April 27, 2012 - 11:20am EST by
2012 2013
Price: 4.23 EPS $0.19 $0.24
Shares Out. (in M): 28 P/E 22.0x 17.0x
Market Cap (in $M): 118 P/FCF 10.0x 8.0x
Net Debt (in $M): -89 EBIT 15 19
TEV (in $M): 29 TEV/EBIT 2.0x 1.6x

Sign up for free guest access to view investment idea with a 45 days delay.

  • Competitor Exit
  • Litigation
  • New Product Launch
  • Semiconductor


Recommendation: Long GSIT

Quick Rationale:

  • Stock price: $4.22
  • $27mm EV ($89mm net cash)
  • Trading at 1.5x ltm adjusted EBIT (ex legal expenses)
  • Catalysts
    • Cypress ITC case wrapped up within next 6 months
    • Samsung exiting market = jump ball on 25% of market share
    • LLDRAM product - could add $40mm rev/year within 18 months
  • Bull case valuation = $17.00
  • Bear case valuation = $5.00


  • Description: GSIT makes SRAM chips for use primarily in networking equipment and military applications.
    • SRAM is different than DRAM. 
    • SRAM is a high mix, low volume product category.  DRAM is the opposite…low mix, high volume.  So SRAM doesn’t have the huge production builds and wild price swings that DRAM suffers from. 
    • These are not consumer products.  The ASP curves are very predictable.  The legacy stuff has very stable pricing.
    • Product cycles are long, 5-6+ years.  GSIT started getting Cisco business in 1998.  We are still shipping parts from designs 11 years ago.
    • The majority is dual sourced.  Some stuff on the bleeding edge is sole sourced.  Take Cisco, GSIT has 120 parts with Cisco, there are 10 that are sole sourced.  7 where no one else could even do it.
  • We estimate TAM for SRAM market to be $800mm+
    • Market share
      • 45% CY
      • 25% Samsung
      • 10% GSIT
      • 10% ISSI
      • ~5% REC
      • ~5% IDT
  • GSIT has ~10% market share. 
    • The #3 player behind Cypress Semi and Samsung. 
  • Samsung has announced they are exiting the business.  (More on that later).



  • Market Cap & EV:
    • $115mm market cap
    • $88mm cash & equiv
    • $27mm EV
  • Financials
    • Income Statement (ltm)
      • $85.6mm revenue
      • 43% Gross Margin
      • 13% EBIT margin = $11.3mm EBIT
      • Legal Expenses = $5.68mm
      • 19.8% adj EBIT Margin = 16.9mm EBIT
      • 33 quarters of profitability = 8 years of consistent profits.
    • Balance Sheet
      • Total Assets = $142mm
      • Total Liabl = $12.8mm
      • Total Equity = $129mm
      • Goodwill & Intangibles = $1.0mm
      • Tangible book = $128mm
  • Valuation
    • 2.5x EV/EBIT(ltm).  (1.5x when you adjust for legal expenses)
    • 0.90x book value
  • Why Down?
    • Cypress Semiconductor launched civil lawsuit and an ITC Complaint against GSIT for patent infringement.  GSIT has been shipping these parts for 6+ years prior to this case coming up.  Should CY win, GSIT will be prohibited from importing related products to the US. The civil case has been put on hold until the ITC case is resolved.  Legal expenses have weighed profits - $3mm spent last quarter. (nearly $6mm over last 9 months)
    • Arrow – a distributor – announced they would stop carrying GSIT and go exclusive to CY in the US.  As quid pro quo, CY dropped Arrow’s name from the ITC complaint.
      • Back story: GSIT is not happy they lost Arrow, but the story is that CY said they were going to fire a distributor.  Arrow and New Horizons were named in the lawsuit.  Arrow is very anti lawsuits, they went to CY and said they’d drop GSIT if they drop the case because they thought that put them in better position.  
      • Mitigation of this issue – GSIT says they are happy with Avnet (another distributor). And Avnet is happy they are now captive.  They’re a good logistical partner.  They’re excited to double their revenue with GSIT.  They know now if they put in the effort they will get the order…since sometimes the effort can result in a different distributor getting the order.  All GSIT reps do the design work to win the design, but the customer really uses a contract manufacturer…so it is easier to sell and distribute if you have an in-house store.
    • Announced they expected to post a loss in Q4 2012 (which includes $3mm of legal expenses).  First quarterly loss in 34+ quarters.  Reason: semis in general expected to be soft in March quarter.  Uncertainty in customers’ purchases as the ITC trial neared (in March).  They don’t want to disappoint, so they guided conservatively.  Military was soft due to slow down in defense spending.
      • CY reported last week.  They discussed soft results for SRAM business in the March quarter, however guided to improvement the rest of the year.


Background on case & additional catalysts:

  • Cypress ITC Complaint
    • Civil case was filed March 30th, 2011
      • Alleges infringement on 5 patents.
    • ITC Complaint was initiated in June 2011
      • Alleges infringement on 3 of the 5 patents + one more ptatnt. 
      • Alleges infringement by 3 of the distributors and 11 GSIT customers
    • Civil case is on hold until ITC ruling is complete
    • GSI filed anti-trust lawsuit against CY in July 2011.  CY requested dismissal.  Waiting on a ruling from the court.
    • In Nov 2011, the PTO re-examined some of the related CY patents ('805 & '477).  The report rejected a number of claims of both patents, including all claims that relate to CY infringement claims against GSI.
    • CY has released 2 distributors and 5 customers (including Alcatel-Lucent & Ericsson) from the case.  We do not know what the agreements between CY and those companies entail.
    • The ITC trial took place in March.  Initial ruling is expected in July 2012 with final ruling in Nov 2012.   
  • If GSIT wins
    • The overhang is removed.  Immediate $1.00-3.00 pick-up in share price
    • Bull case becomes much more probable.
  • If GSIT loses
    • Complaint is about the SigmaQuad family of products. 
      • 30% of total revenue.  US is 2/3 of that.
    • So company could lose 20% of revenues right away.
    • Mitigating factors:
      • Some business sold to military, they probably don’t want to sole source CY…so the military (if for national security) might find a way around to be able to still import.
      • Will telecom equipment customers really want to sole source CY?  Seems possible they might try to get CY and GSIT to come to the table to license the disputed technology. 
      • If an end customer & product is not specifically named in the complaint, the customer could manufacture the product overseas with GSIT products included, and still be able to import them into the US without a problem. 
    • So it could be far less negative than it would first seem.
    • Further, company could invest in R&D to do redesign around CY patents.
  • Additional Catalysts
    • Samsung exiting the market.  They announced it a few months ago.  Last orders will be June/July, last shipments Q4 2012 time frame.
      • Samsung was once 40% of the market (today about 25%).  Still a lot of market share is up for grabs.
      • GSIT should win a fair portion of that business.  We estimate that CY is already 45% of the market.  Doubtful people want them to get a lot bigger…so GSIT will gain more share relative to CY. 
        • If they pick up 30% of the Samsung market share, could be an incremental $60mm in revenue.  ($800 * .25 * .30 = $60mm).  This woudl be nearly 70% growth to GSIT.
    • LLDRAM
      • MU is the only player selling “low latency DRAM.”  Customers are clamoring for a second source.  It’s difficult to do.  LLDRAM is one of the highest margin product lines for MU.
        • LLDRAM goes into telcom equipment (routers, servers, etc…).  Shares similar characteristics to SRAM market.
      • GSIT recently qualified a part and will start shipping it in 9 months (need to win designs, etc…)
      • Their goal is for it to be a $10mm/q revenue business.  This alone would be 50% growth.
      • The TAM for LLDRAM is probably $200-300mm.  It could grow now that there is a second source, people will be more comfortable designing it in.  Pricing will probably fall, but volumes will more than make up.


Quick valuation:

  • Bull case scenario:
    • FY2 -- looking out two fiscal years.
      • $175mm revenue
        • $40mm from LLDRAM
        • $45mm increase in SRAM.  From market share pick up & growth.  50% growth in SRAM from market share increases over 2 years.   (This is slightly lower than what the company has expected.)
      • 25% EBIT margin/ 30% EBITDA
      • $52mm EBITDA
    • 9x EBITDA = $472mm + $88mm cash today + $30mm cash flow = $590mm / 29mm shares = $20.00
    • Discounted back to today = $17.00
  • Bear case scenario:
    • Lose CY case.  Lose all of affected revenue.  Potential = $17mm = 20% of revenues
    • Grow from that lower base = $75mm
      • $20mm from LLDRAM (lower than expectations)
      • $8mm increase in SRAM.  10% growth from market & market share over 2 years.
      • Total revs = $103mm
    • EBIT margin 18% = $18mm // $23mm EBITDA
    • Pay $20mm in damages to CY related to civil case.  (This is a wild guess).
    • 4x EBITDA = $92mm + 69mm cash + 15mm cash flow = $176mm / 29mm shares = $6.00
    • Discount back to today = $5.00
      • * Note: The stock would likely fall further below this price if they lose the case.  I would view that decline to be temporary in nature, assuming these assumptions are accurate.
Additional Details:
  • Management Ownership
    • CEO owns 2.68mm shares.  Worth today about $11.65mm
    • Insiders as a group together own 24% of the stock = 6.5mm shares = $28mm 
  • Samsung
    • They are exiting because this is a small niche market and Samsung has bigger fish to fry.
    • When they leave there will just be CY, REC, ISSI, and GSIT.  ISSI is really in auto, they’re not really a factor.  REC has a few hundred customers but mostly just the large network guys and some people in Japan. 
    • Samsung once had 40%+ of the market but they probably only 25% now.  There are only a couple of sole source designs.  (Example: Juniper.  They are dual sourced CY and Samsung.  When Samsung leaves, it will just be CY…so GSIT will push to get in there.)
  • Customer Concentration Risk
    • Cisco is 40% of revenues.  This is always a risk, the company is levered to networking equipment sales.  Unlikely that Cisco will stop buying from them.
      • GSIT is in many business units.  The 3 most important are CAT 4000, CAT 6000, CRS1 and CRS3.  I think they are GSBU and ISBU.  (The GSBU is the only one with the CSRAM)
    • This risk should be reduced overtime as they pick up additional market share at other customers.
  • Use of cash
    • Company has a buyback in place for $11.3mm.  It is through a 10b-51 plan which allows the company to buy stock during normal blackout periods.
      • The company purchased $4.5mm of common stock at $4.78 in the nine months ended December 31, 2011.
      • That pretty much completed the existing program.  On Jan 25th, 2012, they authorized a new buyback in the amount of $10mm.
    • Potential for acquisitions
      • Company has said, "we will look to add one more product that is still within our core competency.  We want something with higher IP, higher margin, but serve the same customer.  We’d look to a private company that is too small for a customer to consider their technology but would consider it if we owned it. We want to move beyond being perceived as a memory company and become more of a telecommunication service provider. If the right deal came along next quarter, we’d do it.  We’re not worried about the distraction.  We’ve got the LLDRAM and Samsung situations pretty well handled now."
  • Competitive risks on the periphery
    • Keep your eye on MoSys (MOSY).  They claim to be developing products which would leapfrog the SRAM market.  I think the odds are quite low, but important to keep this in mind


  • Cypress ITC case wrapped up within next 6 months
  • Samsung exiting market = jump ball on 25% of market share
  • LLDRAM product - could add $40mm rev/year within 18 months
    show   sort by    
      Back to top