GSI (GSIT) IAccess Alpha Virtual Best Ideas Spring Investment Conference 2026 summary

Technology and product strategy

• Focus on edge AI with proprietary Associative Processing Unit (APU) architecture, enabling compute-in-memory for low power and latency.

• Gemini-II targets edge environments like drones and satellites, offering configurable bit-width and massive parallelism.

• Plato, the next-generation chip, is designed for LLMs at the edge, aiming for sub-10W power consumption and tape-out in H1 2027.

• Over $175 million invested in APU R&D, funded by legacy SRAM business.

• 87 patents filed for APU technology, with ongoing aggressive IP protection.

Market opportunities and competitive positioning

• Edge AI market projected to grow from $20B to $120B by 2030, with target addressable markets doubling in five years.

• Initial deployments focus on defense and government, especially drones and surveillance, with commercial edge applications to follow.

• Proven performance advantage in drone POC, outperforming NVIDIA and Qualcomm in power and latency.

• Legacy SRAM products maintain market leadership, with high ASPs and gross margins, especially in rad-hard applications for satellites.

• Competitors have frozen roadmaps, ensuring continued leadership in high-performance memory.

Financial overview and funding

• Fiscal 2026 revenue expected to increase 25% over prior year, with trailing twelve-month revenue just under $25M.

• Over $70M in cash and cash equivalents after $47M equity raise in October.

• Operating expenses around $7M per quarter, with a recent $3M IP purchase for Plato development.

• High insider ownership at 20% and market cap around $320M.

• No current plans for additional fundraising, but future capital needs anticipated; SRAM business could be considered for sale if opportunity arises.