Vinci, a Palo Alto startup that runs physics simulation for hardware design, has raised a $250 million Series B at a $1.5 billion valuation. Advent, Temasek and Xora co-led the round, with AMD Ventures, Madrona, Eclipse, Khosla Ventures and others taking part. Ten months ago, when Vinci came out of stealth, it had raised $46 million in total, with Xora leading the Series A and Eclipse the seed.
That’s a fast climb for an engineering software company. The pitch explains the hurry. Vinci says it can tell engineers how a chip or package will behave physically while the design is still changing, at a speed and resolution today’s simulation tools don’t reach.
What Vinci Sells
The product is called Continuous Physics Reasoning. Most hardware teams run simulation at checkpoints: a design goes to specialists, the solver runs for hours or days, and the answers come back after some decisions have already been made. Vinci claims deterministic, solver-accurate results in minutes, on manufacturing-scale designs with anywhere from hundreds of millions to more than 15 billion degrees of freedom. It covers thermal, thermo-mechanical and convective fluid behavior so far, and it’s already running on production engineering programs.
The design choice that matters is zero-shot operation. Surrogate models, the usual shortcut in AI-assisted simulation, have to be trained or tuned for specific designs and geometries. Vinci says its system needs no customer-specific training. Under the hood it combines automated design preparation, agentic orchestration, a foundation model for physics and GPU-native physics kernels.
Why Chips Come First
Vinci started in semiconductors because the physics is hardest there. AMD’s Brian Amick, senior vice president of technology and engineering, framed the problem as thermal, physical and electrical behavior interacting across the chip, the package and the board. That’s daily life in advanced packaging. Chiplets, stacked HBM, large interposers and hybrid bonding all push up power density and mechanical stress, and warpage becomes a yield and reliability problem as packages grow. Vinci’s second product, released in February, predicts exactly that: thermo-mechanical warpage at manufacturing scale.
AMD on the cap table is the strongest signal in the announcement. A chip designer backing a simulation startup suggests the existing toolchain is struggling to keep up with the packages it now has to ship.
Going After Incumbent Turf
Simulation is a business the big electronic design automation companies want to own. Synopsys bought Ansys largely for it, and Cadence and Siemens sell their own multiphysics tools. Vinci’s bet is that a GPU-native system can run simulation continuously, inside the design loop, where incumbent tools run it as a separate step.
The plan for the new money runs in two directions. Across physics, Vinci wants to move from thermal, mechanical and fluid behavior to broad coverage of hardware systems, from memory and advanced computing to vehicles, aircraft and satellites. Across engineering, it wants to go from predicting what a design will do to suggesting what should change, and eventually to generating designs from an engineer’s intent.
Those are company ambitions, and the announcement names no customers or benchmarks. The valuation says investors think the prediction engine works. The next round will hang on whether it can start making design decisions.