The physical AI boom is easy to picture as a new factory filled with humanoid robots and autonomous machines. The bigger opportunity may be much less futuristic. It’s making factories built decades ago smarter without replacing the equipment already inside them.
That market is taking shape as manufacturers layer artificial intelligence, sensors, digital twins and, increasingly, robots onto production systems they already own. Recent deployments at PepsiCo and BMW, and the funding flowing to startups such as Augury and Harmoni, suggest physical AI can generate returns without manufacturers rebuilding plants around automation. If plants can add capacity without new construction, that changes equipment financing, project lending and capex cycles.
AI Finds Factory Capacity Manufacturers Already Own
PepsiCo’s collaboration with Siemens shows how that works. Using Siemens’ Digital Twin Composer, built on Nvidia Omniverse, PepsiCo models a plant’s machines and workflows virtually before committing capital to physical change, the company said when it announced the partnership in January. At a U.S. Gatorade plant, the approach lifted throughput 20% in three months and flagged up to 90% of design flaws before construction began, Siemens said.
We’d love to be your preferred source for news.
Please add us to your preferred sources list so our news, data and interviews show up in your feed. Thanks!
Across its operations, PepsiCo estimates capital expenditure reductions of 10% to 15%, extracting latent capacity from existing assets rather than funding new ones, PYMNTS reported. A separate 12-week pilot that merged two legacy sites, one beverage and one snacks, surfaced “flow opportunities, spatial opportunities, things that we were not expecting to see,” Steve Hoinka, PepsiCo’s vice president of global manufacturing strategy and transformation, said at Siemens’ Realize LIVE Americas 2026 conference in June.
That turns the economics of physical AI partly into an asset-utilization story: how much more an existing factory could produce if AI understood how every machine and process interacted. Augury, which uses sensors and industrial AI to monitor existing machinery and flag emerging problems, found that 83% of the 501 manufacturing leaders it surveyed planned to increase AI investment this year, according to its State of Production Health 2026 report, published in June.
Even humanoid robots are increasingly designed around existing production environments rather than purpose-built ones. BMW tested that model at its plant in Spartanburg, South Carolina, where Figure 02 robots spent about 1,250 operating hours over 10 months in 2025 retrieving and positioning sheet metal parts for welding, the company said. The robots supported production of more than 30,000 X3 vehicles. BMW has since moved a newer Figure model into logistics work at the same plant and started a separate humanoid pilot, using Hexagon’s AEON robot, at its Leipzig plant in Germany.
Those robots plug into a factory BMW has already wired for artificial intelligence. The automaker runs more than 600 AI use cases across its business, including factory systems that inspect welds and flag defects in real time, PYMNTS reported.
Startups Target the 280,000 Factories Already Standing
Harmoni is coming at the problem from the opposite end of the market. The company raised $10 million in Series A funding Sept. 9, led by Bessemer Venture Partners, for technology built specifically for existing factories. It is targeting roughly 280,000 U.S. factories that still run traditional machinery and labor-intensive processes, Business Insider reported.
That installed base could change how the physical AI opportunity is measured. North American companies ordered 8,940 robots worth $622 million in the second quarter alone, with order value rising 21.3% year over year, according to the Association for Advancing Automation, as PYMNTS reported. But the physical AI market may ultimately be larger than robot sales alone. Digital twins can optimize old production lines. Sensors can make existing machines observable. Humanoids can fill specific gaps without redesigning the entire plant. The factory of the future may already have been built. The opportunity is making it intelligent.
Add as Preferred Source
