GSI Technology Q3 Earnings Call Highlights

GSI Technology (NASDAQ:GSIT) reported fiscal third-quarter 2026 results and provided updates on its Gemini-II and Plato programs during a conference call held January 29, 2026. Management highlighted year-over-year revenue growth, continued demand from key customers, and progress on edge AI initiatives, including a government-backed proof-of-concept with Israel-based AI company G2 Tech.

Quarterly results and outlook

Chief Financial Officer Douglas Schirle said GSI recorded net revenues of $6.1 million for the third quarter of fiscal 2026, compared with $5.4 million in the year-ago quarter and $6.4 million in the prior quarter. Chairman, President, and CEO Lee-Lean Shu added that revenue increased 12% year-over-year and was up 28.5% on a fiscal year-to-date basis, and that demand “remains solid.”

Gross margin was 52.7%, down from 54.0% a year earlier and 54.8% in the prior quarter. Schirle attributed the margin decline primarily to product mix.

Operating expenses rose to $10.1 million, compared with $7.0 million in the prior-year quarter and $6.7 million in the preceding quarter. Research and development expenses increased to $7.5 million from $4.0 million a year earlier and $3.8 million in the prior quarter, driven primarily by the purchase of IP related to Plato development and associated consulting expenses. Selling, general, and administrative expenses were $2.6 million, compared with $3.0 million in both the year-ago quarter and the prior quarter.

Operating loss for the quarter was $6.9 million, compared with an operating loss of $4.1 million a year earlier and $3.2 million in the prior quarter. Net loss for fiscal Q3 was $3.0 million, or $0.09 per diluted share, compared with a net loss of $3.2 million, or $0.11 per diluted share, in fiscal Q2 and $4.0 million, or $0.16 per diluted share, in the year-ago quarter.

Schirle noted the quarter’s net loss included interest and other income of $3.6 million, reflecting a non-cash accounting adjustment of $6.2 million related to changes in the fair value of pre-funded warrants, offset by $2.8 million of issuance costs tied to the company’s registered direct offering. The company also recorded a $251,000 tax benefit.

For fiscal Q4, GSI guided to net revenues of $5.7 million to $6.5 million and gross margin of approximately 54% to 56%.

Cash position strengthened after registered direct offering

Beginning this quarter, GSI said it is expanding cash-related disclosures in its earnings releases to provide more detail on cash use and generation. Schirle reported cash and cash equivalents of $70.7 million as of December 31, 2025, up from $25.3 million at September 30, 2025, and $13.4 million at March 31, 2025.

For the quarter ended December 31, 2025, net cash used in operating activities was $7.9 million, net cash used in investing activities was $296,000, and net cash provided by financing activities was $53.5 million. Schirle said the increase in cash primarily reflected $46.9 million in net proceeds from the company’s October 22, 2025, registered direct offering. He added that operating cash use included spending for the development and commercialization of Gemini-II and Plato.

Working capital was $71.7 million at December 31, 2025, versus $16.4 million at March 31, 2025, while stockholders’ equity increased to $83.6 million from $28.2 million at the fiscal year ended March 31, 2025.

Gemini-II benchmarking and edge AI positioning

Management emphasized progress on Gemini-II, including a “time-to-first token” (TTFT) benchmark discussed in the day’s press release. Shu said the company reported three-second TTFT performance for a multimodal large language model with text and video input, consuming approximately 30 watts of system power. He said that compared to third-party testing of competitive platforms, Gemini-II delivered up to three times faster first token at lower power on the same workload, and that the results support edge use cases requiring low power and low latency.

Vice President of Sales Didier Lasserre explained why TTFT matters for applications such as drones and unmanned systems, describing TTFT as the time it takes a system to produce the first response. He said a useful TTFT threshold for video surveillance is three seconds, because the system can sample the video image every three seconds; he contrasted that with a 10-second TTFT, which he said could be too slow and cause surveillance to miss events. Lasserre said the company plans to continue improving TTFT over the next five months in preparation for an upcoming Sentinel demonstration.

Lasserre also framed the company’s focus around edge and “physical AI” use cases, stating that for edge inference, performance-per-watt and responsiveness matter more than peak training metrics. He said GSI is not trying to compete in training and data centers and is instead positioning its compute-and-memory APU architecture as a “strong option” for fast, low-power edge AI applications.

Sentinel proof-of-concept with G2 Tech and government funding

Shu said GSI finalized an agreement with G2 Tech for a proof-of-concept tied to a program called Sentinel, which he described as autonomous perimeter security using drones and cameras. Lasserre said the company expects to receive more than $1 million in government funding for the Sentinel POC and that GSI will record the funds as an offset to R&D expense. He said the company plans to use the funding to complete key software milestones, including software development for Gemma 3 12B on Gemini-II, ahead of a planned demonstration to government agencies later in the year. Lasserre added that G2 Tech is receiving additional funds to develop the drone platform for the demonstration.

Lasserre said G2 Tech conducted a competitive evaluation and selected GSI based on Gemini-II’s performance, including what he described as the lowest TTFT at 30 watts. He said a successful government evaluation could lead to a potential Gemini-II design win with G2 Tech and could support additional opportunities with other drone and unmanned system customers beyond the POC sponsors.

During Q&A, management said defense and military markets have been early adopters of the company’s technology, citing prior SBIR wins and ongoing evaluations. Shu also said the work being done for the G2 Tech POC could be leveraged for other markets, including smart cities and robotics.

Customer concentration and product mix details

Lasserre provided a customer and sales breakdown for the quarter. Sales to KYEC were $1.1 million, or 17.9% of net revenues, compared with $1.2 million (22.7%) in the year-ago quarter and $802,000 (12.5%) in the prior quarter. Sales to Nokia were $675,000, or 11.1% of net revenues, compared with $239,000 (4.4%) a year earlier and $200,000 (3.1%) in the prior quarter. Sales to Cadence Design Systems were $233,000, or 3.8% of net revenues, compared with $971,000 (17.9%) in the year-ago quarter and $1.4 million (21.6%) in the prior quarter.

He added that military defense sales represented 28.5% of third-quarter shipments, compared with 30.0% in the year-ago quarter and 28.9% in the prior quarter. SigmaQuad sales represented 41.7% of third-quarter shipments, compared with 39.1% a year earlier and 50.1% in the prior quarter.

Looking ahead, Shu said the company expects strong sales from its largest customers in the first half of calendar 2026. He also said that following the company’s fall 2025 financing, GSI advanced its APU roadmap, began planning hardware development after purchasing required IP, and added contract engineers to support the hardware design team, with the company targeting a Plato tape-out in early 2027.

About GSI Technology (NASDAQ:GSIT)

GSI Technology, Inc is a fabless semiconductor company specializing in the design and development of high-performance memory products. Headquartered in Sunnyvale, California, the company was founded in 1995 and has focused its efforts on content addressable memory (CAM) and high-speed SRAM (static random-access memory) solutions. As a publicly traded company listed on NASDAQ under the ticker GSIT, GSI Technology leverages advanced architectures to meet demanding data-processing requirements.

The company’s core product portfolio includes ternary CAM (TCAM) devices, binary CAM (BCAM) devices and high-speed synchronous SRAM.