Anker Innovations Unveils THUS™ AI Chip Platform and a Multi-Year Plan to Bring Local AI to Consumer Devices

Anker THUS Chip

Anker Innovations

Anker THUS Chip

To think, the device has to carry all those parameters across, many times per second, every single inference. THUS puts the computation where the model already lives. The model never has to move again”

— Steven Yang, Founder & CEO, Anker Innovations

BELLEVUE, WA, UNITED STATES, April 27, 2026 /EINPresswire.com/ -- Anker Innovations has announced THUS™, a proprietary AI chip platform designed to bring local neural-net AI to consumer devices, starting with audio and extending to mobile accessories and IoT devices.

An 80-Year Architecture Meets a New Era
Every chip built today descends from an architecture established in 1945 by mathematician John von Neumann. Its defining logic was divide and conquer: break any problem into discrete steps, translate those steps into code, and execute them one at a time. That sequential nature has a useful consequence — at any given moment, only a small fragment of the code is running. Memory and the processor can therefore be physically separated. Programs live in memory; the CPU fetches one instruction at a time, executes it and advances to the next. For eight decades, across mainframes, desktops and data centers, this model served computing brilliantly.

AI breaks that assumption entirely. A neural network does not divide a problem into steps — it processes end-to-end, drawing on millions or billions of learned parameters all at once. Every inference requires every one of those parameters to make the round trip between memory and the processor. In a data center, that movement is an engineering cost. In a wearable device that is powered by a battery smaller than a fingernail, it is a wall: more than 90 percent of chip power is consumed by simply moving data, leaving almost nothing for actual computation. Additionally, the model is often too large to fit on the device at all.

Introducing THUS™ AI Chip Platform
The answer has long existed in nature. Neurons in the human brain do not separate where information is stored from where it is processed — they do both in the same place. THUS™ applies that same principle to silicon.

Rather than shuttling model parameters between physically separate memory and processor, THUS™ embeds computation directly inside NOR Flash memory cells. The model parameters never have to move. The energy previously consumed by that movement is redirected to actual computation. NOR Flash-based compute-in-memory also requires approximately one-sixth the physical footprint of SRAM-based alternatives, making it viable for the most constrained consumer devices.

“Every AI chip built until now stores the model on one side and does the computation on the other. To think, the device has to carry all those parameters across, many times per second, every single inference. THUS™ puts the computation where the model already lives. The model never has to move again.”
— Steven Yang, Founder and CEO, Anker Innovations

The First THUS™ AI Chip: The World’s First Neural-Net CIM AI Audio Chip
ANKER chose to debut the first THUS™ AI chip in earbuds — the hardest environment for any AI chip. Earbuds operate on milliwatts of power with almost no room for silicon, while running noise cancellation without pause. Until now, that constraint held earbuds to small neural networks of a few hundred thousand parameters at most — not enough to handle complex, real-world noise effectively. The THUS™ chip lifts that ceiling: it supports several million parameters across multiple workloads, delivering up to 150 times more AI computing power on the environmental noise cancellation task than Anker's previous-generation flagship earphones.[1]

The chip's first disclosed feature, Clear Calls, targets a problem every earbud user knows: what the person on the other end of a call hears. Traditional Environmental Noise Cancellation (ENC) relies on signal processing algorithms or small neural networks that break down in loud, complex environments — bars, airports and busy streets. Clear Calls replaces those rules with a large neural network running entirely on-device, anchored by eight MEMS microphones and two bone conduction sensors that isolate the speaker's voice from its physical vibration. The result is a significantly cleaner call, in any environment.
Additional AI-powered features, including Signature Sound and Voice Control, will be detailed at Anker Day on May 21, 2026, in New York.

A Platform for the Future
THUS™ is a multi-year platform designed to bring local AI across Anker’s full product roadmap, from audio devices to mobile accessories and IoT devices. Future chips in the platform are intended to serve additional functions as the roadmap expands.

Availability
The first THUS™ AI chip will integrate into Soundcore's upcoming flagship true wireless earbuds. Full product details will be announced at Anker Day on May 21, 2026, in New York. Learn more about the THUS™ platform and upcoming products at ANKER THUS™.

About THUS™
THUS™ takes inspiration from “Thus have I heard,” a Buddhist phrase that reflects direct transmission, knowledge conveyed as it was experienced, without distortion.

The first THUS™ AI chip was developed through a strategic partnership with a global compute-memory technology company, combining Anker's AI expertise with collaborative hardware design. Anker holds exclusive license to the technology. The chip is fabricated in Germany as part of Anker's global manufacturing network.

About Anker Innovations
Anker Innovations is a global leader in charging and home energy storage technology and a developer of unique, consumer electronic products that support premium audio, mobile entertainment and the emerging smart home space. This innovation is being led by its three key brands: Anker, eufy and Soundcore. More information on Anker Innovations and its various brands can be found at anker.com.

[1] All performance figures are based on internal Anker lab tests. The 150x AI computing power figure applies specifically to the environmental noise cancellation task and reflects the architectural change from rule-based to neural-net AI processing. It is not a general performance claim and does not reflect a process node upgrade. The environmental noise cancellation workload referenced here runs as an on-device neural network model on the first THUS™ chip.

Eric Villines, Chief Communications Officer
eric.villines@anker.com

Adam Weissman
Anker / Soundcore
+1 917-693-3298
email us here

Legal Disclaimer:

EIN Presswire provides this news content "as is" without warranty of any kind. We do not accept any responsibility or liability for the accuracy, content, images, videos, licenses, completeness, legality, or reliability of the information contained in this article. If you have any complaints or copyright issues related to this article, kindly contact the author above.