Always-On Machine Learning Solutions
for Battery Powered Devices
Syntiant is a new breed of a semiconductor company. Combining the best of deep learning and semiconductor design, Syntiant products fundamentally advance machine learning capacities in edge devices.
Larger Networks, Lower Power
Traditional semiconductor solutions are unable to support advanced machine learning applications at the device level today because of the high power costs of moving and processing data.
Focused investments are being made in silicon architectures to reduce this power largely by building chips with memory closer to the computational units. These techniques are often costly and complex, and while they can reduce the data movement penalties, they cannot eliminate them.
Built from the ground up to run deep learning algorithms, Syntiant’s architecture achieves breakthrough performance by slashing memory power consumption. Syntiant’s processors enable devices to be approximately 200x more efficient, providing 20x the throughput over current low-power MCU solutions, enabling larger networks at significantly lower power.
Massively Parallel Architecture
Syntiant breaks the chains of digital data path architectures with massively parallel neural networks ideally suited for high computational workloads required for real time applications.
Syntiant’s solutions are ideal for demanding always-on voice, sensor and image tasks.
Syntiant's ultra-low-power, high-performance NDPs are ideal for always-on applications in battery-powered devices.
ON A CHIP
Reduced time to revenue with full out of the box solutions
ANALOG NEURAL NETWORK
Ideally suited for neural computation with high accuracy at ultra-low-power
AI FOR THE SMALLEST DEVICES
Bringing neural processing to devices from ear pods to mobile phones