of software creation, such as API integration and documentation, align well with what agentic systems can now handle independently.
Developers will still lead on design and creative problem-solving, but Snell and others expect the tedious engineering layers supporting software development will shift almost entirely to agents.
As an example of this, Zafer Sahinoglu, vice president at Mitsubishi Electric Innovation Center noted that projects once requiring 10 engineers and three years to develop can now be completed by two people in months.
The impact of this extends to platforms used throughout manufacturing operations, according to Deloitte. As AI agent capabilities mature, how organizations use and invest in software could shift dramatically. Traditional seat-based and subscription licensing may give way to hybrid models blending consumption-based and outcome-based pricing. In the longer term, Deloitte indicated that sufficiently advanced agentic AI could potentially replace some existing enterprise software packages entirely.
As shop floor teams look to AI to monitor key performance indicators, visualize downtime trends and explore historical performance, manufacturers will boost their investment in their data foundation to not only understand the ‘what’ behind their operations, but the ‘so what’ behind how metrics may affect their operations.
The ongoing data foundation problem
If you’re a regular viewer of Automation’s World content, you’re very familiar with the data foundation issues holding industry back from the potential benefits of numerous advanced manufacturing technologies.
Michael Mills, technical solutions manager at Fluke Corporation captured this challenge by noting: "The predictive maintenance boom of recent years left many plants with a hard lesson: dirty data makes clever models useless."
Mills has seen teams spend months tuning algorithms only to discover fundamental misalignments such as incomplete assessments, outdated asset hierarchies and timing mismatches that create interpretation gaps.
This reality, he said, has pulled attention back to data basics.
“AI is only as effective as the data it is provided,” said Bill Rokos, chief technology officer at Parsec. “Whether through direct integration or augmentation via IIoT, businesses have emphasized ensuring their equipment and processes are capturing the datapoints AI needs to function and enable contextualized querying. As shop floor teams look to AI to monitor key performance indicators, visualize downtime trends and explore historical performance, manufacturers will boost their investment in their data foundation to not only understand the ‘what’ behind their operations, but the ‘so what’ behind how metrics may affect their operations.”
No matter how sophisticated the algorithms, if they don't account for real-world constraints, they won't deliver reliable results.
In light of this change, manufacturing reliability leaders should start by “cleaning before coding,” Mills said, to “define standardized work around data capture, clarify asset criticality and align systems long before deploying machine learning. When prediction finally arrives, it can then rest on a foundation that can carry it.”
This foundational work may be less exciting than implementing cutting-edge AI, but it's essential to deriving significant benefits from any kind of advanced analytics.
In its 2026 manufacturing industry outlook, Laurie Harbour and Mike Devereux, partners at Wipfli Advisory noted that many companies already possess all the information they need to advance their operations. The problem is that many existing systems don't talk to each other, making it nearly impossible to understand current performance or make accurate projections.
However, the technology stack is evolving to address this challenge, according to Mills. Instead of isolated CMMS, SCADA, ERP and condition-monitoring tools, companies are building connected reliability environments where every layer communicates. “These systems do more than detect faults; they also help coordinate maintenance, energy and inventory in a single loop. Predictive capability then becomes the engine for enterprise learning, not just an alert system,” he said.
Given this potential, Mills posed what could be the defining question for 2026: "How much do we trust the data we’re feeding to AI?"
Strategic technology investment in a constrained environment
Regardless of the potential new technologies hold for industry, economic realities will force manufacturers to be highly selective about technology investments in 2026. Harbour and Devereaux with Wipfli recommend evaluating technology from a strict ROI perspective focused on specific business problems that better systems can help overcome in ways that strengthen performance and financials. For this reason, they said manufacturers should focus on three core areas: operational efficiency, data readiness and sales effectiveness.
Rather than passively waiting for commands, AI will act as a collaborative partner that asks clarifying questions, follows up and initiates next steps when appropriate.
“While ROI measurement remains challenging across industries, manufacturing's unique ability to create controlled experiments in production processes gives it a critical advantage in validating AI investments,” added Long. “In 2026, manufacturing teams driving process improvements and supply chain optimizations will increasingly leverage this natural testing capability to demonstrate clear performance improvements before scaling AI deployments. The shift will move from experimental pilots to production applications only after controlled trials prove measurable outcomes — whether in defect reduction, output improvements or operational efficiency gains. This disciplined, evidence-based approach will position manufacturing as a leader in demonstrating concrete AI value, providing a model for other industries seeking to validate their own AI investments.”
Specific investment in cybersecurity was also highlighted by Harbour and Devereaux. Cyberattacks have become a major financial risk for businesses, with manufacturing experiencing an uptick in ransomware attacks that can hold systems hostage. Investing modestly in cybersecurity now can prevent catastrophic losses later.
And for manufacturers pursuing federal contracts, meeting cybersecurity standards like CMMC certification may be required even to have bids considered.
The bottom line for 2026
The manufacturers who will thrive in the near-term aren't necessarily those with the most advanced AI or the latest robots. They're the ones who have built trustworthy data foundations, aligned their technology stacks and deployed automation strategically to solve specific, high-value problems.
As Sahinoglu noted, 2026 will reward the system thinkers. Success requires treating hardware, AI, data infrastructure and human workflows as inseparable elements of a unified strategy.
“Every manufacturer is at a different stage in their digital transformation,” added Rokos. “As software tools like ERP, MES and AI continue seeing an uptick across manufacturing, the non-negotiables will come down to ensuring the people working with that tech daily fully understand it and are aligned with its role and impact. Effective buy-in and training will be the difference between checking a technology box and achieving sustained, impactful improvement across operations.”
Harbour and Devereaux contend that manufacturing leaders should exit wait-and-see mode and focus on what they can control: operational efficiency, data quality and strategic technology deployment. Those who do will position themselves to capitalize when broader economic conditions improve.
About the Author
David Greenfield, editor in chief
Editor in Chief
Leaders relevant to this article: