AetherAI: Singapore's Edge AI Pioneer Bringing Intelligence to Every Device

Date: March 9, 2026

In an era where artificial intelligence increasingly dominates technology discussions, a fundamental challenge persists: how do we bring AI capabilities to the billions of IoT devices that don't have the power or connectivity to rely on cloud computing? AetherAI, a Singapore-based semiconductor startup, believes it has the answer—and the chips to prove it.

The Edge AI Revolution

Traditional AI deployment has relied heavily on cloud infrastructure, sending data from devices to remote servers for processing and then receiving insights back. While effective, this approach introduces latency, raises privacy concerns, and requires constant connectivity—limitations that become critical for applications like autonomous vehicles, smart city infrastructure, and industrial automation.

Edge AI solves these problems by processing data directly on the device, enabling real-time responses without cloud dependency. However, running AI models on battery-powered devices with limited computational resources has proven exceptionally challenging. This is where AetherAI comes in.

Singapore's Answer to Edge Processing

Founded in 2024 by former Nanyang Technological University researchers Dr. Michael Tan and Dr. Priya Sharma, AetherAI has developed a family of ultra-low-power AI accelerator chips specifically designed for edge devices. Their flagship product, the AE-1000, can run state-of-the-art machine learning models while consuming less power than a standard LED bulb.

"Our vision is to enable intelligence everywhere," explained Dr. Tan, who previously led semiconductor research at Singapore's Institute of Microelectronics. "From smart sensors in Singapore's public housing to agricultural drones in Malaysia, we want to power the next generation of intelligent devices."

How It Works

The AE-1000 employs a novel architectural approach that combines traditional neural network processing with specialized hardware acceleration. Unlike competing solutions that require models to be heavily compressed—often losing accuracy in the process—AetherAI's chip can run larger models with minimal performance degradation.

The chip supports popular frameworks including TensorFlow Lite, PyTorch Mobile, and ONNX Runtime, making it easy for developers to deploy existing AI models without extensive retraining. AetherAI also provides a software toolkit that helps optimize models for their hardware, achieving up to 85% efficiency gains compared to general-purpose processors.

Real-World Applications

AetherAI's technology is already being deployed across several sectors in Southeast Asia. In Singapore, the company has partnered with the Housing & Development Board to integrate its chips into smart sensors that monitor air quality, noise levels, and energy usage in public apartments. The system can detect anomalies in real-time and alert residents to potential issues before they become problems.

The company has also secured contracts with Indonesian agricultural technology firm Taniamart to power AI-enabled drones that monitor crop health across vast plantations. These drones can identify pest infestations and nutrient deficiencies without needing to connect to cloud servers, making them practical for remote areas with limited connectivity.

"The response has been overwhelming," said Dr. Sharma. "Every IoT manufacturer we speak to faces the same challenge—they want AI capabilities, but traditional processors can't deliver the performance within their power and cost constraints. We're solving that problem."

Funding and Growth

AetherAI recently closed a $25 million Series A funding round led by Singapore's Temasek Ventures, with participation from global semiconductor investors including Applied Ventures and Samsung Catalyst Fund. The investment will be used to expand the company's engineering team and accelerate production of its second-generation chip.

The company currently employs 45 people at its headquarters in the one-north innovation district and plans to double headcount by the end of 2027. AetherAI is also establishing partnerships with semiconductor fabrication facilities in Taiwan and Malaysia to scale manufacturing.

Singapore's Semiconductor Ambitions

AetherAI's emergence reflects Singapore's broader ambitions in the semiconductor and AI hardware space. The national government has committed billions of dollars to positioning the city-state as a global hub for chip design and AI semiconductor manufacturing.

With global demand for edge AI chips expected to exceed $50 billion by 2030, according to industry analysts at Gartner, Singapore is betting big on capturing a significant share of this growing market. AetherAI represents a promising homegrown contender in this race.

What's Next

The company is already working on its next-generation chip, the AE-2000, which will feature on-device training capabilities—allowing devices to learn from local data without ever transmitting sensitive information to the cloud. This capability could prove revolutionary for applications requiring continuous adaptation, such as personalized health monitoring or adaptive industrial control systems.

As Singapore continues its push to become Southeast Asia's AI hub, companies like AetherAI demonstrate that the island nation's ambitions extend beyond software and services. By tackling the fundamental hardware challenges of AI deployment, Singaporean innovators are laying the groundwork for an AI-powered future where intelligence resides not just in data centers, but in every device around us.

AetherAI is a portfolio company of Singapore's AI Singapore program and a member of the Global Semiconductor Alliance. For more information, visit their website at aetherai.sg.

Source: Straits Times

Related Resources: