March 26, 2015 - 10:19am EST by
2015 2016
Price: 3.49 EPS 0 0
Shares Out. (in M): 96 P/E 0 0
Market Cap (in $M): 333 P/FCF 0 0
Net Debt (in $M): -165 EBIT 0 0
TEV ($): 288 TEV/EBIT 0 0

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

  • Flash storage
  • Technology
  • Insider Buying
  • data services


YCOMBINATOR was nine months early on VMEM.  Today the set-up is far better - the balance sheet is strengthened, insiders are buying gobs of shares on the open market, the turnaround is largely complete, sales are hockey-sticking...and the stock price is lower than when previously posted.

This write-up focuses on the company's revenue trajectory, why the stock price is low, and what will drive the stock price higher.  We include a few comments explaining the product in layman's terms, but for a more detailed technology discussion we refer readers back to YCOMBINATOR's post from last summer.

Before jumping in, we note two quick updates since we wrote the below analysis two weeks ago.

One - over the last dozen days there have been a flurry of Form 4's filed.  Insiders have spent almost $1M of straight cash buying shares on the open market in a couple of weeks, and more Form 4's have been filed just in the last couple days.

Two - we recently came across a recurring survey of approximately 500 CIO's conducted by Enterprise Technology Research on data storage buying intentions.  In the April 2014 survey, only 6% of the CIO’s said they intended to buy VMEM’s product.  In September 2014, the number was a good bit up - to 33%.  In January 2015, 73% of CIO’s said they intended to buy VMEM’s product - #1 in the entire survey.  We think CIO’s are shifting their buy to flash and specifically to VMEM’s solution which offers the best performance vs. cost ratio in the industry (and it’s not close).  This survey is just another data point, to go with our channel checks, etc., indicating VMEM’s business is inflecting towards 100%+ year-over-year sales growth.

Now, without further ado, here’s the write-up.


Violin Memory, Inc. (NYSE: VMEM)

Violin Memory, Inc. (“Violin” or “VMEM”) is an all-flash data storage company headquartered a stone’s throw from Levi’s Stadium (home of the San Francisco 49ers) in Santa Clara, California.  VMEM’s shares are listed on the New York Stock Exchange, currently trade for $3.49 per share (a market cap of $333M), and have strong daily liquidity (1M+ shares).  We believe VMEM shares are significantly undervalued and that the company will be acquired by the end of 2016 for 3x to 10x the current share price.


Two Big Trends in Tech

Over the last half-decade, faster internet speeds have enabled two powerful tech trends.

The first is the move from on premise data storage to the cloud.  Ten years ago, a typical Fortune 500 company would have maintained suites of on-site servers to accommodate that company’s data storage needs.  Today the model is different – faster internet speeds have allowed companies of all sizes to move their data storage to dedicated or shared remote “cloud” data centers that (i) allow for meaningful economies of scale, (ii) enhance performance, (iii) reduce costs, and (iv) can be accessed from anywhere with an internet connection.

The second trend is the shift in the software provider business model to “software-as-a-service” (“SAAS”).  Traditional software was conventionally sold as a perpetual license with an up-front fee at installation; software companies then relied on periodic one-off upgrades to drive revenue.  SAAS has changed the model.  Today, the software sits in the cloud, where customers can access it at any time from any location.  Instead of an upgrade-driven model, SAAS software is updated continually so that customers always have the latest version with no need to purchase an upgrade package.  Customers instead pay a recurring monthly fee to the SAAS provider.

The SAAS model has transformed the tech world.  A wide range of business applications are now run via a SAAS model including software platforms to manage human resources, accounting, billing, expense management, customer relationship management, IT service management, and a host of other processes.  Multi-billion dollar SAAS companies such as (NYSE: CRM), Workday, Inc. (NYSE: WDAY), and ServiceNow (NYSE: NOW) are showing astonishing revenue growth as they displace the legacy on premise incumbents.

As both software and data storage move to the cloud, the demand for incremental storage capacity has been breathtaking: industry-wide, data storage demand is growing 30-40% per year.  And with a three-year depreciation cycle on data storage equipment, companies and cloud providers constantly spend to upgrade and replace equipment.  We estimate total annual spending on data storage equipment is in excess of $30B per year.


Data Storage

The dominant data storage technology today is the hard disk drive (“HDD”).  HDDs are mechanical devices – literally rotating physical disks known as “platters” – that store digital content magnetically.  The data can be written or read via a magnetic head on a movable actuator arm.  HDDs have been the industry standard for nearly forty years since they replaced tape storage in the late-1970’s and early-1980’s.

An alternate technology called flash memory was first developed in the mid-1980’s by Toshiba Corporation.  Flash memory is semiconductor-based – it stores data in transistors on a non-volatile silicon medium that can be electronically erased and re-written – and therefore is subject to Moore’s Law, which postulates that the number of transistors on a circuit doubles every eighteen months.  Put another way, the raw material used to build flash memory (“NAND”) halves in price for the same level of performance every eighteen months.

Historically, flash memory has been much more powerful than HDDs – and also much more expensive.  The cost of flash memory was prohibitive until 2008/09, when declining NAND prices (due to Moore’s Law) allowed flash memory to be competitive for the first time in certain high-performance applications.  As the cost of flash has continued to fall, some storage firms have integrated flash-based solid state drives (“SSDs”) into HDD architectures to create hybrid arrays with better performance than HDDs at only a modestly higher price.

Hybrid arrays, however, are only a temporary bridge to the future.  HDD architectures have inherent physical limitations – disks can only store so much memory and spin so fast.  Integrating SSDs into an HDD architecture may be an acceptable short-term workaround, but the bottlenecks of an HDD architecture (generally physical in nature – limiting the distance the actuator arm must move to read data, etc.) are very different from the bottlenecks of a flash memory architecture (generally electronic in nature, requiring software solutions).

The “holy grail” of the data storage industry is a ground-up all-flash integrated hardware/software architecture built to enhance performance by minimizing the bottlenecks of flash memory.  Due to Moore’s Law, it’s nearly inevitable that flash memory will eventually be cheaper than HDD.  The key is building the highest performance solution at the lowest cost – which can only be done by addressing the bottlenecks inherent to flash, not HDD.

Violin’s solutions fit that description – they are all-flash from the ground-up.  They represent the end goal the enterprise data storage industry is moving towards.  The question is: when will the industry get there?


A Ferrari at a Ford Taurus Price

The $30B data storage market can broadly be divided into two types of storage: warm storage (actively-used storage that requires higher performance as it is accessed regularly) and cold storage (non-actively-used storage such as back-up and archiving that is accessed infrequently).  Within warm storage, the market can be further broken down into tier-zero – the highest performance storage – and tiers one and two, which are also known as primary storage.  Warm storage and cold storage each represent roughly $15B in annual sales.  Of the $15B of annual warm storage sales, tier-zero represents $1B and primary storage represents $14B.

Violin’s data storage solution has historically been the “Ferrari” of the data storage industry – competing in tier-zero as the highest-performance, best-in-class product – but affordable only to customers who could pay Ferrari prices.  This has limited Violin’s total addressable market (“TAM”) to the $1B tier-zero customer group.

Until now.

On February 17, Violin launched its Flash Storage Platform (“FSP”) product.  In our view, this product represents the tipping point in the transition from HDD to flash memory.  The FSP is a fully-integrated hardware/software solution that combines Violin’s patented all-flash hardware array with Violin’s internally-developed proprietary software platform including key features such as deduplication, replication, and compression.

In short, the FSP brings all-flash into the primary storage market – increasing Violin’s TAM from $1B to $15B overnight.  Violin’s FSP solution is:

  • Far higher performance than HDDs (by 5-10x).
  • Far simpler to manage than HDDs (as it is fully-integrated into a single solution).
  • Cheaper on a cap-ex basis than HDDs.
  • Dramatically cheaper on a total-cost-of-ownership basis than HDDs (as all-flash architectures lead to substantial reductions in ongoing operating expenses for power, cooling, and floor space).

Our due diligence checks indicate that Violin is held in extremely high regard by data storage purchasers.  The hang-up has always been price.  With the launch of Violin’s FSP, that is now off the table – the FSP is not only a better solution than HDDs, it also comes at a lower cost.  And as the price of NAND (which is simply a pass-through for Violin) continues to fall, the FSP will continually become more affordable relative to HDD.

Customers can now buy the Ferrari solution at a Ford Taurus price.

We think Violin just changed the game.  All-flash has arrived.


The Violin Story

Publicly-traded data storage company life-cycles tend to follow a similar pattern.  A fast-growing data storage business with a new technology typically burns through cash as it scales its business and hires salespeople.  In need of capital to cover the cash burn, the company decides to go public.

With the company’s financials (and significant operating losses) available for all to see, investors become skittish and customers raise business viability concerns – leading to sales misses and a cratering stock price.  A new management team is brought in to clean house and control cash burn, the business begins to grow again as the market arrives for the new technology, and eventually the company is acquired at a huge revenue multiple by an incumbent firm with distribution.  Past examples include Data Domain Corporation (acquired by EMC Corporation (“EMC”) for $2.4B in 2009), 3PAR Inc. (acquired by Hewlett-Packard Company (“HP”) for $2.35B in 2010), and Isilon Systems (acquired by EMC for $2.25B in 2010).  Each was purchased for at least 10x EV/sales.

Violin went public eighteen months ago in September 2013.  VMEM shares priced at $9 per share on the IPO – and then promptly collapsed to sub-$3 per share as the company’s immense cash burn frightened investors and scared off customers.  VMEM’s CEO was fired less than three months after the IPO in December 2013.

In February 2014, Violin hired Kevin DeNuccio as its new CEO.  In the late-1990’s, Kevin ran the sales force of Cisco Systems, Inc. (NASDAQ: CSCO; “Cisco”) and built it from a few hundred salespeople into a global sales force of 6,000.  He then was hired as CEO of money-losing router company Redback Networks Inc. (“Redback”) and successfully transformed the firm into a profitable business before selling the company to Ericsson in 2007 for $2.1B.  Since then, Kevin has served on a number of boards of directors including those of major technology companies such as SanDisk Corporation (NASDAQ: SNDK) and Juniper Networks, Inc. (NYSE: JNPR) – where he was added to the Board at the behest of well-regarded hedge fund Elliot Management Corp.

Kevin came out of retirement to take the helm at VMEM because he saw an opportunity to grow a disruptive technology business using the same playbook he had followed at Cisco and Redback.  When he arrived, Kevin found a business with outstanding hardware and the raw material for an outstanding software platform, but also a mismanaged firm that had been trying to do too many things at once with an out-of-control cost structure.

Kevin quickly changed the direction of the firm.  He sold the company’s non-core PCIe product line, let go of 40% of the sales force, and restructured VMEM’s internal operations.  Kevin also brought in a new management team including a number of executives he had previously worked with at Redback and/or Cisco.  His main effort was consolidating the company’s significant software investments into a single, comprehensive platform that could be integrated with VMEM’s base hardware.  This effort eventually led to the seminal launch of the FSP product.


A Misinterpreted Quarter

We believe an investment in VMEM is timely given the market’s misunderstanding of the company’s recent Q4 FY2015 financial results (VMEM’s fiscal year-end is in January).

In Q2 and Q3 of the latest fiscal year, VMEM’s largest customer was a top-five global retailer with a recurring $4M buy.  When this customer was informed of the FSP launch, the customer decided to purchase the new FSP product – but first the retailer had to approve the FSP product via a non-industry standard 3-4 month qualification process.  As a result, this top-five global retailer did not purchase any product from Violin in Q4.

When Violin reported its Q4 FY2015 results, revenues badly missed consensus expectations.  In Q2 FY2015, Kevin’s first full quarter as CEO, VMEM reported +3% sequential sales growth to $18.6M.  In Q3 FY2015, VMEM’s top-line accelerated to +17% sequential growth and $21.7M in revenues.  With the global top-five retailer’s recurring purchase delayed by a few months, VMEM’s sales went backwards in Q4 FY2015 by -6% to $20.5M.  VMEM’s shares promptly sold off -20% to a six-month low below $4 per share.

In our view, the market misinterpreted the results.  Backing out the top-five retailer, VMEM’s sales to other customers actually increased +16% sequentially in front of a major product launch – indicating that VMEM continues to deliver on its turnaround and gain traction with customers.  Importantly, we believe the global top-five retailer will return to a $4M quarterly recurring buy of VMEM product no later than Q2 FY2016 as Violin’s solution underpins a mission-critical system for that retailer.  VMEM management guidance for another quarter of at least double-digit sequential top-line growth in Q1 FY2016 does not include any sales to that retailer.

In short, we anticipate accelerating sales growth ahead.  With the launch of the FSP product, we forecast 20%+ sequential top-line growth in Q1 FY2016 to $24-25M.  The return of the global top-five retailer in Q2 FY2016 should provide a $4M kicker on top of continued 20%+ sequential growth to other customers, resulting in Q3 FY2016 sales of $30M+.  Our view is that top-line growth will continue to accelerate from there.


The VMEM Growth Story

There are a few additional tailwinds to the VMEM growth story.

First, Violin is in the process of high-grading its indirect sales network.  Violin’s current distribution model is 30% direct and 70% indirect through channel partners, with VMEM’s indirect partners historically focused on fulfillment and delivery instead of pushing out product.  In anticipation of the FSP launch, Violin spent months re-architecting its indirect channel from “order takers” to “order getters.”  VMEM recently announced a partnership with Arrow ECS and we expect to see additional partnership announcements in coming weeks and months.  This changeover to a sales-led indirect model should accelerate VMEM’s revenue trajectory starting in H2 FY2016.

Second, the transition from tier-zero to primary storage has two significant benefits: (i) Violin’s TAM increases by 15x from $1B to $15B and (ii) Violin’s selling model shifts to a recurring revenue model that serves as a stronger platform for long-term growth.  Tier-zero sales tend to be project-focused sales that do not repeat.  Primary storage sales, however, are recurring as customers must constantly increase their data storage capacity in-line with industry-wide 30-40% annual growth in data storage needs.  As a result, every account win becomes a recurring revenue stream that will repeat in future quarters and can serve as a foundation for additional growth.  The top-five global retailer referenced earlier, for example, spends $10M+ on primary storage each quarter; over time, we think VMEM can increase its share from $4M per quarter to the entire pie.

Third, in late-2014 Violin issued a $120M convertible bond (with a $5.62 per share conversion price) that put to rest any viability concerns from customers.  The company now has nearly $165M in cash on its balance sheet.  We think the offering will help VMEM’s sales on a go-forward basis as Violin’s salespeople can redirect customer conversations away from VMEM’s now-healthy balance sheet and towards the winning FSP product suite.

Finally, in conjunction with accelerating top-line growth we also think VMEM’s gross margins have room to expand.  Today, 25% of VMEM’s sales are in the company’s “pay-as-you-grow” program where VMEM ships the customer (for example) 50 terabytes of storage capacity but the customer only pays for 25 terabytes.  As the customer’s needs increase, the customer can purchase a license on the additional 25 already-installed terabytes.  The initial sale is completed at a lower-than-normal gross margin, but the incremental sale is done at a 100% gross margin.  As upgrades kick in and as VMEM’s software technology becomes an increasingly important part of the hardware/software stack, we think gross margins can rise 1000+bps from the low-50’s to the mid-60’s.


The End Game

It is our view that Violin is first-to-market with a viable all-flash solution in primary storage.  We also think Violin has a 12-18 month head start on competitors – a lifetime in the tech world.

Only two competitors have ground-up all-flash architectures.  One is Texas Memory Systems (“TMS”), a subsidiary of International Business Machines Corporation (NYSE: IBM; “IBM”).  We think TMS is a year or more behind Violin as it does not have software features (deduplication, replication, etc.) necessary to drive down cost.  The other is Western Digital Corporation (NASDAQ: WDC; “WDC”) which picked up the bones of Skyera, a late-to-market all-flash hardware-only solution.  We think Skyera is too far behind to ever be competitive.

Other all-flash competitors, such as privately-held Pure Storage Inc. (“Pure”) and XtremIO (a subsidiary of EMC), use commodity SSDs rather than an integrated bottom-up all-flash architecture.  This allows Pure and XtremIO to be quicker to market but also puts a ceiling on performance vs. cost metrics.  We think Pure will eventually re-architect its offering to be all-flash from the ground-up, but that the process will take at least 1-2 years.  EMC, known for its best-in-class salesforce but not for its R&D capabilities, might never get there.  It is unclear whether smaller potential competitors like Solidfire and Kaminario will even make it to the other side.

It is our view that Violin’s 1-2 year head start may prove insurmountable.  As competitors come to market with their initial all-flash primary storage solutions, VMEM will already be on generation two or three.  We expect the gap to widen, not shrink, between VMEM and its peers.

As a result, we believe Violin will experience staggering sales growth over coming years.  Quarter-over-quarter sales growth of 20% is equivalent to annual sales growth of 100%+.  We would not be surprised to see annual revenue growth eventually reach a level far in excess of that figure.  As VMEM laps its first quarters without the PCIe business, we think headline numbers will for the first time turn positive in Q1 FY2016 with year-over-year sales growth of +40% and then accelerate in Q2 FY2016 to year-over-year sales growth of +70%.

That kind of growth would lead to a range of companies in the data storage industry sitting up and taking notice – bringing us to the end game.  While we think Violin has the potential to follow in the footsteps of data storage provider NetApp, Inc. (NASDAQ: NTAP) and scale from zero to $1B+ in revenues, we think the more likely outcome is that a bidding war emerges and Violin is sold in a competitive auction before year-end 2016.

We think Violin is a rare and extremely valuable asset.  We believe Violin’s hardware architecture alone, based on recent transactions, is worth $500M+.  The integrated software platform that goes with it makes the company’s solution unique.  There are numerous potential acquirers who could purchase Violin and quickly scale the business to $1-3B in sales by pushing Violin’s technology through a better distribution network, including:

  • NAND/HDD Suppliers.  A strong recent trend in tech has been commodity suppliers (of NAND, HDDs, DRAM, etc.) forward-integrating up the value chain.  We believe Violin would be an attractive target for NAND suppliers like Samsung Electronics Co. Ltd. (KRX: 005930), Toshiba Corporation (TSE: 6502), SanDisk Corporation (NASDAQ: SNDK), Micron Technology, Inc. (NASDAQ: MU), or SK Hynix Inc. (KRX: 000660).  We also believe Violin would be an ideal target for Seagate Technology PLC (NASDAQ: STX) or WDC as they look to move away from HDDs and towards flash memory.  We note that Toshiba Corporation currently owns 9% of VMEM’s shares and a joint-venture between Toshiba Corporation and SanDisk Corporation supplies VMEM with a majority of its NAND.
  •  Legacy Data Storage Companies.  Numerous legacy data storage companies may view Violin as a sensible acquisition target including IBM, EMC, NetApp, HP, Dell Inc., and Hitachi, Ltd. (TSE: 6501).
  • Enterprise Software Companies.  A final group of firms is capable of paying “stupid money” for Violin –global enterprise software companies.  Oracle Corporation’s (NYSE: ORCL, “Oracle”) purchase of Sun Microsystems Inc. (“Sun”) in 2010 proved hugely successful after Oracle integrated Sun’s storage hardware with Oracle database software into a single database solution called Exadata.  Enterprise software firms that could go down a similar path include Microsoft Corporation (NASDAQ: MSFT), SAP SE (FRA: SAP), Cisco, Google Inc. (NASDAQ: GOOG),, Inc. (NASDAQ: AMZN), or Oracle itself (should Oracle decide to update the hardware in its Exadata solution).

The math of an investment in Violin is straightforward.  VMEM currently trades at less than 3x EV/sales.  We think that over the next 12-18 months, Violin’s run-rate sales will increase by 2-3x – and attract the attention of a bevy of acquirers, one of whom will pay more than 10x EV/sales to acquire the company.  The end result is a return of anywhere from 3x to 10x the current share price.


Risks and Margin-of-Safety

We highlight several risks to the VMEM investment.

One risk is VMEM’s cash burn, currently at approximately $10M per quarter.  We expect VMEM to steadily move towards profitability – and that the company’s $165M in cash on hand will provide a more-than-adequate buffer for the company to reach profitability and/or be acquired – but should losses increase rather than diminish, our thesis could prove incorrect.  Another risk is VMEM’s sales force, which is fifty-strong but pales in comparison to the distribution power of EMC, IBM, or even Pure (which as a private company can burn significantly more cash than VMEM).  A third risk is the lumpiness in VMEM’s sales, as demonstrated in Q4 FY2015, that may make VMEM’s path less of a straight line and more of a jagged ascent.

It is critical in all investments to have a margin-of-safety.  In the VMEM investment, we think the margin-of-safety is large.  All-flash technology assets command significant value in the storage world.  In late-2013, Cisco purchased WHIPTAIL, an all-flash company with no sales, for $415M.  EMC purchased XtremIO in mid-2012 when it was pre-revenue for $430M.  Both acquisitions were hardware-only purchases.  In our view, Violin’s proven hardware architecture and integrated software platform would command a considerable premium to these purchase prices – before even taking into account Violin’s ongoing business.

Finally, we note that (i) Pure, the closest “comp” to VMEM, recently completed a private-market funding round at a $3B valuation and a multiple of more than 30x EV/sales and (ii) based on information and belief, a number of companies attempted to acquire VMEM pre-IPO for close to $1B but were rebuffed by VMEM’s previous CEO.  With a far stronger business today and a current enterprise value of sub-$300M, we think a considerable margin-of-safety exists for VMEM even in a worst-case scenario.


And a Final Observation

On October 1, 2014, an article was published in The Wall Street Journal with the headline “Meet the CEO Working for $10.15/Hour.”

The CEO in question?  Kevin DeNuccio of Violin Memory, Inc.

On a Form 8-K filed with the SEC at the end of September, Violin disclosed that Kevin had approached the company’s board of directors and asked to take his entire salary and bonus in VMEM shares rather than cash.  Due to minimum wage laws in Santa Clara, CA, Kevin was required to at least receive $10.15 per hour.

Shortly after Kevin agreed to take his salary in shares, VMEM’s COO Ebrahim Abbasi (who worked with Kevin at Redback) asked the company’s board if he too could take his salary and bonus in VMEM shares.  Kevin and Ebrahim – and other VMEM executives – in recent months have also purchased shares in the open market.

We think VMEM’s management team sees the same value in the company we do – and that VMEM executives are intent on maximizing the pay-out should the Violin story play out as we anticipate.



The author of this posting and related persons or entities ("Author") currently holds a long position in this security.  Author may buy additional shares, or sell some or all of Author's shares, at any time.  Author has no obligation to inform anyone of any changes to Author's view of VMEM US.  Please consult your financial, legal, and/or tax advisors before making any investment decisions.  While the Author has tried to present facts it believes are accurate, the Author makes no representation as to the accuracy of completeness of an information contained in this note.  The reader agrees not to invest based on this note, and to perform his or her own due diligence and research before taking a position in VMEM US.  READER AGREES TO HOLD AUTHOR HARMLESS AND HEREBY WAIVES AND CAUSES OF ACTION AGAINST AUTHOR RELATED TO THE NOTE ABOVE.  As with all investments, caveat emptor.

I do not hold a position with the issuer such as employment, directorship, or consultancy.
I and/or others I advise do not hold a material investment in the issuer's securities.


1.  This quarter's (Q1 FY2016) results show large year-over-year sales growth against a weak comp

2.  Sales significantly exceed consensus expectations

3.  Extremely-bearish sell-side shifts its view as company clearly demonstrates it is on road to profitability and stand-alone viability

4.  Eventual possible take-out

-1       sort by    
      Back to top