Flex Executive: Cobots, AMRs & Physical AI Move from Pilot to Scale

Why It Matters Now: The Automation Scaling Imperative

The global industrial automation market has crossed the $210 billion threshold in 2025 and is accelerating toward $226 billion in 2026, but a stubborn gap persists: the vast majority of manufacturers remain stuck in pilot mode. Despite 98% of manufacturers exploring some form of AI-driven automation, only 20% feel genuinely ready to deploy at production scale. The chasm between a compelling prototype and a product shipping reliably across dozens of facilities is where the real industrial drama unfolds — and where PLC-based control systems have become the silent backbone of every credible deployment.

Rodrigo DallOglio, President of Operational Excellence and Transformation at Flex — one of the world's largest contract manufacturers — recently offered a rare, ground-level perspective on what it actually takes to move collaborative robots (cobots), autonomous mobile robots (AMRs), and physical AI from isolated success stories into globally consistent production systems.

The Scale Problem Nobody Talks About

The industrial robotics sector is awash in venture funding and press coverage, yet the bottleneck is not algorithmic sophistication — it is manufacturing discipline. As DallOglio's experience at Flex reveals, physical AI systems face a "sim-to-real" gap that is measurable in cycle time, yield rates, and integration complexity. The fragmentation of legacy PLC architectures across different factories compounds the challenge exponentially.

"Working closely with Teradyne Robotics as an automation partner allows us to scale intelligent automation while supporting increasingly complex manufacturing environments for customers in electronics, industrial equipment, data center infrastructure, and other critical sectors," DallOglio stated, underscoring a fundamental truth: partnerships, not isolated technology bets, unlock scale.

Cobots or Humanoids? The Pragmatic Calculus for 2026

While humanoid robots dominate media attention and venture capital flows, the operational reality on factory floors tells a different story. At Automate 2025, Teradyne Robotics Group President Ujjwal Kumar delivered a keynote that drew a sharp line between technological spectacle and industrial need: cobots and AMRs are delivering measurable improvements in quality, accuracy, productivity, and profitability today — humanoids are not yet ready for widespread manufacturing deployment.

This pragmatism is reflected in Flex's partnership strategy. By standardizing on Universal Robots cobots and MiR AMRs — both Teradyne Robotics companies — Flex is building a repeatable automation architecture across its global footprint. The PLC integration layer becomes critical here: cobots and AMRs must communicate seamlessly with existing programmable logic controllers, MES systems, and factory networks to achieve true interoperability.

Physical AI: Why Manufacturing Is the Real Bottleneck

Physical AI — systems that combine vision, language, and action models to operate in the physical world — represents the next frontier of industrial automation. But the excitement around models like RT-2, PaLM-E, and NVIDIA's GR00T obscures a harder truth: the constraint is manufacturing capability, not model performance. The "sim-to-real" gap, the fragmentation of production ecosystems, and the organizational discipline required to iterate on hardware in near real-time are the genuine barriers to widespread adoption.

Flex's expanded partnership with Teradyne Robotics tackles this bottleneck directly. By manufacturing key robotics components while deploying finished cobots and AMRs across its own facilities, Flex creates a compressed learning environment. Production insights from the factory floor feed directly into design iteration, while PLC-level integration challenges are surfaced and resolved in operational context rather than in isolated lab settings.

The Flexibility-Standardization Paradox

One of the most underappreciated challenges DallOglio and his team navigate is the tension between flexibility and standardization. Contract manufacturers like Flex serve customers with radically different production requirements — from high-mix, low-volume electronics assembly to high-volume data center infrastructure. Each line demands a tailored automation configuration, yet achieving global operational consistency requires standardizing on a core set of technologies.

This is where PLC-based control architectures prove their enduring value. A well-designed PLC infrastructure provides the deterministic backbone upon which flexible automation layers — cobots, AMRs, vision systems, and eventually physical AI — can be reconfigured without destabilizing the entire production environment. Standardization happens at the control and data architecture level, while flexibility is expressed at the device and application layer.

What Comes Next: The Road to 2030

The trajectory is increasingly clear. Collaborative robot shipments are forecast to grow at 17.3% annually through 2030, while the AMR market is projected to more than double — from roughly $3.4 billion in 2026 to over $7 billion by 2031. The industrial automation ecosystem that supports these deployments — PLCs, industrial PCs, safety controllers, edge gateways, and integration services — will expand in lockstep.

Flex's journey from pilot programs to large-scale deployment offers a critical lesson for the broader industry: scaling automation is not primarily a technology problem. It is an operational discipline problem. The companies that bridge the gap between a working prototype and a globally consistent production system — that invest as deliberately in manufacturing capability as they do in AI capability — will define this next era of industrial innovation. PLC-based control architectures, far from being legacy systems, are the foundation upon which that future is being built.