The early results are already measurable. At initial deployment sites, the approach increased throughput by 20% and identified up to 90% of potential design issues before any physical changes occurred, PepsiCo said. PepsiCo estimates the approach can reduce capital expenditure by 10% to 15% by uncovering capacity that already exists inside a facility rather than building new capacity to solve the same problem. “The scale and complexity of PepsiCo’s business, from farm to shelf, is massive, and we are embedding AI throughout our operations to better meet the increasing demands of our consumers and customers,” PepsiCo Chairman and CEO Ramon Laguarta said.
12-Week Pilot Replaced Months of Traditional Facility Planning
The clearest example of what the technology can do came from a 12-week pilot that combined two brownfield manufacturing sites. One ran PepsiCo’s beverage business, the other ran snacks, and the two had always operated separately. “We wanted to bring those businesses together to unlock velocity, efficiencies, capacity,” Steve Hoinka, PepsiCo’s vice president, global manufacturing strategy and transformation, said at Siemens’ Realize LIVE Americas 2026 conference. The challenge, he said, was removing part of one warehouse, sending product straight into a new mixing center, and figuring out whether the combined site could handle new production or packaging without building anything new.
Testing that setup would normally take months. Using the Siemens-Nvidia platform, PepsiCo ran thousands of configuration scenarios in just 12 weeks, all before spending a dollar on concrete, steel or equipment.
Manufacturers Are Treating Simulation as a Capital-Planning Tool
PepsiCo’s framing reflects a broader shift in how large manufacturers approach capacity decisions. Athina Kanioura, CEO of PepsiCo Latin America and the company’s global chief strategy and transformation officer, described the ambition as building toward “a world where every plant and warehouse operates as part of a single, intelligent ecosystem,” where facilities “don’t just respond to demand, they anticipate and then adapt to it,” according to the press release announcing the partnership announcement. Nvidia Founder and CEO Jensen Huang framed the shift in an industry-wide lens. “Physical industries are entering the age of AI,” Huang said. “For companies with real-world assets, digital twins are the foundation of their AI journey.”
The pilots remain limited to select U.S. facilities, with plans to scale globally as the technology matures. What distinguishes this use case from most AI deployments in manufacturing is that the return is measured not in labor saved, but in avoided capital expenditure.
A traditional expansion assumes a company needs new physical assets to hit a capacity target. PepsiCo’s bet is that a meaningful share of that capacity already exists inside the facilities it owns, and that AI is now precise enough to find it before the shovel breaks ground.
For all PYMNTS digital transformation coverage, subscribe to the daily Digital Transformation Newsletter.
