Omer Keilaf, CEO and Founder of Innovizpositions automotive-grade LiDAR as the foundational infrastructure for physical AI and next-generation world models.
This whitepaper will be published in two parts, with the first part being published today and the second part expected to be published next month.
Innoviz Technologies Ltd., a leading supplier of high-performance LiDAR solutions, announced Part I of a new white paper titled “Innoviz and the Rise of Physical AI: Bringing Life to World Models.” This comprehensive white paper details how real-time, high-fidelity 3D perception enables an emerging class of physics-based AI systems and world models.
The white paper, written by Omer Keilaf, CEO and founder of Innoviz, argues that artificial intelligence is entering a fundamentally new phase. After transforming digital systems through software and large-scale language models, AI is now moving into the physical world, powering vehicles, robots, infrastructure, and machines that need to perceive, reason, and operate with reliable accuracy. This change is often referred to as physical AI, and at its core is perception.
Also read: AiThority interview with Arun Subramaniyan, Founder and CEO of Articul8 AI
As computing power continues to accelerate, the white paper argues that the primary bottleneck in the deployment of physical AI is no longer inference power, but access to high-fidelity, real-time 3D data. Cameras capture two-dimensional projections and probabilistically infer depth, but radars have no spatial resolution. In contrast, LiDAR directly measures distance and converts photons into precise shapes.
Recent industry initiatives, such as efforts to develop Industry World Models that incorporate physics, engineering principles, material behavior, and validated industry knowledge, demonstrate how AI is evolving beyond text-trained models to science-based systems that must remain in sync with real-world conditions. Matching these models with reality may require a continuous stream of rich real-time 3D data.
Innoviz’s automotive-grade LiDAR solutions, including platforms such as InnovizSMART and InnovizSMARTer, are designed to deliver high-throughput 3D sensing that can keep next-generation world models up to date with live spatial information from vehicles, robots, industrial environments, and smart city infrastructure.
“AI systems that interact with the physical world should not rely solely on inference. The emergence of world models marks the beginning of a new era of computing rooted in physics, geometry, and verified real-world data. We believe Innoviz is well-positioned to help enable these models by serving as the high-fidelity, real-time perception layer needed for secure and scalable physical AI.” Omer Keilaf said.
Also read: Cheap and Fast: LLM Cascade Strategy (Frugal GPT)
[To share your insights with us, please write to psen@itechseries.com ]
