AI
March 31, 2026
Industrial AI: Where the Hard Work Begins
# AI
# Automation
# Automation & Digitalization
# Digital Twin
# Digital Twin & Simulation
# Ecosystem & Collaboration
# Factory of the Future
# Industrial AI
# Innovation
The tools are ready. Are the organizations?
Industry Signals
Industrial AI is the application of artificial intelligence technologies to optimize industrial processes through enhanced automation, real-time data, and predictive analytics. Since we covered how industrial leaders were rewiring for AI back in October, application has become increasingly beneficial â and challenging â for organizations. We are seeing that the technology itself is not the primary obstacle. Instead, the problem is organizational: integrating AI into existing workflows, governing autonomous systems responsibly, and building the operational discipline to move from pilot to production at scale.
In this edition of Industry Signals, we explore:
- NVIDIAâs overview of what industrial AI is, what it enables, and how companies are putting it to work;
- Siemensâ own activity on the industrial AI front, from CES 2026 to the recent RXD Summit in Beijing;
- Emerjâs analysis of why industrial AI projects stall before reaching production; and
- Deloitteâs look at how AI is accelerating physical product innovation across industries.
ďťż
NVIDIA on What Industrial AI Is and Why It Matters
ďťż
NVIDIAâs glossary entry on industrial AI offers a useful baseline for anyone who wants an explanation of the category before diving into use cases. It defines industrial AI as the application of physical AI and other artificial intelligence technologies to optimize industrial processes using real-time data, predictive analytics, and advanced automation to reduce human intervention, cut downtime, and improve decision-making across the industrial value chain.
Key ideas:
- Digital twins (virtual representations of physical systems) are central to industrial AI in practice. They allow organizations to simulate and verify the performance of AI models and applications before deployment in real facilities, reducing the cost and risk of physical trial-and-error.
- AI-powered automation enables real-time monitoring and predictive maintenance, catching equipment issues before they cause unplanned downtime and allowing continuous optimization throughout the lifecycle of industrial assets.
- Quality control is a primary use case: by continuously monitoring production processes and identifying defects as they occur, AI keeps products meeting standards without adding inspection overhead.
- Sustainability is a big benefit. By optimizing energy and resource consumption, industrial AI helps companies reduce environmental footprint while meeting regulatory and operational requirements.
Siemens Electronics Factory in Erlangen, Germany is a prime example of these kinds of technologies in place, with a reported 30% increase in productivity at the facility. You can watch a video about it here.
ďťż
Siemens on Building the Industrial AI Operating System
ďťż
The first months of 2026 have been active for Siemens on the industrial AI front in assembling the infrastructure layer with partners. Two major announcements, at CES in January and at the RXD Summit in Beijing in March, outline what we are building and with whom.
Key ideas:
- At CES 2026, Siemens expanded our NVIDIA partnership to build the Industrial AI Operating System, a technology stack applying AI across the full industrial lifecycle from design and engineering through manufacturing, operations, and supply chain. In this partnership, we also aim to build the world's first fully AI-driven, adaptive manufacturing sites, starting with the Siemens Electronics Factory in Erlangen.
- The headline product launch at CES was Digital Twin Composer, available on the Siemens Xcelerator Marketplace from mid-2026. PepsiCo is already using it to simulate upgrades across U.S. manufacturing and warehouse facilities, reporting a 20% increase in throughput on initial deployment and 10â15% reductions in capital expenditure through virtual validation.
- Siemens announced nine new AI-powered copilots spanning Teamcenter, Polarion, and Opcenter, covering product data navigation, compliance automation, and manufacturing operations.
- At the RXD Summit in Beijing, 26 new Siemens products were announced, along with the expansion of our partnership with Alibaba to bring computer-aided engineering to customers in China as a cloud service, alongside new edge, automation, and control technologies designed to execute AI-driven decisions on the shop floor.
If you want to go deeper on how Siemens is putting industrial AI to work, the Industrial AI Revolution keynote from CES and the Industrial AI is Scaling Now keynote from RXD are both worth your time.
ďťż
Emerj and IFS on Why Industrial AI Projects Stall Before Production
ďťż
Emerj and IFSâ recent piece on why industrial AI projects stall before production draws from MIT Sloan and NIST research that shows that AI-adopting manufacturers experience a huge initial productivity decline and avoidable maintenance-related costs driven by poor fit between new systems and existing workflows. These are problems AI is meant to address (not create), and precisely where it under-delivers when integration fails. Emerj spoke with Kriti Sharma, CEO of IFS Nexus Black, and Somya Kapoor, CEO of IFS Loops, about how asset-intensive industries can address this.
Key ideas:
- Sharma makes a case for high-velocity deployment, which is grounded in operational proximity. Her team embeds engineers directly on the factory floor to find the data that actually governs the process â SCADA signals, sensor streams, P&ID diagrams â rather than building against a laboratory abstraction. The goal is compressing deployment to production-grade results in roughly three weeks, through tight scope and direct contact with real operational constraints.
- Kapoor focuses on preserving institutional knowledge before it walks out the door with retiring technicians. Her approach starts with back-end tasks that follow clear logic, like inventory replenishment, warranty verification, supplier order management, where digital workers can learn company-specific rules before gradually taking on more autonomy.
- Governance is the third requirement. Kapoor's framework assigns every agent a clear identity and audit trail, uses supervisor agents as a second reasoning layer to catch errors, and routes edge cases to human experts via trigger notifications. Her point is direct: building agents is now the easy part; governing them at scale is where most organizations fall short.
âYou hear about the 5% of AI projects that succeed and the 95% that fail. The difference isnât the model. Itâs whether the work is agentic from the start. When you design the system to take action rather than just analyze, you avoid ending up with pilots that never go anywhere.â â Somya Kapoor, CEO of IFS Loops
ďťż
Deloitte on AI and Physical Product Innovation
ďťż
Deloitte recently made the case that often gets underplayed in AI coverage: generative AI is changing how organizations manage data and processes, and also how they design, prototype, and manufacture physical products. This article draws on Deloitteâs 2025 Tech Value Survey and sector-specific research across life sciences, consumer goods, energy, and industrials.
Key ideas:
- Generative AIâs three core capabilities (simulation, prediction, and optimization) apply to physical product development in ways that are beginning to compress timelines meaningfully. In pharmaceuticals, Deloitte cites evidence of prototype development cycles being cut by up to 70%, with implications for both cost and time to patient.
- Product development is evolving into what the article describes as a triad of collaboration: engineers, operators, and AI systems jointly shaping manufacturing strategy in real time. This changes not just how products are made but what capital investment decisions get made and when.
- The organizations seeing results are investing in continuous learning, and treating rapid experimentation as a standard operating mode.
ďťż
Other Resources
For ongoing coverage of industrial AI in practice, The Industrial AI Podcast is worth a follow. It features engineers, scientists, vendors, and startups across robotics, automotive, process, and automation industries with a focus on making the topic accessible to industrial practitioners.
ďťż
Looking Ahead
Industrial AI rewards organizations willing to get their hands dirty by embedding engineers in operational environments, starting with the workflows that matter most, and building in governance from the beginning. Technology is no longer the hard part. Knowing your process well enough to improve it with AI still is, and that knowledge lives on the floor in the people who run the machines and manage the exceptions. Organizations who succeed with industrial AI are the ones treating that expertise as the starting point.
ďťż
Thatâs a wrap for this edition of Industry Signals. Have a report, use case, or event youâd like to see featured in an upcoming issue? Send a note via PM. Weâre always looking to spotlight whatâs shaping the future of industry, and recommendations from the Xcelerator Community are especially valuable. Your insights and experiences continually shape Industry Signals.
ďťż
Comments (0)
Popular
Dive in
Related
Blog
Industry Signals â October 7, 2025: How Industrial Leaders Are Rewiring for AI
By Industry Signals â˘Â Oct 7th, 2025 ⢠Views 219
Blog
Industry Signals - July 22, 2025: The Future of Work: Taking a Human-Centered Approach to Industrial Intelligence
By Industry Signals â˘Â Jul 22nd, 2025 ⢠Views 306
Blog
Industry Signals - July 22, 2025: The Future of Work: Taking a Human-Centered Approach to Industrial Intelligence
By Industry Signals â˘Â Jul 22nd, 2025 ⢠Views 306
Blog
Industry Signals â October 7, 2025: How Industrial Leaders Are Rewiring for AI
By Industry Signals â˘Â Oct 7th, 2025 ⢠Views 219