Validation and the engineering imperative
they may arrive faster than traditional infrastructure planning cycles expect. For infrastructure leaders, this is not a question of if, but of when and whether today’s investment decisions reflect that trajectory.⁶ That makes complacency risky, particularly for those planning large-scale projects today. The assumption that infrastructure will continue to be governed by the same material constraints that shaped the past century is becoming less secure and assets designed around today’s process assumptions may prove obsolete sooner than expected. The organisations that grasp this early will not just operate better assets; they will be better placed to shape the next cost curves of the physical economy. The boundary of what’s possible We are moving from an era in which AI augments decision-making to one in which it begins to reshape the physical systems those decisions act upon. The boundary between the digital and the physical is shifting — and with it the boundary of what is possible. In a century defined by pressure on energy, resources and resilience, that shift could prove as consequential as any digital revolution. AI is no longer just a software layer sitting on top of infrastructure. It is becoming part of the design, control and operation of the physical world itself.
None of this removes the hard part. Moving from computational discovery to physical deployment introduces a different order of complexity. Engineering is and should remain risk-aware. It is shaped by operational realities, safety standards and the long test of durability in the field. It also operates within liability frameworks that require designs to be understood, not just produced and that will not change simply because the tool generating them is more powerful. The core skill of engineering will not disappear. But it will change. The work will shift, in part, from solving within relatively fixed constraints to exercising judgment over a much wider design space — validating and taking responsibility for solutions generated computationally. That in turn will require new capabilities. Asset owners and engineering firms will need stronger validation frameworks, updated testing protocols and the ability to design and validate new industrial process architectures — alongside closer engagement with regulators as standards evolve for computationally generated materials and processes. There is also a question of timing. Because these advances scale first through computation, simulation and design tools before they scale through construction,
Physical AI, by the numbers 10–20 years
10x–100x faster Early-stage discovery workflows
Millions Billions
Traditional materials discovery
Materials identified computationally
Deployed and multiple billion-dollar valuations: capital is moving quickly into AI-for-materials and discovery platforms⁷
Selected references
1. Karniadakis, G. E. et al. (2021). “Physics-informed machine learning.” Nature Reviews Physics, 3(6), 422–440. https://www.nature.com/articles/s42254-021-00314-5 2. Huang, J. (2025). GTC 2025 Keynote Address. NVIDIA GTC San Jose. Transcript retrieved from Rev.com https://www.rev.com/transcripts/gtc-keynote-with-nvidia- ceo-jensen-huang 3. Hassabis, D. (2024). Accelerating Scientific Discovery with AI. Nobel Prize Lecture, December 8, 2024. Nobel Prize Committee. https://www.nobelprize.org/prizes/ chemistry/2024/hassabis/lecture/ 4. Merchant, A. et al. (2023). “Scaling deep learning for materials discovery.” Nature, 624, 80–85. https://www.nature.com/articles/s41586-023-06735-9
5. Aspuru-Guzik, A. et al. (2018). “The role of AI in chemical discovery.” ACS Central Science, 4(2), 144–152. https://pubs.acs.org/doi/10.1021/acscentsci.8b00338 6. International Energy Agency. World Energy Investment 2025. https://www.iea.org/ reports/world-energy-investment-2025 7. McKinsey & Company. The State of AI: Global Survey 2025.; Tech Funding News (2025), “Ex-OpenAI execs raise $200M at $1B valuation for AI materials science startup backed by a16z.” https://www.mckinsey.com/capabilities/quantumblack/our-insights/the-state-of-ai https://techfundingnews.com/ex-openai-execs-raise-200m-at-1b-valuation-for-ai- materials-science-startup-backed-by-a16z/
6 | GHD | Nexus Magazine
Made with FlippingBook - professional solution for displaying marketing and sales documents online