Ambient Scientific Joins CELUS Design Platform
Ambient Scientific's AI-native processors will be available to engineers on the CELUS cloud electronics design platform from the architecture stage.
Updated on Nov. 13, 2026 in Semiconductors

Fabless semiconductor company Ambient Scientific and cloud-based electronics design automation company CELUS announced a partnership to bring Ambient Scientific's AI-native processors to the CELUS Design Platform. The first covered device is the GPX10 Pro system-on-chip, presented at electronica 2026 in Munich.
Why it matters
Processor choices made after the rest of a system is fixed can force a redesign or reduce product features. Placing AI-native silicon inside the design environment at the start of development aims to avoid that trade-off.
The GPX10 Pro includes ten MX8 AI cores in two power domains, an Arm Cortex-M4F core, and 2 MB of on-chip SRAM, with peak AI throughput of 512 GOPS and always-on keyword spotting below 100 µW. Figures are company-reported, and no third-party benchmarks are cited in the announcement.
The players
Ambient Scientific
A fabless semiconductor company that develops the A-Cube AI-native compute architecture and the GPX10 Pro system-on-chip.
CELUS
A cloud-based electronics design automation company headquartered in Germany, with additional offices in Porto, Portugal and Austin, Texas.
The details
Ambient Scientific devices will appear as CUBOs, the platform's reusable, application-ready design content, letting engineers add always-on AI while defining system architecture. Engineers can check the design against the sensors and supply rails of the board, with the GPX10 Pro supporting eight analog and 20 digital sensors at once.
Timeline
November 13, 2026: Partnership announced and article published.
Electronica 2026, Munich: Partnership and GPX10 Pro presented at Hall C5, Booth 558.
The Tech Race
The partnership extends the CELUS Design Platform, whose library already holds more than 560,000 components, by adding AI-native processors as reusable CUBO content. It places Ambient Scientific's A-Cube architecture in the design stage where chip choices are typically locked in, a competitive position against processor vendors that rely on later-stage integration.
For engineers, the partnership means always-on AI features can be specified at the architecture stage rather than retrofitted later, which may reduce redesign cycles. Product teams using the CELUS platform can evaluate GPX10 Pro-based designs against their sensor and power requirements before committing to hardware.
The takeaway
Choosing a processor early in product design avoids the cost of a later redesign or reduced features. Engineers evaluating AI-enabled hardware should check how a candidate chip fits their sensor and power budget at the architecture stage.
Further reading
For background on the chips at the center of this announcement, see Semiconductors.
Source note: This article includes information reported by EEJournal.



