GSIT - GSI Technology Offers APU Cloud Access Waits For Big SRAM Order
Summary
- Major SRAM order is expected to come in the new quarter, being delayed from the expected delivery last quarter.
- The cloud access at Searchium.ai to the APU is going to become available this quarter, which will start more open-source work with the APU.
- Python compiler is now out for the APU, so developer work from outside the company is finally able to start on the APU and its APIs and libraries.
- Gemini-II is expected to come out in about a year's time from today, and they continue to try to create interest for the APU in defense markets.
- No more China manufacturing plans to limit COVID-19 risks, cash burn should be around $2-3 million per quarter and cash covers it for 8 quarters, and GSIT remains a buy.
Let's talk about one of our VC style exposures in the portfolio, GSI Technology ( NASDAQ: GSIT ). Cash burn continues, and sales growth is still moderate, waiting on a massive prototype order in their SRAM "legacy" business which could double annual sales when it comes through. The APU continues progress on the software side with the release of Python support and the soon to be released Searchium.ai cloud access to the APU. The company is limiting China COVID-19 risk, and there's plenty of cash to cover the company for about 8 quarters in our estimation. It remains a buy, although we see the possibility of further dilution as the next gen Gemini will be done earlies in a year's time.
Key Developments in Q1 for GSIT
Gross profit is up 12% on modest topline growth but this is being absorbed by increases in R&D, so the operating losses are only decreasing by about 6%. The SRAM business continues to create revenue growth, but the major prototype order hasn't kicked in yet which is supposed to supercharge revenues by a factor of 2x on the annual basis. While expected this quarter, and while the parts and product is ready, customers asked for some more specs, and the order is expected to be completed in the quarter after this reported one. This should help bolster cash balances, or at least help move R&D along.
This prototype order could potentially be the same revenue magnitude of all of our radiation hardened orders in the last fiscal year. As a reminder, Rad-Hard is a high-growth profit product for us. And if a few of these prototypes move into production, it could add incrementally to fiscal years 2023 and 2024 revenues.
From a development perspective, a major thing has happened which is now there is a Python compiler for the APU. Moreover, the APU will be delivered by cloud on Searchium.ai as of the coming quarter, and will require institutes and others who want access to subscribe. While the cash flow might not be very significant, it at least means that developers outside of the company can get their hands on it and the open-source process of library building and other development can start on the basis of this APU. They will be adding some of the high level applications useful in ML in the coming quarter as well. Hopefully this will create some access and usability to drum up interest from institutions who might bring forwards some useful applications themselves to create a market for the product. While it's a bit early to call demand-side economics a driver for GSIT, it is something we're thinking about further down the line.
We've also got more info on the Gemini-II developments, which is that it will be about a year before that is done. The designs are finished, but it hasn't been produced yet. For now, they'll have to drum up interest on the basis of the Gemini-I competition results, where it has performed well in some important applications that has apparently created some interest in the defense industry.
The other news is that the company is no longer China-exposed in its manufacturing. Lockdowns in China on the basis of COVID-19 will no longer matter.
Financials and Conclusions
From a cash burn perspective , the maturity of some instruments offset some of the cash burn. They reduce cash burn to a pretty low level, midpoint of $2.5 million per quarter, for the next 4 quarters after this one. We expect that starting in Q2 2023 cash burn will return to higher levels of around $5 million per quarter, and therefore from today's cash balances and ignoring the potential prototype order windfall which will provide lots of revenue and mix improvements for incremental profit, we should have 8 quarters starting from next of cash to cover the cash burn. Not bad at all - there's a chance the fallen price will not matter for the long-term upside for equity depending on dilution timelines.
Overall, the dilution risk is clearly there, and the company has delayed things a lot in the past. It has even done so now, although less egregiously, and this can obviously make dilution more probable. Nonetheless, we see the market being big enough, and some key steps being taken now, where we are alright with whatever happens, also because we realize how speculative this position is and have kept our exposure minimal. A clear buy, and we continue to own this.
For further details see:
GSI Technology Offers APU Cloud Access, Waits For Big SRAM Order