7 Semiconductor Startups Revolutionizing EdgeAI

Edge AI gives wings (of inference) to thin clients, as it puts 32-bit MCUs on steroids & translates directly into a new market for tens of billions of affordable (<$10 each) AI/ML-capable chips


This opens up a whole new market for affordable chips that can do artificial intelligence and machine learning tasks, costing less than $10 each. This is a big change from the expensive chips like TPUs/GPUs/DLAs (deep learning accelerator) currently required for implementing Cloud AI, costing tens of thousands of dollars each. We have compiled a list of 7 semiconductor startups that are making their mark with their AI Chipsets.

“Edge AI gives wings (of inference) to thin clients, as it puts 32-bit MCUs on steroids”

Read More: 3 Reasons Why TSMC Won’t Adopt High-NA EUV Lithography Until 2032

1. Mythic AI

Founded in 2012 by Mike Henry and Dave Fick in California, Mythic is creating a unified hardware and software platform that relies on unique analog compute-in-memory technology to deliver revolutionary power, cost, and performance that will shatter the limits restricting AI innovation. Mythic is making it much easier and more affordable to deploy powerful AI solutions, from the data center to the edge device.


Mythic Analog Matrix Processor (Mythic AMP™): The M1076 AMP utilizes the groundbreaking Mythic Analog Compute Engine (Mythic ACE™) to deliver the compute resources of a GPU at up to 1/10th the power consumption – all in a single chip. Additionally, this means your edge device can run complex AI applications, or multiple applications, at higher resolutions and faster frame-rates for better inference results.

M1076 Analog Matrix Processor: The M1076 Mythic AMP™ delivers up to 25 TOPS in a single chip for high-end edge AI applications.

MM1076 M.2 M Key Card: The MM1076 M.2 M key card enables high-performance, yet power-efficient AI inference for edge devices and edge servers.

2. BrainChip

Although the company is still in the early stages of deployment in a few niche markets. But we believe this company has the potency to go big.BrainChip is involved in edge AI on-chip processing and learning.

The company’s first-to-market neuromorphic processor, AkidaTM, mimics the human brain to analyze only essential sensor inputs at the point of acquisition, processing data with unparalleled efficiency, precision and economy of energy. Moreover, keeping machine learning local to the chip, without the need to access the cloud, dramatically reduces latency while improving privacy and data security.
Akida’s performance per microwatt can be orders of magnitude better than the current solutions on the market.”
In enabling effective edge compute to be universally deployable across real-world applications, such as connected cars, consumer electronics and industrial IoT, BrainChip is proving that on-chip AI close to the sensor is the future for customers’ products as well as the planet.

Read More: 3 Reasons Why TSMC Won’t Adopt High-NA EUV Lithography Until 2032

3. Edge Impulse

Build. Train. Optimize. AI for the edge.

Edge Impulse is an AI platform that focuses on developing and deploying machine learning (ML) models for edge devices. Their goal is to enable developers to build and optimize ML models that can run directly on edge devices, such as microcontrollers (MCUs), Linux CPUs, GPUs, and more.

Data Collection and Analysis: Edge Impulse allows developers to collect data from various sources, including their own sensor hardware, public datasets, simulations, or synthetic data.

Algorithm Development: With Edge Impulse, developers can leverage device-aware feature engineering tools and enterprise-ready AI architectures to develop algorithms optimized for health and industrial use cases.

Edge Deployment and Integration: The platform supports target-agnostic edge deployment, meaning the ML models and algorithms generated with Edge Impulse can be deployed on a wide range of edge hardware

4. DeepX

Korea – DEEPX an AI semiconductor technology company, is announcing that it has surpassed 40 customers for its flagship chip solution, DX-M1—the only AI accelerator on the market to combine low power consumption, high efficiency and performance, and cost-effectiveness.

The groundbreaking solution has been deployed for a hands-on trial to this customer pool, which spans global companies and domestic Korean enterprises across various sectors.

Currently, DEEPX’s DX-M1 has been installed for pre-qualification testing in the mass production and development products of more than 40 global companies, covering robotics and smart mobility, AI video security systems, and AI servers. In particular, the robotics and smart mobility sectors increasingly require state-of-the-art, low-power, high-performance AI chips like DX-M1 that can be embedded in small form factors to enable highly advanced technologies such as autonomous driving and cognition.

Read More: Why are quantum computers taking so long to perfect? – techovedas

5. TensTorrents

Tenstorrent is a computing company that develops processors designed to help in faster training and adaptability to future algorithms. Additionally,they offer scalable and efficient hardware and software solutions for deep learning applications. Some of their key products and solutions include:

Galaxy Systems: These systems deliver dense, high-performance AI compute built on an Ethernet-based mesh of 32 Tenstorrent Wormhole processors, allowing you to scale compute seamlessly.

TT-Buda: TT-Buda This is a high-level software stack that takes in AI frameworks and does a top-down approach, generating a program file to run an AI model on the Tenstorrent hardware.

TT-Metalium: This is a low-level software platform for a heterogeneous collection of CPUs and Tenstorrent devices, giving users direct access to the RISC-V processors, NoC (Network-on-Chip), and Matrix and Vector engines within the Tensix Core. Additionally,it’s great for development customers who want to customize their models, write new ones or even run non-machine learning code.

Tenstorrent is a fab-less AI chip design and software company, which means that they create and design silicon for machine learning, then use a foundry to make the hardware, then work with partners to create solutions. They have a mission of addressing the rapidly growing compute demands for software 2.0. They also offer career opportunities for those interested in the future of AI.

Read More: LLW DRAM: Samsung’s Announced Secret Weapon in the AI Revolution – techovedas

6. Axelera AI (incorporated, 2021)

Axelera offers next-gen computer vision with our powerful edge AI solution that prioritize performance, usability, power efficiency, and cost-effectiveness, particularly in the field of computer vision.

Furthermore,Metis AI Platform, their flagship product, revolutionizes AI inference at the Edge. At the core of this platform is the Metis AI Processing Unit (AIPU), which sets new standards price/performance and performance/Watt. The Metis platform enables computer vision applications to be more accessible and powerful than ever before.

The Voyager SDK, another product by Axelera, is a comprehensive software stack that simplifies the development of AI applications on the Metis AI platform. Moreover, this SDK streamlines the building, optimization, and deployment of models for maximum performance and efficiency at the Edge, without the need for extensive retraining.

Their product lineup includes various hardware solutions, such as the M.2 AI Edge Accelerator Module and the AI Vision Gateway, powered by different processors like the Intel Atom® x6000 CPU, Intel® 12th Gen Core™ i CPU, and 3rd gen. Intel® Xeon® D Scalable CPU. These solutions offer industry-leading performance, minimal power consumption, and easy integration into existing systems. Axelera AI’s mission is to lead the democratization of AI.

Read More: 10 Free AI Tools That Will Blow Your Mind – techovedas

7. Blaize

Blaize, formerly known as Thinci, is a company that develops software and hardware for AI workloads. It is a leading provider of a proprietary purpose-built, full-stack hardware architecture and low-code/no-code software platform.

Additionally,the company enables edge AI processing solutions at the network’s edge for computing in multiple large and rapidly growing markets — automotive, mobility, retail, security, industrial automation, medical devices, and more.

The CEO of Blaize is Dinakar Munagala has more than 22 years of experience in successfully leading global teams building graphics chips.
The company has offices in the US, UK, Japan, and India.

Blaize Graph Streaming processor (in the picture above

The Blaize Graph Streaming Processor is a high-performance data processing platform designed for real-time streaming analytics on large datasets.

Optimized for graph-based workloads and uses stateful analysis to process and analyze data streams in real-time.

Blaize’s platform is built on a custom GPU architecture and can be used for a variety of applications including fraud detection, recommendation engines, and predictive analytics

Blaize Xplorer X1600P PCIe Accelerator

-Blaize 1600 SoC with 16 GSP cores, providing 16 TOPs
-Commercial grade
-Soft ISP available to run on Blaize 1600 SoC
-PCIe Gen 3.0, 4 lanes

Join Our WhatsApp Community


In conclusion, the EdgeAI movement is reshaping the landscape of artificial intelligence, making it more accessible and affordable for everyone. Additionally,these innovative semiconductor startups, such as Mythic AI, BrainChip, Edge Impulse, DeepX, Tenstorrent, Axelera AI, and Blaize, are at the forefront of this revolution, ushering in a new era of powerful, cost-effective, and energy-efficient AI solutions at the edge.

Editorial Team
Editorial Team
Articles: 1898