GSI (GSIT) Investor presentation summary

Company overview and technology

• Specializes in high-performance memory solutions and developed the Associative Processing Unit (APU) for AI applications.

• Pioneers in high-density SRAM and compute-in-memory (CIM) architectures, with 30 years of TSMC collaboration.

• Holds 144 patents, with 85 related to APU technology.

• APU targets edge AI, inference, and HPC, offering ultra-low power and real-time processing.

• Management team has decades of experience from leading semiconductor firms.

Product portfolio and innovation

• SigmaQuad and SigmaDDR SRAMs drive core business, recognized for high density, speed, and low power.

• Radiation-hardened SRAMs serve aerospace, satellite, and defense, with ASP up to $30K and 90%+ gross margin.

• APU platform includes Gemini-I, Gemini-II, and upcoming Plato, supporting LLMs and CNNs for edge devices.

• Gemini-II delivers GPU-class AI at 4x power efficiency, optimized for drones and tactical edge systems.

• Roadmap extends to Plato (12nm), targeting multimodal GenAI and computer vision for edge applications.

Market opportunity and strategy

• Edge AI market projected to grow at 18% CAGR, reaching $16.4B by 2030.

• Edge computing demand driven by need for real-time inference, lower latency, and security.

• Focused on defense and aerospace, leveraging DoD SBIR contracts and established agency relationships.

• Commercialization strategy prioritizes edge AI in drones, military vehicles, and GPS-denied environments.

• Pursuing additional government funding and strategic partnerships for future APU development.