Analog Devices Added Tensilica DSP to SHARC Processors
The new 16nm SHARC-FX architecture boosts audio processing speed by 5x to support neural-network-based inference.
Updated on Sept. 25, 2026 in Semiconductors
Live Poll
Do you believe the latest in-car audio and AI technology will actually improve your driving experience?
Analog Devices has announced the integration of Cadence Tensilica DSP architecture into its new ADSP-SC84x/2184x automotive audio processor family. The rollout leverages 16nm process technology to enhance performance for real-time AI audio tasks.
Why it matters
As automotive cabin systems increasingly rely on neural-network-based features, the shift to 16nm silicon enables higher operating frequencies and efficiency. This architecture update addresses the rising demand for on-chip inference capabilities in next-generation vehicle platforms.
The SHARC-FX core utilizes a 16nm process node, a significant move from the previous 28nm generation. The architecture supports the LPDDR4 memory interface and includes a dedicated hardware security module to manage cybersecurity workloads.
The players
Analog Devices
A semiconductor company focused on data conversion and signal processing technologies for industrial and automotive applications.
Cadence
A provider of electronic design automation software and hardware intellectual property for chip development.
The details
The SHARC-FX core utilizes Tensilica Instruction Extension, a method that allows developers to add custom CPU instructions to the processor for specific tasks. By combining these custom instructions with Cadence compiler tools and Analog Devices' software ecosystem, the processor can execute real-time neural-network-based audio inference. The transition to 16nm lithography—the process of printing microscopic circuits onto silicon wafers—allows for higher transistor density compared to the older 28nm technology.
Timeline
September 25, 2026: Cadence announced the collaboration on the new DSP architecture.
The Tech Race
The SHARC-FX core represents the latest iteration in the evolution of the SHARC+ processor architecture, marking a transition from 28nm to 16nm nodes. This move reflects an industry-wide push to embed neural-network processing into automotive silicon to support increasingly complex audio features.
Automakers will gain access to these processors to enable faster and more efficient audio signal processing in future vehicle models. Developers working with this platform will require integration with the updated Cadence compiler tools to utilize the new instruction extensions.
The takeaway
This development highlights the shift toward putting dedicated AI inference hardware directly into automotive audio signal chains. Watch for how Analog Devices manages the adoption of this 16nm platform across its next-generation automotive product roadmap.
Further reading
For more on the latest advancements in chip design and processing, visit Semiconductors.
Source note: This article includes information reported by DQ.
Live Poll
Do you believe the latest in-car audio and AI technology will actually improve your driving experience?
