Researchers from Lawrence Livermore National Laboratory (LLNL) and Meta have collaborated to create OPoly26, the world’s largest open dataset of atomistic polymer chemistry. This dataset comprises millions of quantum-accurate simulations designed to enhance AI modeling of everyday materials like plastics, films, and batteries. Polymers play a vital role in numerous products, from clothing to electronics, and improvements in polymer science can lead to efficient recycling and environmentally friendly materials.

OPoly26 includes over 6 million density functional theory (DFT) calculations, making it significantly larger and more diverse than existing datasets. This wealth of information empowers AI to recognize patterns in polymer behavior, facilitating rapid advancements in materials design while addressing gaps in polymer data. The dataset is expected to aid the development of safer and more sustainable materials, particularly in light of challenges posed by harmful substances like PFAS.

This partnership leverages LLNL's advanced computational capabilities, utilizing one of the world’s fastest supercomputers to significantly expedite the simulation process. Meta’s robust computational resources contributed to the extensive DFT simulations and machine-learned interatomic potentials (MLIPs) that improve model predictions across polymer and small-molecule chemistry. The combined efforts have shown that incorporating polymer data enhances model accuracy, underscoring the importance of realistic, complex training data.

The OPoly26 dataset is intended as a public resource, democratizing access to data and fostering innovation across research sectors. Researchers are committed to ensuring the dataset’s availability under an open license, maximizing its utility for scientists in academia, industry, and government. This initiative exemplifies how open science and AI can drive significant progress in materials research.

The Jabil Playbook: How an Industrial AI Pacesetter Engineers Extreme Agility
BY COLIN MASSON

Overview

At the 30th Annual ARC Industry Leadership Forum held in Orlando from February 9-12, the spotlight was on Industrial AI, focusing on how end users can derive business value from it. Leading organizations are treating AI as an essential operational capability rather than a mere IT experiment. A notable example is PepsiCo, which effectively utilizes agentic AI and extensive digital twins to handle high-volume operations and logistics.

However, many manufacturing scenarios differ significantly. For instance, contract manufacturers like Jabil face high-mix, low-volume production with rapid changeovers. At the Forum, Chase Christensen, VP and CIO of Jabil, explained how the company applies AI to meet its unique challenges. Jabil collaborates with brands across healthcare, automotive, 5G, and data centers, providing services from design to full-scale production.

What sets Jabil apart from companies like PepsiCo is its tailored use of AI, reflecting the specific demands of its sector. Both organizations emphasize execution over experimentation. Many companies remain in "Pilot Purgatory," where isolated projects fail to scale. Chase highlighted Jabil's strategic approach in his talk "From Charter to Execution." Through a dedicated AI Steering Committee, Jabil ensures that AI initiatives align with business goals and deliver tangible operational and customer benefits. This governance model emphasizes maximizing value, showcasing how AI can drive efficiency and create competitive advantages tailored to diverse manufacturing environments.

Refractory alloys, crucial for advanced defense systems like hypersonic aircraft and nuclear-powered submarines, are engineered materials that maintain strength even in extreme heat. These alloys combine metals such as tungsten, niobium, and molybdenum, known for their high melting points. Their resilient atomic structure resists deformation under intense conditions, making them ideal for components facing severe heat, stress, and radiation.

Traditionally, the design of refractory alloys predates modern technologies like 3D printing and artificial intelligence (AI). While additive manufacturing allows for the creation of complex parts with minimal waste, many existing refractory alloys struggle with cracking and internal defects during the 3D-printing process. To tackle this, a collaborative effort by researchers from Arizona State University and UNSW Sydney employs reinforcement learning, an AI approach typically used in gaming.

This AI systematically explores numerous alloy combinations, evaluating them against criteria like strength and heat resistance, and identifying compositions that are conducive to successful 3D printing. For example, NASA’s GRX-810 alloy, crafted using AI and 3D-printed, exhibits remarkable durability at high temperatures.

The advantages of this research are significant for defense and aerospace sectors. By accelerating materials development, organizations can enhance the performance of next-generation engines and heat-protective systems while drastically reducing material waste from traditional machining, potentially lowering waste to almost zero.

Despite challenges such as limited data and costly materials, the AI model is being developed to improve alloy design, with practical tests planned to validate predictions. This approach represents a significant leap towards faster innovation and adaptability in defense technologies, signaling a new era in materials science.

OMP Unveils Decision-Centric Planning to Enhance Supply Chain Decision Speed

In a groundbreaking shift, OMP, a pioneer in AI-powered supply chain planning, has introduced Unison Decision-Centric Planning, an innovative approach enabling organizations to transition from reactive, process-driven strategies to proactive, event-centric decision-making.

Leveraging OMP's renowned Unison Planning™ platform, this new strategy integrates advanced AI, autonomous agents, real-time scenario modeling, and human oversight to significantly enhance decision-making speed. Organizations can now better anticipate disruptions, evaluate trade-offs, and act decisively in an increasingly unpredictable supply chain landscape.

By moving from conventional planning cycles, which often lag behind market changes, to a dynamic, decision-first methodology, Unison Decision-Centric Planning continuously detects shifts, identifies critical scenarios, and assesses business impacts. This alignment of AI capabilities with human expertise helps organizations shift from merely reacting to efficiently optimizing value.

David Kochanek, Supply Chain Solution Manager at Evonik Oxeno, highlights the practical benefits of this approach. By collaborating with OMP, Evonik transitioned to an always-on, scenario-based decision system that enabled faster responses to disruptions, boosting overall performance. "Unison Decision-Centric Planning has fostered trust in our decision-making process, allowing us to enhance agility and company performance,” he stated.

The transformative potential of this AI-driven approach lies in its event-driven agents, which continuously evaluate opportunities and risks. Organizations can now model hundreds of scenarios, leading to improved service levels, cost efficiency, sustainability, and quicker decision cycles.

Uncover the impact of decision-centric planning on your supply chain by exploring OMP's resources. OMP's Unison Planning™ empowers clients across various sectors, enhancing their ability to navigate complex planning challenges effectively.