Ambient Scientific has provided fresh details of its long-term processor roadmap, outlining ambitions to expand its low-power AI technology beyond IoT devices and into applications including laptops, servers, and network infrastructure.
The company, which specialises in ultra-low-power AI processors for Edge devices, has previously discussed plans to scale its technology into higher-performance markets. However, founder and chief executive GP Singh has now offered a clearer picture of the types of systems future generations of Ambient processors could target.
“This is our first chip. We are actually building an entire road map,” Singh told IoT Insider. “When we started, we wanted to be only in the IoT and Edge devices, but now we have become ambitious. We have plans for more chips that will go in drones, humanoids, laptops, small servers, network appliances.”
Ambient’s current GPX10 processor entered volume production in 2024 and is designed for AI workloads in battery-powered devices such as wearables, industrial sensors, smart rings, agricultural monitoring systems, and personal safety products. The company says its proprietary digital-analogue hybrid architecture, known as DigAn, delivers AI processing at significantly lower power consumption than conventional approaches.
Singh argues that the industry faces growing challenges from the power and cooling requirements of today’s AI hardware.
“GPUs are running out of gas,” he said. “The size is too big, the power consumption is too big. It runs very hot, and to run it, I need a very special setup with liquid cooling.”
Ambient believes reducing power consumption will be critical as AI capabilities move from Cloud infrastructure into everyday devices. The company claims its architecture could enable more advanced intelligence in compact, battery-powered products where conventional processors remain impractical because of cost, size, or energy requirements.
The company plans to unveil a second, more powerful processor at Electronica later this year, with additional chips already under development.
While Ambient has previously spoken about scaling its technology beyond IoT applications, Singh’s latest comments provide a more detailed view of the company’s intended markets. The roadmap now extends from ultra-low-power Edge devices towards increasingly powerful computing platforms, including laptops, small servers, and networking equipment.
Whether the company can successfully scale its architecture into those markets remains to be seen. Competing against established semiconductor vendors will require more than efficiency gains alone, and Ambient must still prove that its technology can deliver at larger scales.
However, as the AI industry grapples with rising energy consumption and infrastructure costs, Singh believes efficiency will become as important as raw performance.
“The chips that can really make it happen, such that you can get very intense, very advanced intelligence in a very small form factor, those chips do not exist today,” he said. “Ambient has brought up the technology that will make the road possible for the next 20 years.”
There’s plenty of other editorial on our sister site, Electronic Specifier! Or you can always join in the conversation by commenting below or visiting our LinkedIn page.
