Companies are moving beyond individual robotics and artificial intelligence projects toward production systems that can monitor quality, adjust operations and coordinate increasingly complex manufacturing processes with limited human intervention. This past week, Siemens and Procter & Gamble announced the global expansion of an AI inspection system. Likewise, Unilever announced that it is expanding its use of AI powered digital twins to optimize manufacturing operations, PepsiCo is deploying AI and simulation technology to increase factory throughput, and BMW is testing humanoid robots on automotive production lines.

Meanwhile, pharmaceutical companies are exploring autonomous laboratories that could transform how new drugs are developed and manufactured. Together, these developments offer a glimpse into how industrial AI is moving beyond isolated applications toward increasingly intelligent, adaptive and autonomous operations throughout the world’s factories.

For manufacturers, the implications extend far beyond operational efficiency. Persistent workforce shortages, pressure to bring production closer to domestic markets and increasingly competitive product development cycles are changing the economics of automation.

Companies that once evaluated robotics primarily as a way to reduce labor costs are beginning to consider whether more autonomous production could determine their ability to compete. The opportunity extends from consumer goods and automotive manufacturing to pharmaceuticals, defense and other industries where production capacity and speed have become commercial priorities. Autonomous manufacturing might soon shift from a technological possibility to an economic necessity.

Manufacturing Is Moving Beyond Individual AI Applications

This past week, Siemens and Procter & Gamble announced the global expansion of an AI inspection system that has reduced manufacturing scrap by 10% to 20%. New installations can be commissioned five to ten times faster than traditional custom vision systems, according to the companies.

Chris Stevens, president of U.S. Automation at Siemens, believes the capabilities required to build highly autonomous factories already exist for certain applications. In an interview conducted at the Ai4 2026 conference in Las Vegas, he described customers approaching Siemens with requests for complete manufacturing operations.

How quickly companies adopt these technologies, and what happens to manufacturers that cannot afford to follow, may prove more consequential than the automation itself.

The Siemens and P&G collaboration illustrates an important change in how manufacturers approach industrial AI. P&G initially developed its own deep learning inspection technology, deploying it internally more than 150 times before partnering with Siemens to expand its capabilities and support broader deployment.

The jointly developed Visual Inspection Cockpit analyzes live camera images of products moving through manufacturing lines. It identifies defects in flexible materials, textured surfaces and complex packaging that can challenge conventional machine vision systems.

Inspection results are processed on computing equipment positioned close to production machinery, allowing the system to trigger alerts or automatically remove defective products without interrupting production. Siemens supplies the industrial computing platform, software infrastructure and tools that allow P&G to replicate the technology throughout its manufacturing operations.

The use of computer vision in assembly and inspection is an already long-proven technology with established implementation. What’s new here is that P&G had already demonstrated that AI could solve difficult inspection problems. The collaboration addresses the commercial challenge of reproducing that success throughout a global manufacturing network.

Other consumer goods manufacturers are making comparable investments.

In June, Unilever announced plans to expand its use of AI powered digital twins through a partnership with Accenture. The company intends to develop more than 40 additional digital twins over 18 months, allowing manufacturing teams to monitor equipment, simulate production scenarios and identify operational problems.

At its deodorant manufacturing facility in Raeford, North Carolina, Unilever reports that a digital twin predicts 95% of process flow restrictions, contributing to a 20% reduction in waste and a 10% increase in production capacity.

Adam Raeburn-James, Unilever’s global vice president for digital business operations, characterized expanding AI throughout the company’s factories as a commitment to product quality, sustainability and employee capabilities.

The investment reflects a broader industry shift. Manufacturers are beginning to view AI as an operational system capable of connecting production equipment, quality control and manufacturing decisions, rather than another isolated software application.

From Adaptive Manufacturing To Autonomous Production

Stevens described the industry’s progression through three stages of manufacturing technology.

Traditional automation performs predetermined tasks, such as transferring components between production stations. Adaptive manufacturing introduces systems capable of responding to changing conditions, including visual inspection equipment that identifies defective products and directs machinery to remove them.

Autonomous manufacturing extends those capabilities by coordinating equipment, material movements and production processes through centralized systems that require less continuous human intervention.

“You can build an autonomous factory today,” Stevens said.

He qualified that assessment according to the manufacturing application. Certain processes can already be automated extensively, but capital requirements, technical complexity and operational constraints limit what companies can implement economically.

The transition is becoming visible in industries beyond consumer packaged goods.

BMW has been testing humanoid robots in automotive production. During a 2025 pilot at its Spartanburg, South Carolina, facility, a Figure AI robot assisted in manufacturing more than 30,000 BMW X3 vehicles, handling over 90,000 components. In 2026, BMW expanded its humanoid robotics program through a pilot at its Leipzig plant, exploring applications in battery assembly and component production.

While these projects do not establish that entire automotive factories are becoming autonomous, they demonstrate how increasingly capable robots are moving from controlled demonstrations into actual manufacturing operations.

The longer term opportunity involves coordinating those capabilities through systems that can make production decisions based on changing conditions.

Pharmaceuticals Are Exploring A More Radical Transformation

For pharmaceutical companies, the commercial implications could be substantial. Automating repetitive laboratory activities may allow researchers to conduct more experiments, generate data faster and reduce the time required to transfer successful processes into production.

The industry is already making considerable investments in related technologies.

In January, Eli Lilly and Nvidia announced plans to invest up to $1 billion over five years in a joint AI research laboratory. Their collaboration will combine pharmaceutical research with accelerated computing, robotics and physical AI to advance drug discovery and production.

For pharmaceutical companies, autonomous laboratories offer the prospect of connecting computational drug discovery with experimental validation and eventual production. The challenge is integrating those activities into reliable operations that meet pharmaceutical quality and regulatory requirements.

The Competition Is Shifting Toward Entire Manufacturing Systems

The implications extend to the companies supplying manufacturing technology.

Stevens said customers are increasingly asking Siemens to deliver integrated manufacturing capabilities rather than individual automation components.

“You have all the pieces, how far do you want to go?” he recalled customers asking.

That changes the supplier relationship. Building a production facility requires coordinating engineering software, industrial controls, electrical infrastructure, robotics and equipment from numerous manufacturers.

Companies capable of integrating those technologies into complete manufacturing operations could capture a larger share of industrial investment.

The shift is visible in the U.S. defense sector. In August, Siemens Government Technologies announced an $80 million Army contract to develop an advanced manufacturing environment at Anniston Army Depot in Alabama.

The project will establish production capabilities for electric motor stator assemblies used in unmanned aircraft systems. Siemens will use AI driven digital twins to design and validate the manufacturing process before commissioning physical production lines.

Even though the project is not a fully autonomous factory, it illustrates how customers are asking technology suppliers to deliver integrated production capabilities, including the digital infrastructure required to build and operate them.

The same approach is appearing in other industries.

PepsiCo announced a collaboration with Siemens and Nvidia in January to use AI and digital twins to simulate factory and warehouse operations before implementing physical changes. At an initial U.S. deployment, PepsiCo reported a 20% increase in throughput. The company identified opportunities to reduce capital expenditures by 10% to 15% through virtual validation and improved utilization of existing manufacturing capacity.

Rather than investing immediately in additional machinery or facilities, manufacturers can use simulation to identify bottlenecks and determine whether existing infrastructure can support greater production.

Automation Could Solve The Labor Shortage And Create A New Employment Challenge

The accelerating interest in automation is occurring amid persistent workforce constraints.

A 2024 study by Deloitte and The Manufacturing Institute projected that American manufacturing could require approximately 3.8 million additional workers between 2024 and 2033. As many as 1.9 million positions could remain unfilled if manufacturers fail to address recruitment and skills challenges.

Stevens described automotive executives discussing production lines that could not operate at full capacity owing to difficulty finding workers.

Under those circumstances, automation becomes a means of maintaining production and expanding capacity, rather than simply replacing employees.

But Stevens identified a more complicated consequence. Once manufacturers introduce robots to fill vacant positions, they may demonstrate that automated operations can produce goods faster and at lower cost than conventional facilities.

Competing manufacturers would then face pressure to make comparable investments, regardless of whether their existing factories have labor shortages.

“What happens to companies that don’t introduce their robotics? Do they stay in business?” Stevens asks.

For manufacturers still evaluating isolated AI projects, the central question is becoming whether their current approach can keep pace with competitors building increasingly connected, adaptive and autonomous operations.