The startup develops deep sub-threshold analog in-memory computation semiconductors that execute data-center-grade AI workloads with significantly lower power consumption and no active cooling. This technology processes data within memory, enabling industries to efficiently handle complex tasks while minimizing energy use.
Funding
$50.5M raised to dateRaised to date based on public sources. This may differ from the amount the company actually raised and is based only on what is publicly available on the internet.
Founders
Product
Problem
Current AI inference hardware relies on digital chips that consume significant power, leading to high operational costs and limiting the scalability and economic viability of generative AI applications, especially in data centers and edge deployments. Existing solutions struggle to balance performance, energy efficiency, and cost-effectiveness, hindering the widespread adoption of AI.
Solution
Sagence AI offers analog in-memory compute solutions that address the power and cost limitations of traditional digital AI inference hardware. By integrating compute and memory on a single device and operating in the deep subthreshold regime, Sagence AI's technology achieves significantly lower power consumption and reduced latency. This approach enables high-performance AI inference at a fraction of the cost and energy, making it economically viable to deploy generative AI applications at scale, from hyper-scale data centers to edge devices. The company's solutions eliminate the need for active cooling and reduce the number of boards and racks required, further lowering acquisition and operating costs.
Target Audience
Sagence AI targets hyper-scale data centers, enterprises with private data centers, and edge AI deployment companies seeking cost-effective and energy-efficient AI inference solutions. Specific industries include retail, healthcare, manufacturing, finance, security, education, and research.
Features
- Analog in-memory compute architecture that performs computations directly within memory cells
- Deep subthreshold operation for ultra-low power consumption
- Multi-level non-volatile memory for dense and power-efficient data storage
- Integrated storage and compute on a single device, eliminating data transfer bottlenecks
- Compile-time resource allocation for consistent, low-latency performance
- Compatibility with existing infrastructure and deployment scenarios
- Deterministic, programmable architecture for real-time defect detection
- 100x lower MAC power compared to traditional digital architectures