GSI (GSIT) Corporate presentation summary

Business overview and strategy

• Specializes in high-performance memory solutions and the Associative Processing Unit (APU) for AI, leveraging over 30 years of semiconductor expertise.

• Transitioned from SRAM leadership to developing edge AI compute platforms, investing $175M in APU R&D and securing 96 APU patents.

• Focuses on edge AI markets where performance-per-watt is critical, targeting smart infrastructure, defense, autonomous drones, and industrial vision.

• Commercialization strategy centers on validating Gemini-II in real-world deployments and expanding into adjacent edge AI markets using reusable software and system integrator partnerships.

• Roadmap includes Gemini-II deployments in 2026 and Plato launch in 2028, with future enhanced APUs planned.

Technology and product highlights

• APU Compute-in-Memory (CIM) architecture reduces data movement and power consumption, achieving up to 98% lower energy use compared to NVIDIA A6000 in Cornell-led testing.

• Gemini-II APU delivers up to 300% better performance-per-watt and fastest response time in edge AI benchmarks.

• Plato APU targets lower power (2–10W), smaller footprint, and higher memory bandwidth for embedded autonomous systems.

• Radiation-hardened SRAM products serve aerospace, satellite, and defense markets, offering high margins and long revenue lifecycles.

Financial performance and position

• Last twelve months (LTM) revenue as of Q1 FY2027 was $25.1M, with $70M in cash and no debt.

• Stable SRAM revenue supports APU development; Q1 FY2027 saw higher R&D expenses for Plato and Gemini-II software.

• High-margin SRAM business provides financial foundation, with Sigma Quad and SigmaDDR products as growth drivers.

• Net loss for the quarter ended June 30, 2026, was $4.87M, with operating expenses driven by R&D and SG&A.

• Balance sheet as of June 30, 2026, shows $77M in cash and $108.4M in total assets.