GT
GSI TECHNOLOGY INC (GSIT)·Q4 2025 Earnings Summary
Executive Summary
- Q4 FY2025 revenue was $5.883M, up 14% YoY and 9% QoQ, with gross margin at 56.1% (vs. 51.6% YoY and 54.0% QoQ); net loss narrowed to $2.2M ($0.09 per share) on reduced operating expenses and SBIR offsets .
- Customer mix improved: KYEC contributed $1.7M (29.5% of revenue) vs. $0.544M (10.6%) a year ago, while Nokia declined to 7.5% of revenue; SigmaQuad shipments were 39.3% and military/defense shipments 30.7% .
- Management announced an initial high-margin radiation-hardened SRAM order from a North American prime contractor and progress on Gemini-II/Plato edge-AI programs; Q1 FY2026 guidance: revenue $5.5–$6.3M, GM 56–58% .
- Street consensus (S&P Global) for Q4 FY2025 EPS and revenue was unavailable; the print landed in-line with prior Q4 guidance ranges for revenue and GM provided in January .
What Went Well and What Went Wrong
What Went Well
- Secured initial order for radiation-hardened SRAM from a North American prime contractor; management expects follow-on orders and highlights materially higher gross margins vs. traditional SRAM .
- Strong SRAM demand drove revenue growth; KYEC exposure rose to 29.5% of revenue, reflecting robust orders tied to AI chip manufacturing systems; gross margin expanded to 56.1% on mix and higher revenue .
- APU roadmap execution on track: Gemini-II production-ready chips and Leda-2 boards targeted by end of Q1 FY2026; SBIR milestones with AFRL/SDA progressing, with deliveries of boards/cards and YOLO algorithms .
What Went Wrong
- Ongoing customer concentration risk: KYEC at 29.5% and Nokia volatility (7.5% of revenue, down from 13.5% YoY); military/defense shipments softened YoY to 30.7% vs. 35.5% .
- Strategic alternatives/funding remain unresolved; management is exploring options (sale of assets, funding, R&D support) with Needham, but “nothing specific to talk about” yet—cash used in operations for FY2025 was about $12.9M .
- Prior-quarter gross margin was pressured by severance/mix; although Q4 recovered, the FY2025 GM fell to 49.4% vs. 54.3% in FY2024 on mix and fixed cost absorption, underscoring sensitivity to product mix .
Financial Results
Quarterly Performance vs Prior Periods
Year-over-Year (Q4 FY2024 vs Q4 FY2025)
Actuals vs Estimates
Values retrieved from S&P Global.*
Customer/Product Mix (Shipments/Concentration)
KPIs
Guidance Changes
Earnings Call Themes & Trends
Management Commentary
- “Our fourth quarter revenue increased 14% year-over-year and 9% sequentially to $5.9 million, reflecting strong demand for our legacy SRAM chips… significantly reduced net loss and lower cash burn for the quarter.” — Lee‑Lean Shu .
- “We secured an initial order for our radiation-hardened SRAM… with follow-on orders expected in fiscal 2026… [this] carries a significantly higher gross margin than our traditional SRAM chips.” — Lee‑Lean Shu .
- “We expect to receive production-ready Gemini-II chips and Leda-2 boards by the end of the first quarter of fiscal 2026… SBIR programs… remain on schedule. We delivered a server with a Leda-2 board to AFRL and will soon ship a Gemini-II card to SDA.” — Lee‑Lean Shu .
- “Adding the integration of a camera interface directly into [Plato]… creates a compact, all-in-one… engine for edge devices… [well] suited for agents requiring object recognition.” — Lee‑Lean Shu .
- “Q1 FY2026… net revenues… $5.5 million to $6.3 million, with gross margin of approximately 56% to 58%.” — Lee‑Lean Shu .
Q&A Highlights
- TAM and positioning: Management has not yet published TAM; Gemini‑II targets search/HPC at the edge (e.g., SAR on drones/satellites), while Plato targets LLM inference at the edge with low power .
- Demand mix: Early interest for Gemini‑II is mainly military/defense; both chip-level and board-level demand pathways exist; hyperscaler engagements are longer-cycle—company is focusing near term on edge deployments .
- Funding/strategic alternatives: Management is pursuing broad options (sale of assets, company funding, R&D funding, product development support); no specific outcomes disclosed yet .
- Cash/CapEx disclosure: FY2025 cash used in operating activities was about $12.9M; CapEx was ~$45k .
Estimates Context
- S&P Global consensus for Q4 FY2025 EPS and revenue was unavailable; actual revenue and EPS were reported but no consensus values were returned by SPGI (S&P Global). Values retrieved from S&P Global.*
Where estimates may need to adjust:
- Guidance implies stable revenue and gross margin into Q1 FY2026 ($5.5–$6.3M; 56–58%), underpinned by continued SRAM demand and improved mix—Street models should reflect sustained margin expansion vs. FY2025 trough and higher contribution from high-margin rad‑hard SRAM .
Key Takeaways for Investors
- Near-term catalyst: Initial radiation-hardened SRAM order with materially higher gross margins and expected follow‑on orders in FY2026—positive for margin trajectory and cash burn reduction .
- Mix-driven margin recovery: GM expanded to 56.1% on product mix and revenue growth; Q1 FY2026 GM guidance of 56–58% indicates maintenance of improvements—constructive for gross profit leverage .
- Customer concentration risk: KYEC at 29.5% of Q4 revenue underscores concentration; any timing variability (tariffs, shipment schedules) could add quarterly volatility .
- Funding overhang: Strategic alternatives remain open with no concrete outcomes; monitor liquidity and burn (FY2025 cash used ops ~$12.9M) against guidance and potential financing events .
- APU roadmap advancing: Gemini‑II/Plato milestones and SBIR-funded work de‑risk execution and broaden edge‑AI addressable markets; partner engagement is increasing .
- Earnings quality: FY2025 included non-recurring elements (sale‑leaseback gain, severance costs, SBIR offsets); Q4 results reflect cleaner trajectory with lower OpEx and margin normalization .
- Trading setup: With guidance in-line and margin resilience, stock reactions likely hinge on incremental contract disclosures (rad‑hard SRAM follow‑ons), APU partnership announcements, and funding clarity—focus on newsflow cadence and customer updates .