MulticoreWare showcases cloud-to-car AI testing workflow at CES
The US artificial intelligence software specialist has demonstrated a real-time cloud-to-car development workflow at CES 2026, showing how advanced driver assistance systems can be tested in the cloud before deployment to vehicles.
MulticoreWare, a California-based provider of embedded and edge artificial intelligence software, said it had demonstrated a full cloud-to-car development and validation workflow at CES 2026 that runs advanced driver assistance system, or ADAS, workloads in the cloud and verifies them on edge hardware.
The company said the demonstration enables original equipment manufacturers and Tier-1 automotive suppliers to test Perception, Planning and Sensor Fusion workloads in the cloud before deploying them to in-vehicle systems.
According to MulticoreWare, the workflow used tools from US chipmaker Qualcomm Incorporated, including Qualcomm AI Hub and AIMET, to quantise an ADAS Perception model from FP32 to INT8. Inference was then validated on the Qualcomm Cloud AI 100, or QCR100, accelerator.
The company said this approach replicated automotive-grade performance and significantly reduced development cycles for Software Defined Vehicles.
“We are excited to partner with Qualcomm, leveraging MulticoreWare’s deep expertise in optimising and quantising AI models with Qualcomm AI toolchains for compute-intensive ADAS and in-cabin systems,” said Vish Rajalingam, vice president and general manager of the Mobility and Transportation business unit at MulticoreWare.
Related reading
- PayPal unveils new analytics tools for advertisers at CES
- AMD unveils new Ryzen AI processors and desktop chip
- NVIDIA unveils AI inference memory platform powered by BlueField-4
Rajat Sagar, vice president of product management at Qualcomm Incorporated, said: “We are pleased to work with MulticoreWare to demonstrate the power of the Qualcomm AI Hub. This collaboration highlights how our unified AI toolchain enables developers to seamlessly optimise and verify ADAS models from the cloud to the car, accelerating time-to-market.”
MulticoreWare said the workflow offers benefits including faster model conversion for edge performance, on-demand global validation without on-premises hardware, compatibility between cloud and automotive accelerators, and access to a growing library of pre-integrated models.
The Recap
- MulticoreWare demonstrated cloud-to-car AI workflow at CES 2026 show.
- Models converted from FP32 to INT8 using AIMET for inference.
- OEMs can validate workloads in cloud before edge deployment.
