Industrial automation's AI gap: why 80% of U.S. factories still run without robots

Eight in ten U.S. manufacturing facilities operate with zero automation. That figure, cited by Intrinsic Chief Technology Officer Brian Gerkey and reported by Manufacturing Dive, sits in sharp contrast to the volume of AI and robotics announcements filling trade media this year. The gap between vendor activity and factory-floor reality is the defining tension in industrial operations right now.

Belief is high, deployment is low

A Deloitte 2025 Smart Manufacturing and Operations Survey, cited by Manufacturing Dive, found that 92% of manufacturers believe smart manufacturing will be the primary driver of competitiveness over the next three years. Yet the same research shows that only a small share say AI is widely deployed across their operations today. Jeff Burnstein, president of the Association for Advancing Automation, put it plainly in comments to Manufacturing Dive: interest is high across the board, but execution is where things get difficult.

The barriers are structural. Cost, integration complexity, workforce skills gaps, and uncertainty about ROI all slow decision cycles. For a mid-sized plant manager, the leap from spreadsheet-driven scheduling to an AI-connected production line is not a software purchase, it is an operational transformation that touches every system on the floor.

Labor shortage is forcing the conversation

Honeywell CEO Vimal Kapur framed the urgency differently in a June appearance on CNBC's Mad Money. Rather than leading with efficiency gains, he pointed to demographics. Net workforce size is not going to be increasing, it is going to be decreasing over time, Kapur said, according to CNBC. Operators and technicians are already in short supply across Honeywell's customer base in sectors ranging from hospitals and data centers to LNG plants and semiconductor fabs.

That framing matters for procurement teams building business cases. Automation justified purely as a cost cut faces a different approval process than automation justified as a capacity expansion that offsets an unfillable headcount gap. Kapur also argued that Honeywell's existing sensor, controls, and software infrastructure already captures enormous volumes of operational data, and that AI now makes that data actionable in ways that were not previously possible, according to CNBC.

The timing is deliberate. Honeywell completed the spin-off of its Solstice Advanced Materials business last fall and is separating its aerospace unit, leaving a pure-play automation company that spans building controls, industrial software, and process automation, as CNBC reported. The restructuring concentrates investment and management attention on exactly the segment where AI opportunity is accelerating.

Vendors are meeting manufacturers closer to zero

For the 80% of facilities that haven't automated at all, the starting point matters enormously. A cluster of product launches in July 2026 reflects vendors trying to lower that entry point rather than sell only to already-sophisticated plants.

ABB extended its all-compatible drive platform to deliver localized motor control in severe operating environments, heat, dust, moisture, without requiring a separate protective enclosure, according to Automation International. That matters operationally because enclosure costs and installation complexity are common reasons drive deployments get deferred in brownfield facilities.

ARBOR Technology introduced a rugged edge AI platform targeting machine vision, intelligent transportation, and smart manufacturing applications, Automation International reported. Edge platforms that process vision data locally, without a cloud round-trip, reduce latency and data-transmission costs, two friction points that have slowed machine vision adoption in high-speed lines.

Avalue Technology took a different angle, expanding its integration with the Torizon platform to support embedded Linux lifecycle management and cybersecurity compliance for industrial systems, per Automation International. Cybersecurity compliance is increasingly a procurement gate for OT systems, particularly as manufacturers face sector-specific regulations and customer audits. Separately, the company also introduced fanless edge systems optimized for performance per watt, an important metric in facilities where thermal management and energy costs both constrain hardware choices.

On the sensing and vision side, Cambrian Robotics and SensoPart announced a partnership combining SensoPart's VISOR hardware with Cambrian's AI software to build a 3D handling system for industrial gripping and positioning, Automation International reported. The collaboration addresses one of the most persistent automation pain points: reliably picking and placing parts that vary in orientation or geometry, a task that has historically required expensive custom tooling or extensive robot programming.

What this means for your team

• Audit your automation baseline against the 80% figure: if your facilities are in the zero-automation cohort, prioritize identifying one high-labor, high-variance process as a contained pilot rather than a plant-wide program.
• Reframe budget justifications around capacity and workforce resilience, not headcount reduction, the Honeywell framing of automation as a revenue-generation tool is more likely to clear executive approval in a tight labor market.
• Evaluate new edge AI and drive platforms (ABB, ARBOR, Avalue) for brownfield compatibility specifically: ask vendors for evidence of deployment in facilities with legacy PLCs and without clean-room conditions.
• For vision and gripping applications, assess partnerships like Cambrian-SensoPart that bundle hardware and AI software, they can reduce integration risk compared to assembling components from separate vendors.