Nvidia has agreed to acquire SoftBank’s Arm Limited for $40 billion. The combination brings together Nvidia’s AI computing platform with Arm’s advanced processor designs to create what Nvidia calls the premier computing company for the age of artificial intelligence (AI). The move will help Nvidia accelerate innovation while expanding into large, high-growth markets. SoftBank will remain committed to Arm’s long-term success through its ownership stake in Nvidia, expected to be under 10%.
“In the years ahead, trillions of computers running AI will create a new internet-of-things that is thousands of times larger than today’s internet-of-people,” said Jensen Huang, Founder & CEO of Nvidia. “Uniting Nvidia’s AI computing capabilities with the vast ecosystem of Arm’s CPU, we can advance computing from the cloud, smartphones, PCs, self-driving cars and robotics, to edge IoT, and expand AI computing to every corner of the globe.
Arm will remain headquartered in Cambridge, and plans call for expanding its site by building an AI research facility supporting developments in healthcare, life sciences, robotics, and self-driving cars. To attract researchers and scientists from the U.K. and around the world, Nvidia will build a state-of-the-art AI supercomputer powered by Arm CPUs at Arm Cambridge.
Nvidia says that Arm will continue to operate its open-licensing model while maintaining the customer neutrality that has been foundational to its success, with 180 billion chips shipped to-date by its licensees. Arm partners will also benefit from Nvidia’s innovations.
Arm technology has been used in healthcare, life sciences, and robotics, but it is also a big part of Nvidia’s self-driving car computer, according to Huang. For instance, the Drive AGX Pegasus and Xavier developer kits come with 8-core Carmel CPUs (central processing units) based on the ARM v8 ISA (instructional set architecture).
Startup Voyage uses Pegasus in its third-generation robotaxi launched in August. The G3, pitched as a self-cleaning driverless vehicle, is a Chrysler Pacifica Hybrid minivan retrofitted with self-driving technology. It also has a COVID-19-friendly addition in an ambulance-grade ultraviolet light disinfectant system to keep passengers healthy.
Voyage used the Drive AGX platform to power its previous-generation vehicles’ Shield automatic emergency braking system. With the latest G3 robotaxi, the startup is employing the 320-TOPS (tera operations per second) Pegasus to process sensor data and run redundant deep neural networks simultaneously for driverless operation.
The G3 is Voyage’s first production-ready robotaxi, meaning it will not require a safety driver. It features three deeply integrated, proprietary, self-driving technologies—the Commander brain, Shield collision mitigation system, and Telessist remote operations solution. Those are integrated into the purpose-built Pacificas developed specifically by FCA for the integration of automated technology.
The Pegasus is deeply embedded into the G3’s compute platform. It uses Nvidia’s Xavier system-on-a-chip (SoC), which contains 9 billion transistors to process vast amounts of data, as well as thousands of safety mechanisms to address random hardware failures. Inside the Xavier SoC, there are six types of processors—ISP (image signal processor), VPU (video processing unit), PVA (programmable vision accelerator), DLA (deep learning accelerator), CUDA GPU (graphics processing unit), and CPU. They provide the compute necessary to run multiple diverse and redundant deep neural networks simultaneously for robust perception, prediction, and planning.
The G3 is designed to be a safe, shared vehicle for passengers. COVID-19 devastated ride-hailing businesses globally, as human drivers pose an inherent transmission risk. The pandemic challenged Voyage to build a better-shared vehicle to keep riders safe. The company partnered with auto supplier GHSP to adapt that company’s grēnlite solution using ultraviolet-C light to destroy pathogens (such as viruses, fungus, and bacteria) inside the vehicle. It does this by inactivating the pathogen’s DNA and halting its reproductive cycle.
G3 development prioritized not only performance and safety but also cost and scalability, according to Voyage. The company rigorously evaluated suppliers and components to drive down costs and deliver a vehicle at 50% of the prior generation’s cost. Sensor and compute costs are reduced by 65% and 25%, relative to the G2, helped in part by Nvidia’s platform.
The G3 represents a major milestone for Voyage, putting the company on the path to delivering profitable robotaxi services. It expects to continue reducing vehicle costs in future generations as the self-driving industry migrates down the cost curve.
