PhysicsX

AI is changing engineering fast, but the reality of deploying it into real-world physical systems is still complex, operational, and deeply human. On June 4, we’re hosting the London edition of #BuiltToDeploy with Atomico: a breakfast bringing together engineers, technical operators, and builders working at the intersection of AI and the physical world. The conversation will focus on what happens after the demo: integrating AI into production workflows, earning trust from domain experts, and making systems usable under demanding constraints. If you’re building in this space — come join us! 🔗 Apply here: https://luma.com/yqlkii89

Last Thursday, we hosted #BuiltToDeploy in New York — a breakfast bringing together forward-deployed engineers, technical leads, and operators working at the edge of one of the hardest problems in AI: getting it to actually work in the real world. The conversation was refreshingly candid, focusing on the reality of deploying AI into complex physical systems: where trust breaks down, why deployments stall, how engineering roles are changing, and what it takes to build products for users who immediately know when something doesn’t make sense. A big thank you to our panelists — Sam Hiscox (Field CTO, PhysicsX), Anthony Mainero (Software Engineer, Nominal), and Davìd Markey (Air Dominance & Strike, Anduril Industries) — for the honesty, technical depth, and practical perspective they brought to the discussion. And thank you to everyone who joined us and pushed the conversation further. The quality of the room made the morning what it was. #BuiltToDeploy is coming to London next. More soon.

Last week, Jacomo Corbo joined the 55th St. Gallen Symposium — three sessions, one consistent signal: Europe knows it needs to transform. The conversations kept returning to the same tension. Europe has extraordinary industrial depth and a regulatory and policy environment that shapes global standards. What it lacks is a track record of building frontier technology at the pace and scale the moment demands. The risk isn't that Europe falls behind on ambition. It's that ambition without capability compounds the gap. Industrial AI is where this plays out most acutely. Not AI as a layer of automation applied to existing workflows, but engineering that is AI-native from first principles. That's what changes the structural ceiling: expanding the design space, deepening optimization, maximizing performance, compressing the development cycle. For critical industries like aerospace & defense, automotive, semiconductors, energy, and advanced manufacturing, this is the defining capability question of the next decade. At PhysicsX, we're working to answer it with an AI-native platform for engineering — built in Europe, used by leading industrials globally to solve their hardest challenges. #PhysicsAI #IndustrialAI #Innovation #DeepTech

The next industrial revolution will be AI-native. Who builds that infrastructure — and where — is one of the most consequential questions of this decade. This week, Jacomo Corbo is at the 55th St. Gallen Symposium, in three sessions on what Europe must get right to compete at the frontier. 🔗 Access the full program: https://lnkd.in/dKuwvF2V #PhysicsAI #DeepTech #Engineering #IndustrialAI #StGallenSymposium

Last Monday, our Founder & Chairman Robin Tuluie Ph.D. joined the UK Secretary of State for Business and Trade Peter Kyle, HMA Andrew Mitchell, and senior leaders from across European industry in Berlin to mark the UK's entry into the EU AI Champions Initiative. The conversation in the room reflected something we see in our own work: Europe's industrial base is not waiting. The proprietary data locked inside asset-heavy industries — aerospace, energy, advanced manufacturing — represents a structural advantage that is only valuable if you can act on it. That's precisely what physics AI is built to do. Europe has deep engineering expertise. The question is whether it can match that with the tempo required — moving like an F1 team rather than a legacy development program. We think it can. Congratulations to the EU AI Champions Initiative team on a significant milestone, and thank you to all who made the evening possible. #PhysicsAI #IndustrialAI #EUAICI General Catalyst Clara Koester Dr. Aline Vedder Kirstin MacLeod

The hardest part of AI in engineering isn't building the model. It's getting it into production. On May 14, we're bringing together forward-deployed engineers and technical leads from defense, aerospace, and industrial systems for a closed-door breakfast at the PhysicsX office in New York — hosted with Nominal and General Catalyst. The conversation will be honest and grounded: what deployment actually looks like in the field, from integrating AI into mission-critical workflows to navigating the friction that slows teams down. If you're building in this space, come talk with people who are doing the same. 🔗 Apply to join: https://luma.com/isev24qj

Your CFD solver isn’t the bottleneck. Running it a thousand times is. Fourier Neural Operators shift the balance. They learn the solution map once — then solve new PDE instances in milliseconds, not hours. Our team breaks down how FNOs actually work: the architecture, the Fourier intuition, where they genuinely excel, and just as importantly, where they break down. If you’re evaluating neural surrogates for engineering simulation, this is worth your time. 🔗 New deep dive by Gurpreet Singh Hora, Prakhar Kapoor, Ph.D, and Albert Matveev on the PhysicsX blog: https://lnkd.in/eDafepSm #PhysicsAI #SimulationEngineering #CFD #NeuralOperators 

Last week, we welcomed Dorothy Chou, Director of the Public Engagement Lab at Google DeepMind, to the PhysicsX London office for a fireside chat with Shamal Thakar on navigating public trust in technology innovation. A few threads that stayed with us: • Trust needs to be earned, not assumed. It's built through consistent behavior over time, especially when consistency is inconvenient. When gaps open up between what organizations claim and what they actually demonstrate, trust degrades quietly and quickly. • Good governance isn't primarily about constraints. It's about defining the outcome the technology is trying to achieve, and structuring the right incentives to get there. That means looking beyond penalties to tools like procurement requirements and advance market commitments, which are often more durable than reactive regulation, and still significantly underused. • The organizations most likely to hold trust over time aren't the ones with the sharpest messaging. They're the ones willing to say: here are our assumptions, here's who we consulted, here's what we still don't understand, and here's the recourse we're prepared to commit to. Real trust is visible when it's costly — when it means moving more slowly, opening up to scrutiny, or walking away from profitable but corrosive use cases. • Leadership sets the tone. The best leaders bring people with them, including the public, and particularly skeptics. Organizations that do this well don't just earn trust. They build the kind of legitimacy that holds up under pressure. Thank you, Dorothy — an honest, wide-ranging conversation that left the room with plenty to think about. More of this!

This Thursday at HANNOVER MESSE, our Field CTO Sam Hiscox takes the Center Stage alongside Markus W. Hacker (NVIDIA), Christian Piechnick (Wandelbots), and Maja Himmer (T-Systems International) to discuss what it actually takes to move industrial AI into production — from infrastructure to application layer. Sam will be speaking to how customers leverage the PhysicsX platform to embed physics AI directly into real-world engineering workflows, the operational practices that unlock impact at scale, and why sovereign European infrastructure like the Industrial AI Cloud matters. 📍 Hannover Messe, Hall 25, Center Stage 📅 Thursday, April 23 🕖 10:05-10:35 CEST 🔗 https://lnkd.in/d32sF2Fp

Large Physics Models are the new frontier — compressing industrial development cycles and redefining how engineering decisions get made. 👉 PhysicsX Co-Founder & CEO Jacomo Corbo shares his perspective in the new IEEE Spectrum feature: https://lnkd.in/eZGqVigX