What is the timeline for formalizing system integrator relationships, and how mature does the SDK need to be before integrators can effectively use Gemini-II?

Didier Lasserre explained that the timeline depends on the market and integrator. GSI initially performs the work themselves in new markets to prove the APU solution's strength. Once proven, it's easier to attract integrators. The AI-assisted SDK is critical for integrators to address diverse customer needs with slight customizations, and its anticipated fall release is key. “And for them to be able to address all the different customers with their slight unique customization that SDK will be important or the AI SDK. So that's why we're working hard on that. And as we discussed, we anticipate having it in fall of this year.”

What are the implications of the radiation hardening test results for the aerospace and defense sector, and could it open new revenue streams?

Didier Lasserre affirmed that it will absolutely increase revenue opportunities, as there's a growing need for AI in space and GPUs historically perform poorly under radiation. He highlighted Gemini-II's low power advantage, which is crucial for space applications. The successful radiation testing, showing no single event latches, makes the standard chip suitable for harsh environments and has already led to discussions with SBIR partners. “Absolutely. It will absolutely increase the opportunity for revenue. So if you look at what's happening now, and I'm sure you've read plenty, they're trying to put more and more intelligence and AI into space itself. And historically, GPUs have not fared well under radiation testing and environments.”

GSIT

GSI Technology (GSIT) Q1 FY27 earnings call: guidance and Q&A

Earnings call· Jun 2026 (Q1 FY27)· Information Technology· Jul 30, 2026

Source: earnings call transcript

Executive summary

GSI Technology Q1 FY27 — Advancing AI Platform with Gemini-II POCs and PLAY-DOH Development

GSI Technology is progressing its transition to an AI platform company, with Q1 FY27 marked by disciplined execution on Gemini-II proof-of-concept projects and the development of the next-generation PLAY-DOH processor. The company is focused on securing design wins and expanding into adjacent Edge AI applications, supported by an upcoming AI-assisted SDK. Despite increased operating expenses due to R&D investments, GSI maintains a strengthened cash position to fund its strategic priorities.

Highlights

Q1 FY27 revenue of $6.3 million was at the midpoint of guidance and stable sequentially.
Cash and cash equivalents increased to $77 million as of June 30, 2026, up from $67.2 million at fiscal year-end.
Successfully completed all deliverables for Phase I of the Sentinel POC, with a time-to-first-token (TTFT) of 2.4 seconds.
Completed radiation testing on a standard Gemini-II device, showing operational capability in harsh space environments with no single event latches (SEL).
AI-assisted SDK development significantly reduced application development time from one year to less than one month in one case, with alpha release targeted for fall.

Concerns

Gross margin for Q1 FY27 was 53.4%, below the guided range of 56%.
Operating expenses rose to $8.8 million, up from $5.8 million in the prior year, primarily due to Gemini-II software and PLAY-DOH chip development.
Operating loss increased to $5.4 million in Q1 FY27, compared to $2.2 million in the prior year quarter.
Anticipate using approximately $4 million of cash per quarter, or $16 million annually, for development.

Guidance & targets

Category	Target	Timeline	Confidence
Net revenues	$5.7 million to $6.5 million high materiality	Q2 · FY27	High
Gross margin	approximately 53% to 55% medium materiality	Q2 · FY27	High
Annual cash usage	approximately $16 million annually medium materiality	Annual	Medium
Quarterly cash usage	approximately $4 million per quarter medium materiality	Quarterly	Medium
PLAY-DOH tape-out	early calendar 2027 high materiality	CY27	High
AI-assisted SDK alpha release	this fall medium materiality	Fall 2026	High
Smart City Phase I deployment	complete in November medium materiality	November 2026	High

Product announcements

Product	Type	Timeline
Gemini-II production chip	milestone	ready for customers
AI-assisted SDK	launch	alpha release targeted for this fall
PLAY-DOH processor	roadmap	tape out in early calendar 2027

Deals & partnerships

G2 Tech and participating defense agencies Collaboration on Sentinel POC

GSI successfully completed all deliverables for Phase I of the Sentinel POC. A laboratory test was conducted in collaboration with G2 Tech and defense agencies. G2 Tech is leading the field demonstration.

Shinshu County, Taiwan Smart City project Phase I deployment

GSI continues to meet Phase I deliverables, shipping Gemini-II chips, servers, and software for 20 cameras. Phase I deployment is on track for November. Phase II would expand to 80 cameras and add audio.

U.S. Army SBIR Phase II program

Launched U.S. Army Phase II program with a formal kickoff meeting. Preliminary work has begun on developing a ruggedized Gemini-II system for Army applications.

Space Development Agency SBIR program for radiation testing

Successfully completed radiation testing on a standard commercial Gemini-II device, showing it remained operational under harsh space environment radiation levels with no single event latches (SEL).

Risks & headwinds

Gross margin below guidance Q1 FY27

53.4%, below guided range of 56%

Mitigation:Product mix can fluctuate based on timing and composition of customer shipments; sequential improvement of 100 bps noted.

Increased operating expenses Q1 FY27

$8.8 million, up from $5.8 million in prior year period

Mitigation:Primarily due to necessary investments in Gemini-II software and Plato chip development; partially offset by SBIR funding, though lower this quarter.

Higher software development costs due to FX Q1 FY27

Higher software development cost from our Israel-based team

Mitigation:Due to the strengthening of the Israeli shekel relative to the U.S. dollar; noted as variable in nature.

Cash usage for development Ongoing

approximately $4 million of cash per quarter or about $16 million annually

Mitigation:Existing cash resources of $77 million provide sufficient runway; maintaining disciplined capital allocation.

Customer demand variability Ongoing

customer demand can vary from quarter-to-quarter, which can impact both our revenue and gross margin

Mitigation:Expect continued shipments to customers supporting AI chip design and semiconductor manufacturing, but order patterns may continue to vary.

What to watch in Q2 FY27

Q2 FY27 Revenue

Q2 FY27

Current $6.3 million (Q1 FY27)

Target Within $5.7 million to $6.5 million

Why it matters

Revenue performance indicates the stability of SRAM business and early signs of APU adoption, crucial for overall financial health.

For the upcoming second quarter, we expect net revenues in the range of $5.7 million to $6.5 million

Find out whether GSIT delivers this. Inve re-checks your case every morning.

Q&A highlights

How transferable is the technical work from current POCs to new applications, and what does this mean for expanding the POC pipeline without increasing engineering headcount?

Didier Lasserre confirmed high transferability, citing the time-to-first-token (TTFT) parameter and vision language model features as examples. He noted that adapting software from Sentinel to Smart City was not a complete overhaul. The upcoming AI-assisted SDK will further enhance this by increasing deployable algorithms.

“Yes, what's nice is when we went from the Sentinel drone POC to the Smart City, it wasn't a complete new hall. We were able to leverage a lot of what we've had in the past. And that will be true for some of these other POCs that we're looking to engage in.”

asked by David Williams · answered by Didier Lasserre

2 min read 6 chapters

Detailed narrative

Transition to AI Platform

GSI Technology is actively transitioning from an SRAM chip company to an AI platform company, focusing on its APU technology. The Gemini-II production chip is ready for customers, targeting edge AI applications that require lower power than GPUs but higher performance than small processors. The company aims to fill a critical gap in the market with its power-efficient, high-performance APU.

Proof-of-Concept (POC) Progress

The company made significant progress on several POCs and SBIRs in Q1 FY27. The Sentinel POC successfully completed all deliverables for Phase I, demonstrating a time-to-first-token (TTFT) of 2.4 seconds in lab tests. The Smart City project in Taiwan is on track to complete its Phase I deployment in November, involving 20 cameras and potentially expanding to 80 cameras and audio in Phase II, which would generate additional hardware and software licensing revenue.

Software Development Acceleration

GSI is developing an AI-assisted SDK, expanding on its previous AI compiler vision. This SDK is expected to significantly accelerate AI model development and deployment on Gemini-II. Internal testing showed a reduction in application development time from one year to less than a month for a representative case, with an alpha release targeted for this fall.

This tool is crucial for supporting a broader range of customer applications and enabling partnerships with system integrators.

Radiation Hardening for Space Applications

Radiation testing on a standard commercial Gemini-II device successfully demonstrated operational capability under harsh space environment radiation levels, achieving the required threshold for radiation-hardened applications with no single event latches (SEL). This opens up new market opportunities in the aerospace and defense sector, particularly for AI-based processors in space vehicles where low power consumption is also critical.

Next-Generation Processor (PLAY-DOH)

Development of the next-generation PLAY-DOH processor is on track, with a tape-out anticipated in early calendar 2027. This chip is designed to further reduce power consumption, targeting 10 watts or less per device, which is expected to be a significant advantage for demanding edge AI applications.

Financial Performance and Capital Allocation

Q1 FY27 revenue was $6.3 million, aligning with guidance, but gross margin was 53.4%, slightly below the guided range. Operating expenses increased year-over-year to $8.8 million due to investments in Gemini-II software and PLAY-DOH development, leading to an operating loss of $5.4 million.

The company ended the quarter with a strengthened cash position of $77 million, up from $67.2 million, primarily due to ATM proceeds and stock option exercises. Management expects to use approximately $4 million of cash per quarter to fund its strategic priorities.

Summary of the company's earnings call; see how call summaries are prepared. Not investment advice.