PepsiCo rolls out AI across China operations

PepsiCo is rolling out AI across its operations in China and other markets as it looks to improve efficiency, Asia-Pacific chief Anne Tse said in an interview.

Tse said PepsiCo is using AI for precision agriculture, manufacturing management, and consumer analytics, allowing it to expand capacity without proportional increases in headcount, though it is still hiring in China as it opens new plants.

She said the company is adjusting its China portfolio by targeting premium, nutrient-dense products, reducing sugar and sodium while blending local ingredients to align with Chinese culinary traditions, while managing geopolitical and cost risks by sourcing 95% of ingredients locally and using commodity hedges.

PepsiCo announced in January a multiyear collaboration with Siemens and Nvidia to apply AI and Siemens’ digital twin technology to plant and supply chain operations.

🔗 Source: Bloomberg

🧠 Food for thought

Implications, context, and why it matters.

PepsiCo has been building its AI strategy for years

Before the recent work with Siemens and Nvidia, PepsiCo was already using AI across the business.
Research and development (R&D) teams use it to read social media trends, which sped up new items such as seaweed snacks and immunity-boosting Propel drinks 1.
On factory lines, a system checks Cheetos for consistent shape, crunch, and the “Cheetle” cheese dust coating 1.
At some Frito-Lay plants (PepsiCo’s snack-food manufacturing division), predictive maintenance stopped unexpected machine breakdowns for a full year and moved work to scheduled interventions 1.

The industrial metaverse focuses on getting more from physical assets

The Siemens and Nvidia work aims to unlock “hidden capacity” in existing facilities, many of them decades old and hard to upgrade in the real world 2.
PepsiCo uses digital twins (virtual replicas of real-world facilities and processes) to test and confirm facility layouts, catching up to 90% of potential issues before physical changes begin 3.
Early U.S. deployments delivered a 20% increase in throughput plus a 10% to 15% reduction in capital expenditures (CapEx) by tuning current assets 3.
For other capital-intensive industries, AI and digital twins can raise output from legacy infrastructure faster and cheaper than building new capacity from scratch.