Anna Chapman

Jeffrey Milton Bernstein

Tony Reynaud

- GSI Technology announced that a Cornell University-led paper confirmed its Associative Processing Unit (APU) Compute-In-Memory (CIM) architectures can achieve **GPU-level performance** for large-scale AI applications.  **[1126741_45b19c1a017149acb0f98bb774e1ac94_0]** 
- The study found that the Gemini-I APU delivered comparable throughput to NVIDIA’s A6000 GPU on RAG workloads while consuming over **98% less energy** than a GPU.  **[1126741_45b19c1a017149acb0f98bb774e1ac94_0]** 
- GSI Technology's CEO, Lee-Lean Shu, highlighted the potential for compute-in-memory to disrupt the **$100 billion AI inference market**, noting that the company's second-generation APU, Gemini-II, offers roughly **10x faster throughput**.  **[1126741_45b19c1a017149acb0f98bb774e1ac94_1]**

GSI TECHNOLOGY (GSIT)