SandboxAQ

A major milestone for American semiconductor manufacturing. Today, SandboxAQ announced a definitive agreement with the U.S. Department of Commerce for a $500 million CHIPS award to help tackle one of the most urgent challenges in American manufacturing: securing the critical materials and chemistries that underpin semiconductor production. Using Large Quantitative Models (LQMs), SandboxAQ will advance innovation across four material categories essential to the future of the industry: • PFAS-free process chemicals • Advanced catalysts • Rare earth-free magnets • Next-generation battery systems Across all four categories, the opportunity is the same: move critical materials from discovery to industrial deployment faster than ever before and strengthen the foundation of American semiconductor manufacturing. Read the full release here: sandboxaq.com/post/sandboxaq…

The semiconductor supply chain is only as resilient as its most vulnerable link. Today, more than 90% of the supply chain for certain critical magnetic materials used in semiconductor equipment is concentrated in foreign nations of concern. A disruption could put U.S. fabrication facilities at risk and expose just how much the industry depends on access to strategically important materials. Strengthening domestic materials innovation is about more than manufacturing capacity. It’s about building supply chain resilience and securing the capabilities needed to maintain U.S. semiconductor leadership. Watch the full conversation on YouTube to hear Shalini Sharma discuss the importance of rebuilding these strategic capabilities: youtube.com/watch?v=EZIHlk-E…

At Techonomy 26, SandboxAQ’s Luca Ferrara, who leads AQNav, broke down why GPS vulnerabilities are becoming an increasingly urgent challenge for commercial aviation, defense, financial markets and autonomous systems. In a new feature from Worth Magazine, Ferrara explains how GPS jamming and spoofing threaten critical infrastructure and why reliable, alternative positioning and navigation technologies are becoming essential. AQNav takes a different approach: using quantum sensing and AI-powered software to read the Earth’s magnetic field, providing positioning that is passive, globally available, jam-proof and spoof-proof. Read the full feature:worth.com/gps-the-1-6-billio…

SandboxAQ Essential Concepts – How Do You Design a Better Chip Layer? Every semiconductor chip is built layer by layer, using ultra-thin films of specialized materials. Getting those layers right starts with designing the molecules that make them possible. The challenge is finding a molecule that can travel as a gas, arrive intact on a surface and bond on contact. It also needs to remain stable long enough to reach the surface without breaking down or leaving contaminants behind. SandboxAQ’s large quantitative models (LQMs) use physics-based data to screen millions of potential molecules before they ever reach the lab. This could help accelerate the discovery of ultra-pure process chemicals, including new materials for extreme ultraviolet (EUV) lithography, where vapor-based deposition methods are becoming increasingly important. See how AI is helping researchers design the next generation of semiconductor materials in our infographic below 👇

What does it take to turn scientific breakthroughs into businesses that can scale? On the latest episode of the fAQ podcast, Alex Belsley, who leads Strategic Finance at SandboxAQ, shares how his journey from quantum photonics research to deep tech finance shaped the way he approaches building and scaling emerging technologies. The conversation explores the challenges of bringing complex technologies from the lab into the real world, the role of strategic capital allocation in accelerating innovation, and how to build sustainable businesses around scientific breakthroughs. Alex also shares how competitive chess has influenced his approach to decision-making and reflects on a family legacy rooted in early space exploration. Tune in to hear how science, strategy, and a long-term perspective come together in the business of deep tech: youtube.com/watch?v=DLP3KM3m…

That’s a wrap on AIDDD—and what an incredible week it was! One of our highlights was a full house for Mary Pitman’s keynote: "Building the Scientific Foundations of Agentic Drug Discovery." Mary’s central message struck a chord with the crowded room: workflow automation alone isn't enough. To make autonomous discovery truly impactful, we have to rethink the underlying science itself. As Mary said "Autonomous science requires methods designed not just to execute, but to evolve." She also shared what we’ve been building here at SandboxAQ to support this shift, including nonequilibrium chimeric switching, absolute binding free-energy calculation methods, and knowledge graphs designed to connect fragmented scientific data. The energy carried over to our poster sessions, where Amogh Sood and Rae Qi joined Mary Pitman to showcased three research areas: 🔸 Integrated AI and Physics-based methodologies 🔸 Accurate and Scalable free energy calculations 🔸 Scaling Binding Affinity Prediction A massive thank you to everyone who filled every seat for the keynote and stopped by our posters to exchange ideas. Want to continue exploring autonomous discovery workflows? Connect with Mary Pitman,Amogh Sood, and Rae Qi directly—we’d love to keep the conversation going!

SandboxAQ Essential Concepts – How Sensitive Is Sensitive? Imagine a camera lens that could capture all of Mount Everest in a single frame, then zoom in far enough to see a single human hair at the summit. That’s one way to think about the extraordinary dynamic range of quantum sensors: detecting incredibly faint signals without losing sight of much stronger ones in the same environment. Consider the magnetic fields produced by the human body. The heart’s magnetic field can be roughly a million times weaker than Earth’s, while signals from the brain can be smaller still. Detecting those signals against all the surrounding magnetic noise is the challenge. That level of sensitivity could enable new possibilities, from measuring biomagnetic signals for medical applications, to navigation where GPS isn’t available, to probing tiny magnetic fields inside living cells. See what that level of sensitivity can unlock in our infographic below 👇

SandboxAQ Essential Concepts – Where Quantum Sensing Goes Next Magnetic fields were only the beginning. A new generation of quantum sensors is expanding what we can measure with extraordinary precision, from acceleration and rotation to pressure, temperature, gravity, and more. And there isn’t just one way to build them. NV centers in diamond, trapped atoms, superconducting circuits, 2D material defects, and other emerging platforms are each pushing quantum sensing in new directions. That growing toolkit could unlock applications across navigation, medicine, exploration, and materials science, including navigation without GPS, next-generation medical imaging, sensing underground and underwater, and probing materials at the atomic scale. What’s especially exciting is that the field is still opening up. As new sensing platforms mature, so will the range of signals we can detect and the places we can put that precision to work. Explore where the technology could go next in our infographic below 👇

We are at the AI Driven Drug Development Summit in Boston this week. Come and find us! Here is where, and what to chat with us about: Tomorrow, 7:00 PM — The 19th Hole at Swingers. Drinks, mini golf, and friendly conversations about using AI to accelerate drug discovery. A few places left. Registration link here: the19thholewithsandboxaq.spl… Thursday, 2:30 PM — Dr. Mary Pitman, our Head of Simulation and Biologics, will be giving a talk on Building the Scientific Foundations of Agentic Drug Discovery. On Wednesday and Thursday, we will also be presenting three posters. Come and hear our scientists talk about their work: Dr. Mary Pitman— Integrated AI and physics-based methodologies for in-Silico Target Identification Dr. Amogh Sood — Accurate and Scalable Free Energy Calculations via Nonequilibrium Chimeric Switching (NEX) Dr. Rae Qi — Scaling Binding Affinity Prediction with Cofolding and Free Energy Calculations: Application to E3 Ligase Modulators All week during the conference, our scientists, along with Danielle Scheffey, Radha Patel, and Bill Fitzgerald will be around. Message us in Swapcard for a meeting!

Can you explain your research without a single slide? Science’s Dance Your Ph.D. contest is open for its 19th year, and SandboxAQ is sponsoring it again. The challenge is to turn your research into dance, with no PowerPoint and no jargon. Last year’s winner danced the piezoelectric effect, how certain materials produce an electric charge when squeezed or stretched. That’s the bar this year. You don’t have to be a current student. You can be just starting your Ph.D. or have finished it decades ago. There are four categories: physics, biology, chemistry, and social science. Each category winner receives $750, and the overall winner gets an additional $2,000. Submissions are due January 29, 2027, and we'll announce the winner by March 2027. At SandboxAQ, our work in AI and quantum technology depends on making complex science clear to people outside the lab. This contest rewards that skill, and we’re looking forward to the entries. Read the full details and enter here: science.org/content/article/… Learn more about the SBAQ residency program: youtube.com/watch?v=WuBHQwbB… and subscribe to our AQ Newsflash newsletter for more updates on science, research, and innovation: linkedin.com/newsletters/aq-… #DanceYourPhD #SciComm #STEM

This past Monday evening in Berlin, we announced that Apheris and SandboxAQ have launched the Virtual Cell Network. Our goal is to build the world's largest pharma-led virtual cell network. Late-stage failures are among the costliest outcomes in drug development. Virtual cell models predict how a compound changes a cell, helping teams spot problems earlier. To support decisions in drug programs, these models need data from human-relevant cell systems, measured at the relevant dose and duration, against efficacy and safety endpoints. Much of that data sits inside pharma companies, and no single company holds enough of it alone to train a model that performs reliably on new compounds and cells. In the Virtual Cell Network, members train AQCell on their proprietary data to broaden its applicability domain, then fine-tune it downstream to make it specific to their own programs. Raw data never leaves any company’s environment. SandboxAQ provides AQCell. Apheris operates the network and delivers the models through the Apheris platform, so pharma teams can customize them and integrate them into their drug programs. Apheris also runs federated networks for co-folding, ADMET, and antibody developability. Thank you to everyone who made this launch happen. At SandboxAQ: Nadian Harhen , Bill Fitzgerald, Andrew McLaughlin, Adam Lewis, Zane Beckwith, Nihit Pokhrel, Danielle Scheffey, Radha Patel, and Jennifer Stewart. At Apheris: Robin Röhm, Lewis Mervin, Julian Schoenauer, Iva Lukac, José-Tomás (JT) Prieto, and Max Nisslein. Founding members can join until the end of Q1 2027. To learn more or talk to our scientists, see apheris.com/networks/virtual… and sandboxaq.com/models/aqcell.

That's a wrap on The Battery Show North America 2026! We appreciated the conversations about how AI and physics-based modeling can help address practical challenges in battery materials discovery and performance prediction. We were also honored to receive the Software and Digital Innovation of the Year award - a recognition of our team's work bringing AQVolt and Large Quantitative Models to battery R&D. Thank you to everyone who visited our booth and shared their challenges and ideas. We look forward to building on these conversations. #TheBatteryShow #BatteryInnovation #AI #SandboxAQ

Just one week out, and we would like to meet you there. Bring your thoughts to this. Which of your computational methods would you let an agent call unsupervised? Dr. Mary Pitman, our Head of Simulation and Biologics, will be discussing this with you during her talk on Building the Scientific Foundations of Agentic Drug Discovery. She will talk about how we should build scientific methods for an agentic future that is robust without human in the loop tuning, flexible across diverse problems, and designed to operate natively within autonomous workflows. Mary’s talk will be on Thursday October 29 at 2:30 PM. We also have three posters accepted. Come and hear our scientists talk about their work: Dr. Mary Pitman— Integrated AI and physics-based methodologies for in-Silico Target Identification Dr. Amogh Sood — Accurate and Scalable Free Energy Calculations via Nonequilibrium Chimeric Switching (NEX) Dr. Rae Qi — Scaling Binding Affinity Prediction with Cofolding and Free Energy Calculations: Application to E3 Ligase Modulators Our scientists, along with Danielle Scheffey, Radha Patel, and Bill Fitzgerald, will also be at The 19th Hole on Tuesday the 27th — come and chat in person! Registration link here: the19thholewithsandboxaq.spl…

On Thursday, @switch_ai_ General Manager, Mohammed Aboul-Magd will discuss what's next for enterprise AI: moving beyond individual agents to a workforce where humans and agents collaborate, with shared context, governance, and flexibility. Continue the conversation at booth number 738 near Stage 4.

Join us in Berlin on Monday, October 19th, the evening before NVIDIA GTC Berlin! SandboxAQ and Apheris are hosting an evening for the life sciences and AI community and we’d love for you to be there. Together, we’ll share the story behind our new collaboration: how SandboxAQ’s physics-grounded Large Quantitative Models and Apheris’ federated networks can bring AI to the proprietary data that drug discovery depends on, without that data ever leaving the building. Our COO Andrew McLaughlin, Apheris CEO Robin Röhm, and key scientists behind the work will kick things off with a short scientific program. Then it’s time to connect with R&D leaders, drug-discovery AI researchers, founders, and investors over great food, drinks, and music. Location: zazou, Berlin. Französische Str. 15, 10117 Berlin, Germany Date and Time: Monday, October 19, from 7:00 PM CEST If you’re in Berlin or heading to NVIDIA GTC Berlin, you’re invited. Register here: luma.com/8mw452pu

Your AI workforce is growing. So is @switch_ai_ 🛡️ Switch Trust: Assess agent risk and apply guardrails. 📊 Switch ROI: Track agent costs, performance and value. More visibility. More control. Explore the expanded Switch AI: switchagents.ai/ Read the announcement: prnewswire.com/news-releases…

SandboxAQ Essential Concepts – What Is AQVolt, and Why Does Battery Discovery Need It? How can we discover better battery materials faster? AQVolt is a physics-grounded AI and simulation platform designed to help predict material properties, identify promising battery compositions, and optimize performance. By combining physics-based methods with AI-driven insights, AQVolt can screen large numbers of candidates before they reach the lab. The result? Researchers can focus lab time on stronger candidates, accelerating the cycle from prediction and discovery to validation and optimization. AQVolt doesn’t replace the lab. It helps make lab time more targeted and productive. Explore the diagram above to see how 👇

Battery Technology recapped the 2026 Battery Awards, held at The Battery Show North America, which celebrated the innovations and companies driving the battery industry forward. Among this year’s honorees was SandboxAQ’s AQVolt, named 2026 Battery Software and Digital Innovation of the Year. We’re proud to see our work recognized alongside the innovations shaping the future of battery technology. As the industry works to improve battery performance, safety, and longevity, AQVolt's capabilities can help inform smarter decisions across the battery lifecycle. Read the full article from Battery Technology for a look at this year’s award winners and highlights from the event: batterytechonline.com/batter…

SandboxAQ Essential Concepts – How Can We Simulate and Predict Battery Performance? How can we better predict how a battery will perform over time? Battery performance depends on processes happening across multiple scales, from atoms and material properties to interfaces and full cells. Connecting these scales can help us understand everything from ion migration and surface reactions to cycle life, temperature evolution, and operating windows. The goal? Combine simulation, testing, and learning to predict battery behavior, identify opportunities for improvement, and accelerate the development of better-performing cells. Explore the diagram above to see how modeling across scales can help us better understand and predict battery performance 👇

🎉 What a night in Detroit! Last night AQVolt took home the 2026 Battery Software and Digital Innovation of the Year award at The Battery Show North America — and the team is still at Booth 4503 today! Ang Xiao, Aatif Misbah, Alexandre Fleury, Chuhong Wang, QJ Li, and Scott Healey will be providing demos all day at the booth. Come see our award winning solution for yourself! A few more ways to meet the team this week: Wednesday evening. We are hosting a happy hour today, 5:00–7:00 PM EDT at 607 Shelby St. Register here if you have not yet: sandboxaqhappyhour.splashtha… Thursday afternoon: Dr. Ang Xiao, Sandbox’s Technical Lead of AI for Materials Science, will be giving a talk on “AI Meets Battery Science: Accelerating Materials Discovery with Large Quantitative Models” on Thursday at 2:30pm EDT. Additional details here. Add this to your agenda if you have not! Congratulations to our team who built AQVolt. Come meet them in person!

Being directionally right about a cell is a long way from being a model you can fully trust. Figuring out where foundation models belong in a discovery workflow means working through the space between those two points. That is the subject of "Where Foundation and Virtual Cell Models Are Earning Their Place in Discovery" at Data Driven Pharma West, moderated by Jordan Crivelli-Decker, Head of Machine Learning & Informatics at SandboxAQ. Hajo Schiewe will also be around to answer your questions! If you are attending DataDrivenPharma, join the conversation on Thursday, October 15, 3:45 pm!