AVEVA's Garg: Industrial AI adoption up 78% in two years

AVEVA has been applying AI to industrial applications for more than 20 years, mostly with narrower tools such as a proprietary anomaly detection model. What changed recently, in Garg's telling, is the arrival of general-purpose foundation models, an explosion in physical AI that accelerates robotics and autonomous systems, and agentic AI that automates at the software level. The result, she says, is not gradual growth but a step-function shift in adoption — one study points to almost a 78% increase over just two years.

The safety stakes differ from purely digital AI because industrial systems interact with physical equipment in hazardous environments, where an unexpected decision can affect safety, reliability, and critical infrastructure. Garg notes that newer models are by design harder to explain and can change behavior as they learn, which is why AVEVA leans toward guardrails — bands of operation, limits on where automation is allowed — and toward a human supervisor role rather than a human operator role. She is candid that this balance is still being worked out.

The company is also pursuing adjacent problems where AI and sustainability intersect. With Idaho National Laboratory, AVEVA has worked on an AI grid resilience project aimed at helping operators manage a grid with more intermittent renewables. Garg separately chairs the IEEE P7100 working group, which is trying to establish a single methodology for measuring AI's environmental footprint across electricity and energy consumption, resource usage, water consumption, and carbon, in part so reporting and oversight can have a common basis.