CATEGORY · 54 COMPANIES
AI Hardware / Devices
Consumer AI devices, wearables, and on-device AI.
#CompanyValuationARRHeadcountFresh
01RRain AIEnergy-efficient neuromorphic AI hardware company developing in-memory compute technology and accelerator chips. Creates chips that reduce AI inference and training energy consumption using spiking neural networks and memristive devices.——~43green02VVaire ComputingVaire Computing develops specialized AI inference accelerators optimized for large language models. The company's focus is cost-per-token efficiency and low-latency deployment for production inference workloads.$200M+—30-60green03MMeta Ray-Ban / OakleyAI smart glasses————04SSES AISES AI specializes in AI-discovered high-energy, high-power density lithium-metal and lithium-ion rechargeable batteries for electric vehicles, drones, urban air mobility, and energy storage. The company uses Molecular Universe, its proprietary AI agent, to accelerate material discovery and develops Avatar, an AI-powered battery health and safety management system.———green05GGroqCustom inference silicon (the Linear Processor Unit) and the GroqCloud low-latency inference service.—~$500M projected 2025 revenueNot separately reported—06HHumane AI PinWearable AI pin (wound down)————07PPlaud NoteAI voice recorder————08HHyperbolic LabsDecentralized GPU inference network enabling cost-efficient model serving and fine-tuning. Provides distributed compute infrastructure for LLM inference, image generation, and video processing by aggregating spare GPU capacity globally, bypassing traditional cloud bottlenecks.Estimated $100-200MEstimated $2-5M (early stage)~30green09FFriend.comAI companion necklace————10RRebellionsSouth Korean AI chip startup focused on inference chips (Atom, Rebel-Quad). Designs fabless AI chips optimized for inference workloads in data centers and edge applications.$2.34 billion—200+ estimatedgreen11LLattice SemiconductorPublic semiconductor company specializing in low-power, reprogrammable FPGAs (Field Programmable Gate Arrays) for edge AI inference, IoT, and datacenters. CertusPro-NX and MachXO5-NX optimize AI workloads.Public (NASDAQ: LSCC)$500M+—green12IION Storage SystemsION Storage Systems develops next-generation solid-state batteries using a patented 3D ceramic architecture that eliminates flammable liquid electrolytes and requires no external compression or cooling. The technology addresses safety and energy density demands for consumer electronics, electric vehicles, drones, and robotics.———green13YYC Bookface (Ycombinator Demo Day company) / Scale ComputeScale Compute provides distributed compute orchestration and GPU resource management for AI teams. The platform optimizes multi-GPU training and inference, offering dynamic resource allocation to reduce infrastructure costs and complexity in managing heterogeneous compute clusters.——estimated 5–10green14CCerebras SystemsWafer-scale AI accelerator maker (WSE-3) and inference cloud provider.—$510M revenue 2025 (+76% YoY)Not separately reported—15FFractileFractile develops SRAM-based AI inference accelerators using Memory Compute Fusion Architecture to co-locate memory and compute on a single die, reducing reliance on expensive HBM (High Bandwidth Memory). The company claims 100x faster inference and 10x cost reduction versus NVIDIA GPUs.$1B+—20-50green16RRabbit R1Large Action Model companion device————17CCrusoe EnergyCrusoe Energy is developing specialized hardware infrastructure for AI and high-performance computing, focusing on flare gas utilization and energy-efficient GPU compute. The company combines power generation with AI compute optimization.$1.5B+~$276M (2024); projected ~$1B (2025)estimated 150-200green18SSamsung Galaxy AIOn-device phone AI————19HHailoEdge AI processor startup developing specialized chips optimized for high-performance deep learning on edge devices. Partners with Raspberry Pi Foundation and major IoT/edge device manufacturers for deployment of efficient on-device AI.———green20AAyar LabsAyar Labs develops silicon photonics-based optical I/O solutions that use light-based interconnects to address data movement and bandwidth bottlenecks in AI data center clusters. The company's technology enables high-speed, energy-efficient data transmission between AI accelerators.$1B+—~100+green21CCelestial AISanta Clara-based AI infrastructure company developing photonic interconnect technology for data center networks. Celestial's technology enables high-speed, low-latency GPU interconnects, addressing critical bottlenecks in AI compute infrastructure scaling.$2.5B—200+green22AApple IntelligenceOS-level AI across Apple devices————23IInspur Information<cite index="18-4,18-5">Inspur Information is a leading domestic AI server supplier powering training and inference clusters, with racks and integrated systems tuned for large-model performance in Chinese data centers.</cite> The company serves as critical infrastructure provider for China's sovereign AI stack and large-scale model training operations.Public company (SZSE: 000977)Not specified in search resultsNot specifiedgreen24EEtched AIEtched AI is a hardware startup designing specialized AI inference chips. The company focuses on custom silicon optimized for specific AI model architectures to improve throughput and reduce latency compared to general-purpose GPUs.——~25green25SSpiNNcloud TechnologiesNeuromorphic supercomputing provider deploying specialized systems for brain-inspired AI workloads. Built 656,640-core neuromorphic system at Leipzig University simulating 650M neurons.———green26AAxelera AIAxelera AI develops specialized hardware and software for efficient AI inference at the edge. The company's Metis AI processor and platform optimize energy efficiency and latency for computer vision and machine learning tasks in resource-constrained environments.$600M+—100-150green27UUnconventional AIHardware startup designing neuromorphic and analog chips optimized for AI workloads. Pursues biology-inspired, energy-efficient computing using spiking neurons, oscillators, and mixed-signal designs to reduce power consumption by orders of magnitude versus digital GPUs.$4.5B——green28MMythicMythic develops power-efficient compute hardware and software for artificial intelligence inference at the edge. The company designs specialized processors and neural processing units optimized for running AI models on devices with minimal power consumption, enabling on-device AI capabilities.$1+ billion——green29LLightelligenceHybrid photonic-electronic computing company developing silicon photonics accelerators for AI and data centers. Offers near-packaged optics (NPO), co-packaged optics (CPO), and photonic computing accelerator cards. First publicly-listed AI silicon photonics chip company, debuted Hong Kong Stock Exchange April 2026.HKD 77.94B (~$10B USD equivalent at listing)——green30CCelero CommunicationsCelero Communications develops coherent DSP (Digital Signal Processing) platforms for high-bandwidth, low-power optical connectivity in AI data centers. The company enables AI systems to scale across accelerator clusters and cloud data center connections.~$500M+ (estimated)—~30-50green31NNeurophosNeurophos develops optical processing units (OPUs) using metasurface modulators and composite materials to accelerate AI inference. The company claims its photonic approach achieves 100x the performance and efficiency of traditional GPUs by routing photons through optical components for matrix multiplication at light speed.~$500M+ (estimated Series A)—~50green32CCambricon<cite index="31-1,31-2">Cambricon, founded in February 2015 and headquartered in Beijing, is an AI chip developer established by Chen Tianshi and has received financings from Heli Capital, Chinese Academy of Sciences Venture Capital and Alibaba, among others.</cite> <cite index="30-10,30-11,30-12">Cambricon has mastered core technologies in intelligent chips and foundational system software, with its 7nm Shenwei 590 chip delivering 80%-90% of NVIDIA A100's performance while achieving over 20% higher energy efficiency, with mass production begun for Alibaba Cloud and Tencent Cloud.</cite>—~RMB 1.111B (Q1 2025, 4,230% YoY)Not disclosedgreen33FFuriosaAISouth Korean AI chip designer specializing in domain-specific inference accelerators. Develops the RNGD (Renegade) AI accelerator using proprietary Tensor Contraction Processor architecture, achieving 3x better performance-per-watt than Nvidia H100 GPUs for LLM inference. Secured design win with LG AI Research.$735M——green34PPasqal<cite index="31-4">Built on Nobel Prize-winning research, Pasqal has been designing and building high-performance hardware and cloud-ready software since 2019 to address complex challenges in optimization, simulation, and artificial intelligence.</cite> <cite index="31-3">Pasqal is a global leading quantum technology company pioneering the development of neutral-atom quantum computers for industry, science, and governments worldwide.</cite> <cite index="31-5">Headquartered in France, Pasqal employs over 275 people including 70 phDs and serves over 25 clients and partners.</cite>$2B—275+green35HHorizon Robotics<cite index="34-31,34-32,34-33">Horizon Robotics is a Chinese technology company headquartered in Beijing that develops artificial intelligence chips used in self-driving cars and advanced driver assistance systems, founded in 2015 by several former Baidu employees including Yu Kai who previously led Baidu's self-driving car division.</cite> <cite index="30-15,30-16">The company provides automotive-grade intelligent driving computing solutions and software/hardware development platforms, and is the first and largest enterprise in China to achieve front-end mass production of automotive-grade intelligent in-vehicle computing solutions.</cite>IPO price HK$15.99, currently trading as of May 2026~RMB 2.384B (2024, 53.6% YoY growth)Not publicly disclosedgreen36OOLIX ComputingPhotonic AI chip startup developing Optical Tensor Processing Unit (OTPU) for AI inference. Uses hybrid photonic-electronic architecture with optical interconnects optimized for latency and power efficiency compared to silicon-only SRAM architectures. Positioned for inference-specific acceleration.$1B+——green37EEnfabricaEnfabrica designs high-bandwidth network interface controller fabric optimized for accelerated computing clusters. Its Accelerated Compute Fabric (ACF) SuperNIC chips provide multi-port 800-Gigabit-Ethernet connectivity for GPU servers in AI data centers.———green38EEnvision AI<cite index="30-2,30-3">Ally is an accessibility-first AI assistant that reads print, describes scenes, scans documents, searches the web, checks calendars, weather, and more, empowering thousands of people who are blind, low vision, or anyone who benefits from assistive technology</cite>. <cite index="30-7,30-8">Envision Glasses bring award-winning software to smartglasses with real-time text recognition, voice controls and hands-free video calling</cite>.———green39BBipedSmart harness using computer vision and gyroscope-based obstacle detection to alert blind and visually impaired users to collision risks via audio feedback. Real-time autonomy assistance inspired by self-driving car navigation principles.——~8green40QQuantum ArtTrapped-ion quantum computing company developing full-stack quantum systems targeting 1,000+ qubits. Raised $150M including $100M Series A plus $40M extension, positioning it as leading Israeli quantum player. Partners with leading institutes.——~60green41AAragoArago develops photonic AI accelerators using hybrid digital, analog, and optical design for energy-efficient, high-throughput AI workloads. Its deterministic architecture maximizes memory transfer efficiency and includes PyTorch-compatible software stack.———green42QQ-FactorStealth quantum computing startup building neutral atom-based quantum computer targeting 1M+ qubits. Emerged April 2026 with $24M seed from NFX, TPY Capital, Intel Capital. Founded by physicists from Weizmann and Technion. Unique technology approach.——~15-20 (est.)green43SSpeedataSpeedata develops an Analytics Processing Unit (APU) designed to accelerate big data analytics and AI workloads. Its APU targets Apache Spark workloads and other major data analytics platforms with specialized hardware acceleration.———green44OOpenLightSilicon photonics design company specializing in photonic application-specific integrated circuits (PASICs) using heterogeneous integration of indium phosphide and silicon photonics. Process Design Kit (PDK) validated at Tower Semiconductor, used by 25+ companies for production-grade PASICs across AI, datacom, and quantum applications.———green45SScintil PhotonicsHeterogeneous integrated photonics company developing silicon photonic co-packaged optics (CPO) for AI data centers. Created LEAF Light, industry's first single-chip DWDM (Dense Wavelength Division Multiplexing) light engine integrating multi-wavelength lasers with silicon photonics. Originated from CEA-Leti research institute.———green46AAira<cite index="34-3,34-7">Aira combines AI with real human visual interpreters to provide visual assistance for blind and low-vision users via smartphone camera or compatible smart glasses</cite>. <cite index="12-6">Customers, called Aira explorers, pay a monthly subscription fee to access trained, employed visual interpreters</cite>, with <cite index="12-8">an AI component called Access AI now free to all users</cite>.—null—green47SSiFiveSiFive is a fabless semiconductor IP company providing RISC-V CPU cores and system IP for AI data center infrastructure. The company licenses customizable processor cores designed to address memory bandwidth bottlenecks and orchestration in agentic AI systems, with support for CUDA and NVLink Fusion.$3.65B~$150M+ (2025 revenues >50% YoY growth)~508green48PParrots Inc. (Polly)Wheelchair-mounted AI voice assistant and cognitive support system offering real-time visual assistance, telecare, and predictive support for users with severe motor impairments and neurological conditions (ALS, cerebral palsy). Integrates with Alexa, joysticks, eye-tracking.——~8green49IInfiniLinkSilicon-photonics semiconductor startup building chiplets and high-speed optical serializers for AI data centres. Uses light instead of copper to move data between chips, delivering high bandwidth with lower power consumption.———green50XXscape PhotonicsXscape Photonics develops silicon photonics solutions using multi-color wavelength platforms ('ChromX') to deliver hundreds of wavelengths on a single optical fiber, providing flexible, efficient optical interconnect for AI data center scaling and power reduction.~$200M+ (estimated)—~20-30green51GGlovatrix (Fifth Sense)<cite index="18-1,18-6,18-7">Fifth Sense by Pune-based startup GlovatriX is an AI-powered 'glove/smartwatch' with sensors that use machine learning to map and learn sign language gestures, which are then translated into speech in real time</cite>. <cite index="15-17,15-18">Glovatrix is focused on enabling seamless communication between people of all abilities, with gloves that translate sign language gestures to speech and text using sensors and AI algorithms</cite>.———green52XXRAI GlassAR smart glasses providing real-time AI-generated captions and translations for deaf and hard-of-hearing users in conversations, presentations, and events. Transforms audio into visual subtitles on eyewear.——~6green53GGlidance (Glide)<cite index="31-7,31-8,31-11">Glide is a groundbreaking assistive technology for people who are blind or have low vision with two wheels connected to the ground and an adjustable telescoping handle that physically guides users, leveraging cutting-edge AI, sensors and robotics technology purpose-built for blind and low vision mobility</cite>. <cite index="31-16">Glide is the world's first autonomous, self-guided mobility aid designed for people who are blind or have low vision</cite>.———green54VVisualfyAI-based sound recognition platform converting environmental sounds into visual and vibration alerts for deaf and hard-of-hearing users. Visualfy Home detects doorbells, fire alarms, baby cries, and other critical sounds; Visualfy Places extends to public transportation and venues.€8.4M—~10green