Engineering AI Startup Vinci Raises $250M at $1.5B Valuation

Physics simulation has long been a bottleneck in hardware engineering. Traditional solvers like Ansys and Siemens Simcenter — software that solves physics equations directly to calculate how a design will actually behave — are accurate but can take hours or days to validate a single design. AI alternatives that emerged in recent years, so-called “surrogate models,” are fast but have to be trained on data from a specific product family, so they need retraining whenever the design changes even slightly. Vinci is raising money on a third path: a “zero-shot” foundation model for physics.

Engineering AI Startup Vinci Raises $250M at $1.5B Valuation

Vinci said it has raised a $250 million Series B, co-led by Advent, Temasek, and Xora, with AMD Ventures, Eclipse, Khosla Ventures, Madrona and others participating. The round values the company at $1.5 billion.

Reasoning About Physics While the Design Is Still Being Made

Vinci’s core technology is what it calls Continuous Physics Reasoning. Rather than validating a finished design with a separate simulation step, it calculates how a product will actually behave in near real time while the design is still being created. The architecture combines automated design understanding and preparation, agentic orchestration, a “Foundation Model for Physics,” and GPU-native physics kernels — and the company says it can be applied directly to new designs without per-design or per-geometry retraining. It says the platform analyzes manufacturing-scale designs ranging from hundreds of millions to more than 15 billion degrees of freedom in minutes rather than hours or days. That “zero-shot” claim, however, is the company’s own description — it hasn’t been independently verified by a third party.

Vinci first commercialized the technology in semiconductor thermal, thermo-mechanical, and convective-fluid simulation — one of the most physically demanding engineering environments there is. The new funding will be used to expand into other physical phenomena and engineering disciplines, including memory, computing, vehicles, aircraft, and satellites. “Designing advanced systems requires engineers to understand how thermal, physical and electrical behavior interact across the chip, package and board,” said Brian Amick, SVP of technology and engineering at AMD. “By bringing fast, accurate simulation into the design process, teams can evaluate tradeoffs earlier, iterate more quickly and make better architectural decisions while there is still time to act.” AMD Ventures is a strategic investor in this round, and Amick’s comment appears to be grounded in AMD’s engineering organization actually testing Vinci — but it hasn’t been disclosed whether AMD has signed on as a paying customer.

A Carbon Alum CEO and a Robotics Co-founder

Vinci was co-founded in late 2023 in Palo Alto by CEO Hardik Kabaria and co-founder Sarah Osentoski. Kabaria holds a Stanford PhD in computational geometry and meshing and joined 3D-printing company Carbon as an early employee, rising to VP of Software there. Osentoski is a robotics and machine-learning researcher who previously served as SVP of Engineering at Iron Ox and COO of Mayfield Robotics.

Vinci first disclosed its funding when it exited stealth last December, revealing a combined $46 million in seed and Series A funding. The $36 million Series A was led by Xora, Temasek’s deep-tech investing platform, with Khosla Ventures and Eclipse participating; no valuation was disclosed at the time. With this Series B, Vinci’s total disclosed funding has grown to roughly $296 million.

Competitors: A Race Among Physics Foundation Models — With Overlapping Backers

Vinci isn’t the only startup betting AI can replace physics simulation. UK-based PhysicsX raised a $300 million Series C in June, more than doubling its valuation to $2.4 billion. That round was also led by Temasek, meaning the same investor is backing two rival “physics foundation model” companies at once. PhysicsX targets manufacturing broadly across industries, a wider scope than Vinci’s semiconductor-first focus.

CFD-focused Luminary Cloud raised a $72 million Series B last September, with NVIDIA’s venture arm NVentures participating. Switzerland’s Neural Concept raised $100 million in a Series C led by Goldman Sachs Alternatives last December, pitching CAD-integrated generative design and simulation. Flexcompute also competes in CFD and electromagnetic simulation, while Rescale occupies a different layer entirely, providing cloud and HPC infrastructure for simulation workloads rather than an AI model of its own.

The semiconductor design landscape looks different again. Rather than building its own physics foundation model, EDA market leader Synopsys chose to partner with NVIDIA to accelerate chip design and simulation broadly with GPUs. Startups like ChipAgents and Ricursive Intelligence are focused not on physics simulation but on automating chip design and architecture itself through agents and reinforcement learning. Vinci’s bet on physics reasoning during the design process sits adjacent to, but distinct from, that layer of the stack.

Share

Leave a Reply

Your email address will not be published. Required fields are marked *