Revisiting Our 2026 Industrial Transformation Predictions: What Has Held Up So Far?

In January, we published ten predictions for industrial transformation in 2026. With the end of the year approaching, we have revisited four to see how they compare with developments so far. The picture is mixed: some are taking shape, others have unfolded differently than how we expected, and a few still lack the evidence to call it either way. Two months before the end of 2026, we’ve found that:

• AI agents are arriving, though role-based design is still catching up.
  We predicted that “at least half” of asset management and AI analytics vendors would launch persona-based AI agents. Agent launches (of products, not just previews) have been frequent: UptimeAI launched task-based agents in March and Augmentir released agent-building tools in September. Meanwhile, Seeq introduced Agent Q – a natural-language AI analyst – alongside tools for building custom agents, while IFS Ultimo introduced planner, technician and safety manager digital workers in May. Of these offerings, IFS Ultimo’s map most directly to job roles, while the others mostly target tasks or leave customers to build their own workflow.
• AI-powered robots are in training, not quite yet being used in production.
  We predicted that “robotics will have its ChatGPT moment in manufacturing”. And we have seen some moves in this direction. In late June, BMW announced that Figure 03 humanoid robots will take on logistics sequencing at its Spartanburg plant. In September, Boston Dynamics reported that its Atlas humanoid robot is “now in training” at its application centre at Hyundai’s Georgia plant, where it is learning tasks such as sequencing parts. It also announced that a new building that will expand the centre by about tenfold is due in 2027. Training centres and planned deployments show intent and investment, but current evidence of foundation-model robots doing production work in variable manufacturing settings is thin.
• Microsoft and AWS supplied platforms, while partners shipped the sector-specific suites.
  We predicted that a hyperscaler would launch “a fully integrated, sector-specific industrial AI suite”. At HANNOVER MESSE in April, Microsoft positioned Azure IoT Operations and Microsoft Fabric Real-Time Intelligence as the data foundations for industrial agents rather than launching a new suite. Meanwhile, Schneider Electric unveiled agentic manufacturing capabilities built on Azure AI. In the same month, Infor announced manufacturing agents built on AWS and sold through its own Velocity Suite, and QAD showcased ChampionAI agents built on Amazon Bedrock. AWS’s October launch, the open-source Physical AI Toolchain, gives developers a route from synthetic data to edge deployment for intelligent machines. For now, the hyperscalers supply the data, model and agent services, and the sector-specific suites come from their partners.
• Plants are halting, but not because of prompt injection.
  We predicted that “five Fortune 500 manufacturing plants will halt operations after prompt injection attacks” on agentic AI workflows, but up to this point in 2026 such cases have not been publicly reported. Some plants did halt in 2026 due to cybersecurity breaches, but the disclosed causes were not prompt injection. Fairlife, owned by Coca-Cola, paused US production after a ransomware event in July, and Stryker restored most manufacturing operations roughly two weeks after a March cyber attack that its SEC filing linked to a malicious file. Cyber attacks are rising, but public disclosures point to established vectors rather than attacks on AI workflows; a weeks-long cybersecurity-related stoppage like the one Jaguar Land Rover faced in late 2025 remains a real threat.

The direction is clearer than the timing

Specialized AI agents are becoming more common, and hyperscalers are reaching manufacturers through their partners. Our predictions on robotics and attacks on agentic AI workflows remain ahead of the available evidence. For industrial buyers, these developments are useful signals for planning, even where the timing remains uncertain.

For more information about our 2026 predictions, read Market Insight: 10 Predictions For Industrial Transformation In 2026 And Beyond. For a look at what the future holds, register for our 2027 industrial transformation predictions webinar.