Every hardware design contains decisions made before the team fully understands their physical consequences. An architecture is chosen. Components are placed. Materials are selected. Engineers bring experience and judgment to those choices, but the detailed physics often arrives later, when changing course is harder and more expensive.
That order of work limits what engineers can build. When every answer requires substantial preparation, compute and specialist time, teams have to decide which questions are worth asking. Promising ideas can be set aside because there is no practical way to investigate them within the schedule.
We built Vinci to change that. Today, we announced a $250 million Series B at a $1.5 billion valuation. The investment gives us the resources to accelerate a goal that reaches well beyond any single analysis: shortening the distance between what engineers can imagine and what they can build.
What waiting for physics costs
The cost of slow analysis extends into decisions made long before a simulation begins. Teams simplify designs so they can evaluate them. They add margin where uncertainty remains. They carry assumptions forward because testing every alternative would take too long.
Consider a change in component placement. It may improve one aspect of a design while changing how heat moves through the board. Differences in temperature and material expansion can also affect warpage. The engineer needs to understand those interactions while the placement is still a choice. If that understanding arrives after other decisions depend on it, even a small change can require substantial rework.
This becomes more consequential as engineers design chips, packages, boards and systems together. A decision about power affects cooling requirements. A packaging choice changes the physical constraints elsewhere. The number of interactions grows, while the time available to investigate them remains limited.
AI adds urgency to this problem. As engineering workflows become increasingly automated and agentic, they will be able to propose and act on more design choices. Each of those choices still has physical consequences. Generating possibilities is useful only if engineers can determine which ones will work.
Physics inside the design process
Continuous Physics Reasoning makes trustworthy physical understanding available throughout design. Engineers can investigate a change, understand its consequences and use that understanding to decide what to do next, while the design is still moving.
Making this practical requires changes to the entire computational process. Vinci automates design understanding and preparation, combining agentic orchestration with our Foundation Model for Physics and GPU-native physics kernels. The platform delivers deterministic, solver-accurate results at manufacturing resolution and works across new designs without customer-specific training or fine-tuning.
Those properties matter together. Engineers need enough physical detail to evaluate the design they intend to manufacture. They need repeatable answers they can trust. And they need to investigate a new design without first building a model trained specifically for it.
We chose semiconductors as our first proving ground because their physical demands leave little room for compromise. Vinci is already running on production engineering programs, with commercial capabilities across thermal, thermo-mechanical and convective fluid behavior. Manufacturing-scale analyses that once required hours or days can be completed in minutes.
For an engineering team, that changes how much it can learn before committing. More alternatives become practical to evaluate. A question that would once have been deferred can inform the next design decision. Engineers can pursue ideas that would otherwise have been too costly or time-consuming to investigate.
What we are building toward
Continuous Physics Reasoning gives us the foundation for a broader ambition. Once physical understanding can keep pace with design, we can work toward helping engineers determine what should change to improve a product, and eventually toward creating designs from human intent.
That is a demanding engineering problem. A useful recommendation has to account for how changing one part of a system affects the rest. It must work within constraints imposed by materials, manufacturing and the other disciplines involved in building a product. Extending physical understanding into that kind of engineering intelligence is central to our long-term work.
This financing will accelerate progress across both physics and engineering. We will expand physical coverage beyond the capabilities available today, develop more capable agentic engineering workflows and deepen integrations with the systems our customers use. We will also invest in the people and computing infrastructure needed to bring Vinci into more production programs.
The need extends well beyond semiconductors. Engineers developing vehicles, aircraft and satellites face different physical problems, but they share the need to understand the consequences of decisions while those decisions can still change. Our ambition is to make that understanding part of the infrastructure on which hardware engineering runs.
I am grateful to our customers, investors and team for helping us pursue this work. Our customers are putting Vinci to use on real engineering programs, where the value of an answer depends on whether it can inform the work in front of them.
That is the standard we intend to keep building toward. Engineers should be able to pursue an idea because they can investigate it, understand its constraints and improve it. Giving them that ability throughout design can expand the range of products they are able to create. That is what this next chapter of Vinci is for.
