IHW

What IHW does

IHW (Intelligent Hardware Korea / 인텔리전트 하드웨어 코리아), founded in 2021 and headquartered in Seongnam, South Korea, is a semiconductor company developing analog AI inference accelerators under the proprietary Analog Processing Unit (APU) architecture. The core technology performs matrix multiplication (MAC) operations in-memory using analog circuitry, eliminating the "memory wall" that consumes most of the energy in conventional digital AI accelerators such as GPUs, TPUs, and NPUs. The company claims its analog approach achieves digital-comparable inference accuracy while delivering materially better power efficiency, with the architecture validated through a peer-reviewed Nature publication (doi.org/10.1038/s41565-020-0655-z).

IHW's commercial product family is marketed under the Infertron brand, segmented by deployment target. Infertron-E/P (Phase I) targets on-sensor and on-things applications such as image processing, voice processing, security and surveillance, commercial drones, and autonomous vehicles, with Infertron-E specified at greater than 60 FPS at minimum cost and energy. Infertron-S/U (Phase I-III) targets on-device detection models and server/high-performance edge LLM inference. Use cases called out include medical diagnostics, robotics, drones, banking/stock analytics, security and surveillance, and generative AI workloads.

The go-to-market is B2B and OEM/embedded, with planned revenue streams from chip/hardware sales (APU and Infertron SKUs), IP and architecture licensing to semiconductor OEMs, and partnership-based platform development. The company's most material commercial signal to date is a March 2025 co-development agreement with Microchip Technology combining IHW's APU with Microchip's Membrain memory IP for edge AI/ML inference. IHW is pre-commercial, with 11-50 employees, and has raised approximately $15.3M cumulatively across a 2022 seed round led by Bridge Alliance Partners, a Feb 2025 Series 1 led by Mirae Asset Venture (with DSC Investment, HB Investment, Kolon Investment, and Woori Venture Partners), and a May 2025 grant from South Korea's Ministry of Science and ICT (MSIT). Kim Tae-hoon is CEO.

Name

IHW

Legal name

인텔리전트 하드웨어 코리아 (IHWK)

Website

https://ihw-ai.com

Company type

Private

Founded year

2021

Operating status

Operating

Headcount range

11–50 employees

Short description

IHW (Intelligent Hardware Korea) is a South Korean semiconductor company that builds analog AI inference chips (APU and Infertron product family) targeting edge, mobile, on-sensor, and server workloads as a lower-power alternative to GPU/TPU/NPU. It sells and licenses its analog compute IP to OEMs and partners, notably Microchip Technology.

Ownership category

akta.pro rank

Product category

AI Inference Semiconductors

NAICS

Semiconductor and Related Device Manufacturing (334413), Semiconductor and Other Electronic Component Manufacturing (3344)

SIC

Electronic Computers (3571)

akta.pro primary industry

AI Inference Hardware & Edge Accelerators (HDAAAAAC)

akta.pro secondary industries

AI Accelerators (GPUs/TPUs/NPUs/ASICs) (HDAAAAAA), Compute & Acceleration ICs (CPU/GPU/AI/FPGA/ASIC) (HDAHABAC)

Keywords

• Analog AI chips
• AI inference accelerators
• Edge AI hardware
• In-memory computing
• Semiconductor IP licensing

Headquarters

HQ city

Seongnam

HQ country

South Korea

HQ region

Asia

Offices1 record

Markets served

GTM type

B2B

Offering type

Hardware or Manufacturing

Cost components

Technology or R&D, Personnel, Operations, Marketing or Sales, Supply Chain

Revenue model

1. APU Chip / Hardware Sales: Sale of analog AI processing units (APUs) and Infertron product line hardware to semiconductor manufacturers, edge device makers, and AI hardware companies.
2. Technology Licensing / IP Licensing: Licensing of APU architecture and analog computing IP to semiconductor OEMs and chip manufacturers for integration into their own AI products.
3. Platform Development (Partnership-based): Co-development of analog compute platforms with partners (e.g., Microchip Technology) for specific vertical applications such as edge AI/ML inference, potentially generating revenue through joint products or component supply agreements.

Go-to-market motion1 record

Distribution channels1 record

Marketing channels3 records

Core offering

IHW designs and sells AI inference accelerator chips built on proprietary Analog Processing Unit (APU) technology, performing matrix computation in analog memory to deliver high energy efficiency and digital-comparable accuracy. Its Infertron-E/P and Infertron-S/U product lines address on-sensor/on-things and on-device/server AI inference respectively, targeting OEMs and AI hardware manufacturers across mobile, edge, cloud, and vertical applications including autonomous driving, medical diagnosis, robotics, and generative AI.

Product overview

IHW develops a comprehensive portfolio of AI inference accelerators based on proprietary Analog Processing Unit (APU) technology. The product lineup includes the Infertron-E/P series for ultra-efficient on-sensor/on-things AI applications (Phase I) and the Infertron-S/U series for high-performance on-device and server/edge AI applications (Phase I through IIII). These products target markets including image processing, voice processing, autonomous vehicles, medical diagnostics, commercial drones, robotics, security & surveillance, and generative AI applications (LLMs).

Differentiator

Problem solved

Functional benefit

Brands

• Infertron: AI inference hardware product line including Infertron-E/P (on-sensor/on-things AI) and Infertron-S/U (high performance AI solutions).
• APU (Analog Processing Unit)

Products and services

• Infertron-E An ultra-efficient AI inference chip for on-sensor and on-things applications, delivering over 60 FPS performance with minimum cost and energy consumption. Designed for semiconductor OEMs and edge device manufacturers building image processing, voice processing, security & surveillance, commercial drone, and autonomous vehicle applications.
• Infertron-P A cost-effective AI inference chip for on-sensor and on-things deployments, optimized for low cost and low energy consumption in compact, embedded intelligence applications. Pairs with Infertron-E as the Phase I product tier of IHW's roadmap.
• Infertron-S A high-performance AI inference chip for on-device applications, delivering high performance with low cost and low energy consumption. Targeted at high-performance on-device AI detection model deployments and related workloads.
• Infertron-U An ultra-high-performance AI inference chip for server and high-performance edge computing, delivering ultra performance with low cost and low energy consumption. Targeted at server providers running LLM-based applications and large-scale detection model inference.
• APU (Analog Processing Unit) technology Proprietary Analog Processing Unit (APU) technology — analog in-memory computing architecture that performs MAC operations in-memory to eliminate the memory wall bottleneck and deliver digital-comparable inference accuracy. Available to semiconductor OEMs and AI hardware manufacturers as licensable IP and integrated chip technology for edge, mobile, and server AI inference.

Quantifiable outcome

• Infertron-E delivers >60 FPS at minimum cost and energy for on-sensor AI applications
• +1 more outcomes

Segments3 records

Ideal customer profiles3 records

Technology focussed Yes

API detail

Has API

No

API docs

API detail

Core technology

AI maturity

App detail

AI capability8 records

Feature5 records

Partnerships

One partnership is on record.

• Microchip TechnologyflagshipStrategic or Co-development Partner · 13 March 2025Microchip Technology (Chandler, Arizona, USA) and IHW are collaborating to develop an analog compute platform for edge AI/ML inference. Microchip supplies its Membrain memory technology to the partnership. The collaboration targets the development of a commercially viable analog computing platform for edge AI/ML workloads, combining Microchip's memory IP with IHW's APU analog computing technology. The announcement was made on March 13, 2025.

Scale indicators1 record

Recent moves5 records

Expansion highlights6 records

Broad incumbents

• Tenstorrent: Broader AI compute company (RISC-V based training and inference accelerators) with overlap to IHW's LLM inference ambitions. Less specialized than IHW at the ultra-low-power edge tier, but a comparable AI accelerator player.

Emerging players

• Lightmatter: US-based photonic AI compute company pursuing non-digital acceleration for AI workloads. Adjacent to IHW in attacking the memory-wall and energy-efficiency bottlenecks, but via silicon photonics rather than analog in-memory.

Direct peers

• Hailo: Israel-based fabless AI chip company delivering edge AI inference processors for automotive, smart cities, retail, and industrial IoT. Overlaps with IHW's Infertron-S/U on-device and HP-edge inference positioning.
• Syntiant: US-based fabless semiconductor company producing ultra-low-power neural decision processors for always-on edge AI in IoT, hearables, and battery-powered devices. Directly comparable to IHW's Infertron-E/P on power/cost targeting.
• Mythic AI: US-based analog in-memory AI inference chip startup developing matrix-multiply-in-NRAM accelerators for edge devices. Closely comparable to IHW in both technology approach (analog in-memory MAC) and target use cases (low-power edge AI).
• GrAI Matter Labs: European fabless company producing neuromorphic-inspired edge AI inference chips. Competes with IHW in low-latency, low-power on-device AI for industrial and consumer edge applications.
• Kneron: US/Taiwan-based edge AI SoC company focused on on-device neural processing for IoT, smart devices, and automotive. Comparable to IHW on edge inference targeting and OEM/embedded go-to-market.
• Anaflash: Korea-based analog AI chip startup targeting ultra-low-power on-device inference. Direct geographic and technological peer to IHW within the Korean analog-in-memory ecosystem.
• BrainChip: Australia-listed neuromorphic processor company offering ultra-low-power on-device AI inference. Competes with IHW's Infertron-E/P in ultra-low-power edge AI, although via an event-based neuromorphic rather than analog-in-memory approach.
• Sima.ai: US-based edge AI inference chip company targeting computer vision and multimodal inference at the edge. Comparable to IHW's Infertron-S/U in the on-device and HP-edge tier.

Market position

Strengths5 records

Weaknesses5 records

Competitive moat4 records

Key risks6 records

Key highlights6 records

Customer concentration

Revenue estimate

Valuation estimate

Number of profiles

Profiles1 record

Funding overview

Funding rounds4 records

Investors11 records

Funding detail is available on the Subscription and Enterprise plan.Contact sales →

M&A

Investments

M&A and investment is available on the Subscription and Enterprise plan.Contact sales →

Frequently asked questions about IHW