Google Finance
Filing
Dec. 01, 2014
2015 Q2 Data
1.Major Business.
SRAM and LLDRAM manufacturer. SRAM is high end fast response memory used mostly in server or router compare to DRAM which used in PC. Fast SRAM has a market size around $600m to $800m/year. LLDRAM is a newer RAM developed by the company and I guess its sales is still not significant.
The company was founded at 1995 by its CEO Shu, Lee-Lean who was working for Sony for 5 years before that. Later at Aug. 2009 the company actually acquired the SRAM business from Sony for $5.2m.
2.Balance sheet.
Recent price is $5.16. Shares 23.7m. Market Cap around $122m. Tangible book value $4.25. Cash $63m. No debt.
3.Credit facility.
Not important
4.Financial data by years.
5.Insider holding, options, Insider trading info, share buy back.
CEO: 2.4m shares. Including 500k options.
Currently there are 5m+ options outstanding average price around $5.2. 6m+ more options to be issued.
6.Management compensation.
Top five around $2.5m/year.
7.Employee numbers. Revenue/Employee. Compensation/Employee.
At Mar. 2014 has 138 employees. 42 in R&D, 18 in Sales, 10 in Admin, 66 in Manufacturing.
8.Industry comparison.
9.Auditor
PricewaterhouseCoopers LLP
10.Major events.
(1)The major player in SRAM is Cypress which has annual sales in SRAM around $200m to $300m. Since 2011, it filed 4 patent lawsuits against GSIT and GSIT won first two. The other two have been combined to one and is still going on. As a defence, GSIT filed antitrust lawsuit against the company and currently is still in process. This has already cost GSIT extra $20m+ legal expense and will continue in incur in near future.
11.Comments.
(1) There are two major problems of the company.
One is the Cypress litigation issue it is still unknown how much will cost in the future. One good part of this is the company if the company win the antitrust lawsuit, it should get $30m+ in return. Personally I feel it has a higher chance to win than lost. Anyway, this is temporary issue and won't drive it out of business.
The other one is the revenue is down for the past 3 years as well. The major contributor of this is Cisco which purchased more from the company after Sony acquisition. Toped at 2011 and then down since. It might or might not related to the litigation. If remove Cisco sales from revenue, they are around $45m $50m for the past several years and actually increased a little bit at 2014.
(2) In cash flow, there are others item $3m to $5m which is combine of stock options expense($2m/year), write down of inventory($700k to $2m), and amortization of bond premium($1.2m to $800k). Options expense is around $2m/years.
(3)Go forward, estimate $50m revenue, 45% gross margin, it would generate $22.5m gross profit. Assume it would be $25m. Assume $10m in SG&A. $10m in R&D. it would leave $5m in EBIDTA.
Use just $1m in Capx. $0.5m in tax. The real income is still just $3m to $4m. We could add some of the cash flow item and the interest income($0.5m/year) for about $5m. It is still less than $10m.
(4) The stock is priced that either the LLDRAM will pick up the revenue, or they could win the antitrust lawsuit. With the heavy payment of management and big options pay out, it is not really cheap.
12.Links
Aug 28, 2015
Q1 2016 Data
Current price $4.63, Market Cap: $105m.
For 2015, the company's gross profit is $25m which is close to my estimate. However, R&D is actually $12m which makes the real EBIDTA is just $4m.
On positive side: first, it solved the litigation with Cypress, which means eventually its litigation expense would go down. Secondly at Q1 2016, the company's revenue and gross profit both are up for $1m.
Going forward, it might be able to make $28m in annual gross profit. Using $10m SG&A and $12m in R&D. It would generate around $6m in EBIDTA. $4.5m in real income. Using 9 times P/E. It is cheap at $40m.
There is a mistake in previous study that the large cash balance should be counted in valuation. If using 20% discount rate with 2 years time frame, the current value is $40m.
On a cash flow basis, it should be able to add around $3m to $4m to real income. Which added another $30m to $40m value.
When added discounted cash to the real earning power, its real worth is actually over $100m.
Risks:
1. Litigation expense might last very long time, this will the company operate in loss for quite a while, however, currently the FCF-WC is neutral.
2. RAM business might go down again.
May 12, 2016
Q4 2016 Data
Current price $3.85, Market Cap: $90m.
Full 2016 revenue $52.7m, Gross profit $26.7m. Legal expense: $6.7m. SG&A including legal is $17.6m. R&D is $12m. Cash+investment: $66m
Finally the legal expense down to $200k at Q4 2016. The company made a small loss and a positive cash flow at Q4 I guess.
The company acquired an Israel company MikaMonu at Nov. 2015 who is specialized on associative computing which enable memory computing to solve bus bottleneck between cpu and memory. It does make sense to me although it might not work at the end. GSIT paid $5m + $2.5m in future. MikaMonu will not have any income contribution until 2019. The company has less than 10 employee and I estimate $2m extra expense for GSIT.
In future, the company will have $24m to $25m running rate of SG&A+ R&D. It profit should be able to maintain at above $25m level. So it would be just profitable. Cash side, it might be little more given $3m level in the others non-cash items.
Overall, with litigation expense is finally down. The company is more likely to make a profit in the future. Currently stock price is just about 85% of book value which is quite cheap. However, whether the new acquisition will work remains unknown and takes a long time to see the result(at 2019).
PricewaterhouseCoopers LLP
10.Major events.
(1)The major player in SRAM is Cypress which has annual sales in SRAM around $200m to $300m. Since 2011, it filed 4 patent lawsuits against GSIT and GSIT won first two. The other two have been combined to one and is still going on. As a defence, GSIT filed antitrust lawsuit against the company and currently is still in process. This has already cost GSIT extra $20m+ legal expense and will continue in incur in near future.
11.Comments.
(1) There are two major problems of the company.
One is the Cypress litigation issue it is still unknown how much will cost in the future. One good part of this is the company if the company win the antitrust lawsuit, it should get $30m+ in return. Personally I feel it has a higher chance to win than lost. Anyway, this is temporary issue and won't drive it out of business.
The other one is the revenue is down for the past 3 years as well. The major contributor of this is Cisco which purchased more from the company after Sony acquisition. Toped at 2011 and then down since. It might or might not related to the litigation. If remove Cisco sales from revenue, they are around $45m $50m for the past several years and actually increased a little bit at 2014.
(2) In cash flow, there are others item $3m to $5m which is combine of stock options expense($2m/year), write down of inventory($700k to $2m), and amortization of bond premium($1.2m to $800k). Options expense is around $2m/years.
(3)Go forward, estimate $50m revenue, 45% gross margin, it would generate $22.5m gross profit. Assume it would be $25m. Assume $10m in SG&A. $10m in R&D. it would leave $5m in EBIDTA.
Use just $1m in Capx. $0.5m in tax. The real income is still just $3m to $4m. We could add some of the cash flow item and the interest income($0.5m/year) for about $5m. It is still less than $10m.
(4) The stock is priced that either the LLDRAM will pick up the revenue, or they could win the antitrust lawsuit. With the heavy payment of management and big options pay out, it is not really cheap.
12.Links
Aug 28, 2015
Q1 2016 Data
Current price $4.63, Market Cap: $105m.
For 2015, the company's gross profit is $25m which is close to my estimate. However, R&D is actually $12m which makes the real EBIDTA is just $4m.
On positive side: first, it solved the litigation with Cypress, which means eventually its litigation expense would go down. Secondly at Q1 2016, the company's revenue and gross profit both are up for $1m.
Going forward, it might be able to make $28m in annual gross profit. Using $10m SG&A and $12m in R&D. It would generate around $6m in EBIDTA. $4.5m in real income. Using 9 times P/E. It is cheap at $40m.
There is a mistake in previous study that the large cash balance should be counted in valuation. If using 20% discount rate with 2 years time frame, the current value is $40m.
On a cash flow basis, it should be able to add around $3m to $4m to real income. Which added another $30m to $40m value.
When added discounted cash to the real earning power, its real worth is actually over $100m.
Risks:
1. Litigation expense might last very long time, this will the company operate in loss for quite a while, however, currently the FCF-WC is neutral.
2. RAM business might go down again.
May 12, 2016
Q4 2016 Data
Current price $3.85, Market Cap: $90m.
Full 2016 revenue $52.7m, Gross profit $26.7m. Legal expense: $6.7m. SG&A including legal is $17.6m. R&D is $12m. Cash+investment: $66m
Finally the legal expense down to $200k at Q4 2016. The company made a small loss and a positive cash flow at Q4 I guess.
The company acquired an Israel company MikaMonu at Nov. 2015 who is specialized on associative computing which enable memory computing to solve bus bottleneck between cpu and memory. It does make sense to me although it might not work at the end. GSIT paid $5m + $2.5m in future. MikaMonu will not have any income contribution until 2019. The company has less than 10 employee and I estimate $2m extra expense for GSIT.
In future, the company will have $24m to $25m running rate of SG&A+ R&D. It profit should be able to maintain at above $25m level. So it would be just profitable. Cash side, it might be little more given $3m level in the others non-cash items.
Overall, with litigation expense is finally down. The company is more likely to make a profit in the future. Currently stock price is just about 85% of book value which is quite cheap. However, whether the new acquisition will work remains unknown and takes a long time to see the result(at 2019).
