Executive Summary / Key Takeaways

• GSI Technology is undergoing a strategic transformation, pivoting from its legacy SRAM business to become a leader in AI edge computing with its Associative Processing Unit (APU) technology.
• The company's compute-in-memory architecture, exemplified by Gemini-II and the upcoming Plato chip, offers a significant competitive advantage in power efficiency and performance for demanding edge AI applications like drones and satellites.
• A recent $50 million equity financing provides crucial capital to accelerate the commercialization of Gemini-II and advance the development of Plato, positioning GSI Technology for high-growth markets.
• While the legacy SRAM business provides stable revenue and funds APU development, the company faces risks related to customer concentration, supply chain dependencies, and the inherent challenges of new product development and market adoption.
• GSI Technology's disciplined execution, coupled with strong government program engagement and validated technological differentiation, underpins its potential for long-term shareholder value creation in the burgeoning edge AI processor market.

The Evolution of GSI Technology: A Strategic Pivot to AI at the Edge

GSI Technology, founded in 1995, has historically been a provider of high-performance semiconductor memory solutions, primarily focusing on Very Fast SRAMs for diverse sectors including networking, telecommunications, test equipment, military/defense, and aerospace. A significant turning point in the company's trajectory occurred in fiscal 2016 with the acquisition of MikaMonu Group Ltd., an Israel-based company specializing in in-place associative computing. This strategic move marked GSI Technology's principal objective to develop Associative Processing Units (APUs), signaling a profound shift in its core business focus.

The rationale behind this pivot is clear: the networking and telecommunications market, a traditional stronghold for SRAM, has been in decline and is expected to continue this trend. Consequently, GSI Technology has strategically leveraged the revenue generated from its stable SRAM business to fund the intensive research and development required for its new APU solutions. This dual-pronged approach allows the company to maintain a revenue base while investing in a high-growth future. The broader industry landscape is witnessing an unprecedented surge in Artificial Intelligence (AI), particularly at the "edge"—devices and systems that process data locally rather than relying solely on centralized cloud infrastructure. This shift demands entirely new compute architectures that prioritize power efficiency, low latency, and on-device intelligence, areas where GSI Technology's APU technology is designed to excel.

The Technological Edge: Compute-in-Memory for AI

At the heart of GSI Technology's strategic transformation is its proprietary Associative Processing Unit (APU) technology, built on a compute-in-memory (CIM) architecture. This design fundamentally differentiates GSI Technology from traditional Von Neumann architectures, which separate processing from memory, leading to bottlenecks and higher power consumption. GSI Technology's CIM approach integrates computation directly within the memory, enabling significantly faster and more energy-efficient processing, particularly for data-intensive AI workloads.

The tangible benefits of this technology have been independently validated. A research paper published by Cornell University in mid-October 2025 verified that GSI Technology's Gemini-I chip performed on par with NVIDIA's A6000 on certain AI tasks, while consuming approximately 98% less energy. This stark difference in power consumption underscores the disruptive potential of GSI Technology's approach, especially for power-constrained environments.

Building on the foundation of Gemini-I, the company's Gemini-II chip represents a substantial leap forward. Gemini-II is designed to deliver 8x the memory and 10x the performance of its predecessor. It is engineered for complex edge AI capabilities, operating at an extremely low power profile of around 15 watts, making it ideal for applications in drones, defense systems, robotics, and mobile platforms where traditional GPUs struggle to compete. In proof-of-concept (POC) engagements, Gemini-II has demonstrated first-response times up to three times faster than alternative solutions, a critical advantage in applications like drone and defense systems where milliseconds matter for mission endurance. Gemini-II is also a flexible bit processor, capable of handling operations from 1-bit to 32-bit or larger efficiently, which is particularly well-suited for large language models (LLMs) at the edge.

Looking further ahead, GSI Technology is developing Plato, its next-generation APU. Plato is designed to address the rapidly expanding market for LLMs and Agentic AI applications at the edge. Leveraging the Gemini-II architecture, Plato aims to deliver data center-level performance using very low power, projected to be as little as 4 watts and a maximum of 12 to 15 watts. Plato will integrate a camera interface directly into the chip, along with enhanced connectivity features, enabling it to interface with a wide range of sensors. This makes Plato particularly well-suited for AI agents requiring object recognition and real-time decision-making in the physical world. The company plans for the tape-out of the Plato chip in early calendar 2027, with a strategic focus on acquiring necessary IP and expanding software development teams to accelerate its time to market.

This technological differentiation forms a strong competitive moat for GSI Technology. The ability to deliver GPU-class performance at a fraction of the power translates directly into higher value for customers in specialized markets. This can lead to higher average selling prices (ASPs) and superior gross margins for GSI Technology's APU products, enhancing its financial performance and market positioning in high-value niches. The company's radiation-hardened and radiation-tolerant SRAMs, which carry significantly higher gross margins than traditional SRAMs, further exemplify its ability to command premium pricing in specialized, high-reliability applications.

Financial Performance and Strategic Funding

GSI Technology's recent financial performance reflects its ongoing strategic pivot and disciplined operational management. For the second quarter of fiscal year 2026 (ended September 30, 2025), the company reported net revenues of $6.4 million, a substantial 41.6% increase compared to $4.6 million in the same period of fiscal year 2025. This growth was primarily driven by strong market momentum for its leading SRAM solutions, with the overall average selling price of units shipped increasing by 19.2% and unit shipments rising by 19% year-over-year. For the first six months of fiscal year 2026, net revenues reached $12.7 million, up 38% from $9.2 million in the prior year period.

Profitability metrics have also shown significant improvement. The gross margin for Q2 fiscal 2026 was 54.8%, a notable increase from 38.6% in the prior year quarter. For the first six months of fiscal 2026, gross margin expanded to 56.5% from 42.5% in the comparable period, primarily due to a favorable product and customer mix and the leverage of fixed overhead on higher shipment levels.

The company has also demonstrated effective cost control, reducing its net loss to $3.2 million in Q2 fiscal 2026, down from $5.5 million in the prior year period. For the full fiscal year 2025, GSI Technology meaningfully reduced its net loss by 47% to $10.6 million from $20.1 million in fiscal 2024, driven by a 35% reduction in operating expenses.

Liquidity has been significantly bolstered by strategic financial initiatives. As of September 30, 2025, GSI Technology held $25.3 million in cash and cash equivalents with no debt, indicating a strong balance sheet. This position was further strengthened by the sale and leaseback of its Sunnyvale headquarters in June 2024, which generated $11.3 million in cash. Between May and August 2025, the company raised an additional $14.3 million net through an At-the-Market (ATM) offering. Most critically, on October 21, 2025, GSI Technology closed a registered direct offering, raising approximately $50 million in gross proceeds (before $3 million in offering expenses) through the issuance of common stock and pre-funded warrants. This substantial capital infusion is being strategically deployed to accelerate the hardware and software development for both Gemini-II and Plato, including acquiring necessary IP for Plato and expanding software teams.

Management believes this enhanced liquidity will be sufficient to meet its cash needs for working capital and capital expenditures for at least the next 12 months, providing a critical runway for its APU roadmap.

Market Opportunity and Competitive Dynamics

GSI Technology is strategically targeting high-growth markets where its APU technology offers a decisive competitive advantage. The global edge AI processor market is projected to reach $9.6 billion by 2030, with the drone market, a key initial focus for GSI Technology's edge AI strategy, projected to reach $2.7 billion by 2030. These markets demand high volumes of compact, power-efficient devices, aligning perfectly with the core strengths of GSI Technology's compute-in-memory architecture.

In the broader semiconductor and AI landscape, GSI Technology operates as a specialized niche player compared to industry giants like NVIDIA (NVDA), Micron Technology (MU), Intel Corporation (INTC), and Rambus Inc. (RMBS). While NVIDIA dominates the data center AI hardware market with GPUs that can consume up to 2kW of power, the edge market requires a fundamentally different compute architecture. GSI Technology's Gemini-II, with its 15W power profile and GPU-class performance on specific AI tasks, directly addresses this need where competitors "simply cannot compete." Its faster time-to-first-token is a critical differentiator in drone and defense applications.

GSI Technology's competitive strategy is not to directly challenge these incumbents in their core data center domains, but rather to carve out a leadership position in specialized edge AI applications. The company's engagement with U.S. government agencies through Small Business Innovation Research (SBIR) programs is a testament to this strategy. Current contracts include prototype agreements with the Space Development Agency (SDA) for Next-Generation APU-2 development (recently expanded to $2 million, including radiation-hardened testing for APU3), and a $1.10 million SBIR with the U.S. Air Force Research Laboratory (AFRL) for specialized algorithms for APU2 in space. A recent U.S. Army contract for Gemini-II in edge computing AI solutions, particularly for 1-bit LLMs with low power and latency, further solidifies its defense market penetration. These government agreements not only provide crucial funding but also serve as validation and early adoption pathways for its cutting-edge technology.

Despite its innovative APU technology, GSI Technology faces competitive pressures in its legacy SRAM business, which is characterized by price erosion and intense competition from larger players with greater resources and manufacturing capabilities. The company is also exposed to risks associated with customer concentration, with two customers accounting for 45% and 30% of accounts receivable at September 30, 2025. Furthermore, its dependence on single-source suppliers like TSMC (TSM) for wafers and ASE (ASX) for packaging, particularly given geopolitical tensions in Taiwan, presents a significant operational risk. Supply chain constraints, exacerbated by tariffs shifting assembly operations to Taiwan, have already impacted SRAM lead times, leading to a stable rather than growth outlook for SRAM revenue in the near term.

Outlook and Execution Roadmap

GSI Technology's outlook is firmly centered on the accelerated commercialization of its APU products and the expansion of its software ecosystem. For the third quarter of fiscal year 2026, the company expects net revenues in the range of $6.0 million to $6.8 million, with a gross margin of approximately 54% to 56%. While SRAM revenue is projected to remain stable for the remainder of fiscal year 2026 due to extended lead times, the focus is on converting APU proof-of-concept projects into commercial success.

Pilot shipments of Gemini-II are anticipated in the first half of calendar year 2026, with the potential for substantial revenues from these POCs in the second half of calendar year 2026. The company is actively investing the proceeds from its recent $50 million financing into acquiring intellectual property for Plato and expanding its software and application teams for both Gemini-II and Plato. This includes developing robust software libraries, algorithms, and compiler tools to enhance developer accessibility and flexibility, crucial for broader market adoption. The tape-out of the Plato chip, a major milestone, is targeted for early calendar year 2027. GSI Technology is also actively engaging strategic partners for Plato to secure funding, collaborate on testing, and guide future enhancements, recognizing the estimated $50 million development cost for the program.

The company's strategic review, initiated in May 2024, continues to explore various options to maximize shareholder value, including potential equity or debt financing, asset divestiture, technology licensing, or a sale of the company. This ongoing evaluation underscores management's commitment to securing the necessary capital and strategic alignment to fully realize the potential of its APU technology.

Conclusion

GSI Technology stands at a pivotal juncture, strategically transforming from a legacy SRAM provider into an innovative force in the burgeoning AI edge computing market. Its compute-in-memory APU technology, particularly Gemini-II and the upcoming Plato chip, offers a compelling value proposition with demonstrated advantages in power efficiency and performance against traditional GPU solutions. The recent $50 million equity financing provides critical fuel to accelerate this strategic pivot, enabling the company to invest in its technological roadmap and pursue high-growth opportunities in defense, drone, and satellite applications.

While the company's disciplined operational execution and strong government partnerships provide a solid foundation, investors must consider the inherent risks of new product development, customer concentration, and supply chain vulnerabilities. GSI Technology's ability to successfully convert its technological prowess into widespread commercial adoption, particularly for Gemini-II in the near term and Plato in the longer term, will be paramount. The company's future hinges on its continued ability to execute its ambitious APU roadmap, leverage its unique technological differentiators, and effectively navigate the competitive landscape to capture a significant share of the rapidly expanding edge AI market.