Insider Brief
- Applied Brain Research closed an oversubscribed seed round led by Two Small Fish Ventures to accelerate development of its TSP1 state space accelerator chip and edge AI software.
- The TSP1 enables real-time, low-power on-device voice AI for speech recognition and synthesis, delivering full vocabulary performance at under 30mW without cloud connectivity.
- ABR is targeting applications across AR, robotics, wearables, medical devices, and automotive, addressing latency, privacy, and power constraints in edge AI.
PRESS RELEASE — Applied Brain Research (ABR), the leader in low latency real-time on-device voice AI solutions, announced the successful closure of its oversubscribed seed funding round. The investment round was led by Two Small Fish Ventures (TSF), with Eva Lau joining the ABR board. The investment will fuel continued development and commercialization of ABR’s breakthrough TSP1 state space accelerator chip and software models for edge AI inference applications. ABR is a unique full-stack edge AI company, delivering patented state space models and purpose-built silicon to enable unmatched real-time, low-power voice AI fully on-device.
The oversubscribed seed round closed following the successful release of ABR’s TSP1 chip, the company’s first-generation state space accelerator purpose-built for real-time time series processing at the edge. The TSP1 enables on-device automatic speech recognition (ASR), text-to-speech (TTS), and advanced digital signal processing applications — capabilities that have traditionally required cloud connectivity or significantly more powerful hardware.
“Voice is becoming the default interface for the next wave of edge devices but using cloud voice AI solutions is a terrible experience, plagued by delays, reliability and privacy concerns while also being very expensive. And most edge devices are constrained by their battery life and computational resources limiting on-device capabilities,” said Kevin Conley, CEO of Applied Brain Research. “We’re elevating the experience of super responsive on-device voice interaction with low power solutions that also significantly reduce device operating costs.”
As was seen the past week at CES, ABR’s technology addresses critical pain points across multiple product segments where reliable, low-latency voice processing is essential but cloud connectivity is impractical, unreliable, expensive, and poses privacy concerns. There was clear resonance for ABR solutions in several categories:
- Augmented Reality Glasses: Elevating on-device voice control and responsive agent integration
- Robotics: Enabling zero delay cloud agent integrations and eliminating crippling operating costs
- Wearables: Unlocking a new level of AI voice control possibilities
- Medical Devices: More powerful biosensor on-device inference that protects sensitive data
- Automotive: Consistent in-vehicle voice experiences regardless of network connectivity
The TSP1’s architecture, optimized for state space model inference, represents a fundamental advance in efficient AI computation, enabling complex neural network inference at a fraction of the power budget required by conventional approaches. Full vocabulary speech-to-text and text-to-speech are now possible at under 30mW of power consumption, 10–100 times less than other available alternatives.
“The embedded AI market is at an inflection point,” said Eva Lau, General Partner at Two Small Fish Ventures. “Applied Brain Research has demonstrated that sophisticated voice AI doesn’t require the cloud. Their technology unlocks entirely new product categories and use cases that weren’t previously viable.”
ABR is currently engaging with partners across the AR, robotics, wearables, medical device, and automotive sectors. The company offers both the TSP1 chip and optimized software models tailored to specific application requirements.
About Applied Brain Research
Applied Brain Research Inc. (ABR) is an AI solution company focused on dramatically reducing the cost, latency and power consumption of edge AI applications. The company’s patented State Space Model (SSM) technology improves the user experience and economics of everything from AR glasses to sophisticated robots. ABR’s SSM technology solves the computational resource constraints that limit edge AI solutions by enabling deployment of powerful real-time AI, alleviating the constraints, latency and costs of cloud-based AI. ABR licenses SSM software for existing hardware applications and sells its accelerator ASIC, the TSP1 for industry leading lowest power levels.
Additional details about ABR’s SSMs and the TSP1 can be found at https://www.eetimes.com/applied-brain-research-demos-first-silicon-for-state-space-models/
Corporate information can be found at www.appliedbrainresearch.com.
