The company’s physics-driven AI software, already deployed at leading semiconductor companies, is reported to operate up to 1000 times faster than conventional simulation tools, at higher accuracy and without needing customer data.
PALO ALTO, Calif.—Describing itself as “the pioneer of Physics-Driven AI for hardware design and simulation,” Vinci emerged from stealth mode in December to unveil a physics-driven AI system that operates like a team of hardware engineers, running thousands of verified simulations in hours rather than weeks.
One of Vinci’s co-founders, Hardik Kabaria, is an expert in computational geometry whose doctoral work at Stanford helped solve one of the hardest problems in simulation—automating high-fidelity meshing for complex, real-world geometries. The company’s other co-founder is Sarah Osentoski, a pioneer in large-scale machine learning and autonomous systems, according to a company release.
Vinci thermal simulation of an open source semiconductor chip. (Image: Vinci/Business Wire)
Together, they unite two rarely connected domains: deep, physics-based simulation and production-grade AI. Their expertise is reported to have drawn an exceptional engineering team, uniting top industry talent with some of the field’s leading researchers.
“The result is a physics-driven AI platform that pairs the accuracy engineers rely on with the scale and automation the next decade of hardware design requires,” the company stated in the release.
Accelerating workflows and delivering accuracy
According to Vinci, rising system complexity in areas such as advanced chip packaging, and 2.5D/3D IC is pushing traditional finite element analysis (FEA)-based simulation tools beyond their limits in speed, resolution, and accuracy. Traditional simulation workflows are time-intensive, break down on full manufacturing-resolution geometry, and rely on narrow domain expertise that is rapidly becoming a talent bottleneck.
As it emerged from stealth, following more than two years in development, Vinci announced the public debut of its physics-driven AI system to solve this critical gap.
“At Vinci, our goal is to let any engineer see how their design will perform once built,” said Kabaria, CEO of Vinci, in the release. “Vinci empowers engineers to simulate how designs will perform in seconds instead of days, doing so at a fraction of the compute cost. On next-generation geometries that conventional tools must simplify, such as nanometer-scale components on centimeter-scale dies, Vinci maintains full-fidelity accuracy.”
Vinci’s agentic system is reported to combine proven physics methods with an AI model to deliver 1,000-times faster simulations, without meshing, without hallucinations, and with guaranteed accuracy. While many AI solutions in this space remain aspirational, Vinci’s system is already deployed, powering next-generation design programs at three leading semiconductor manufacturers.
Pre-trained and production-ready, the system operates securely behind customer firewalls, requires no training on proprietary data, and delivers verified results immediately upon deployment, the company said in the release.
“On top of these deployments, more than 10 semiconductor companies have also independently benchmarked Vinci’s results against their traditional FEA solvers and experimental data,” the release stated. “In every case, Vinci’s simulations matched or exceeded the accuracy of established methods and, in several instances, correlated closely with experimental data, all while delivering results in a fraction of the time.
Founded in 2023 and headquartered in Silicon Valley, Vinci is backed by Xora, Khosla Ventures, and Eclipse.
“Few teams combine deep physics expertise with the ability to ship real, production-ready software,” said Charly Mwangi, partner at Eclipse, in the release. “Vinci’s technology is already demonstrating value in the field—accelerating workflows and delivering accuracy that engineers can trust.”
“Vinci has demonstrated the ability to deliver lightning-fast, high-accuracy simulations without requiring customer data for some of the world’s most complex physical devices, state-of-the-art semiconductor packages,” said Phil Inagaki, managing partner and chief investment officer at Xora, in the release. “Soon, we believe that Vinci’s platform will deliver not only simulation, but also co-design capabilities across a broad range of physics and hardware products, which will result in a radical expansion of what has been traditionally viewed as the EDA market.”