Oracle trades for less than 8x my estimate of the next 12 months’ levered free cash flow, net of cash.  EBITA has grown at a 20% annual rate the last five years.  EBITA grew 11% in FYE 5/31/09.  Management is shareholder conscious, and openly states that they will actively seek to prevent free cash flow per share from ever declining through buybacks, acquisitions and expense management.

The Business

72% of new software license revenue is database and middleware software.  Database software is used to collect, store and enable access to data across an enterprise.  For example, a retailer can use database software to collect sales and inventory data from stores and relay the data to headquarters.

28% of new software license revenue is application software.  Examples include supply chain management and CRM applications. 

One-time license fees are charged for new licenses.  Thereafter, recurring fees are charged for updates and product support. 

The hardware business was acquired when Oracle bought Sun Microsystems on 1/27/10.  Oracle paid $5.4 billion for Sun.  At the time Oracle stated that Sun would generate $1.5 billion of cash pre-tax in the first year and $2 billion in the second year.  Sun cash pre-tax was closer to $2.8 billion in the 5/31/11 fiscal year, and $1.8 billion in the first nine months of the 5/31/12 fiscal year.  The acquisition is viewed by some as a failure because many of the legacy Sun server products are likely in terminal decline.  I think management would say that they are intentionally exiting commodity hardware products.  At any rate, the $5.4 billion price was relatively immaterial, and Sun has likely earned back most of the cost after two years.

The hardware competency that came with Sun enabled Oracle to develop an integrated database/server and middleware/server product, Exadata and Exalogic, that will be doing $1 billion of run-rate revenue exiting the 5/31/12 quarter (growing 100%+ year over year).  According to management, Exadata and Exalogic have gross margins closer to the margins of the software business (software is bundled with the hardware) than the legacy hardware business.  Exadata and Exalogic are high-performance server products designed for users with intensive data processing needs like insurance companies, they do not compete with commodity servers. 

If you assume half of R&D is related to new license fees and half is related to product updates, and sales & marketing is the “cost of sales” of new license fees, here are the contributions to pre-G&A earnings:

Updates & Support       75%

New Software Licenses  12%

Hardware                       7%

Services                         6%

Competitive Threats

One fear is that cheaper “big data” database applications will destroy Oracle’s database business.  One example is Hadoop, an open-source database application.  Hadoop is used by Facebook and YouTube because it is better suited for unstructured data like user-generated posts and videos that don’t fit a predictable category like February revenue for store #753.  Also structured databases like Oracle aren’t adept at processing very large data sets.  My thinking about this threat is that it’s true that Oracle won’t participate in the big data market, Facebook will never use Oracle.  However, I think most companies don’t have the engineering expertise to run an open source application that Facebook does, and those companies will always want full-service support for their mission critical database software.  Oracle has a near 100% attachment rate for support services.  Also I’m not aware of any situation where open source has gained significant penetration in an enterprise setting.  I think enterprise customers prioritize reliable support, regular security patches, and application stability by a wide margin over cost.  Hadoop has also been around since at least 2008, so this isn’t a new threat.  It’s also worth remembering that Oracle’s customers have critical data stored in their databases, and the risks inherent in switching are high.

Gartner estimates Oracle’s database revenue market share at 48% and growing, more market share than the next five competitors combined, and 2.5x the next largest, IBM.

Management

In 1996 Larry Ellison formed a group to take over Apple and reinstate his friend Steve Jobs as CEO, but Jobs decided to sell NeXT to Apple instead:

Ellison: “If we don’t buy [Apple], how can we make any money?”

Jobs: “You don’t need any more money.”

Ellison: “Well, I may not need the money, but why should some fund manager at Fidelity get the money?  Why should someone else get it?  Why shouldn’t it be us?”

Since 1993 Oracle has deployed 100% of its free cash flow into buybacks, acquisitions and dividends.  Over the last seven years Oracle has spent 75% of FCF on acquisitions and 20% on buybacks (net of option exercises) and 5% on dividends (first dividend was declared in March 2009).  For the last seven years Oracle has maintained net cash averaging 30% of EBITA.  The net cash position is 85% of EBITA today.  Oracle would have to spend 110% of free cash flow on acquisitions and/or buybacks to end FY2015 with 35% net cash/EBITDA.  Spending 110% of free cash flow would require about $60 billion of acquisitions and buybacks over the next three years.  It seems like an unmanageable amount of cash.  I think this bloated cash position is less likely to persist at Oracle than at other tech companies.  Oracle is a relatively active acquirer.  Larry Ellison is also exceedingly greedy, so I think he understands that building cash at negative real rates isn’t going to help him overtake his nemesis Bill Gates on the Forbes list.  I think he could overtake Carlos Slim if he can get the stock to $70 in a reasonable case.

For the record, I realize how stupid I sound talking about a cash/EBITA ratio.  I’d love to see Oracle achieve net leverage, it’s a very leverageable cash flow stream.  But we have to operate within the confines of reality.

Here’s my analysis of Oracle’s major acquisitions in recent years.  EV is market cap less net cash.  NTM revenue is the sell-side consensus estimate for the 12 months post-acquisition for the targets.  Free cash flow is my own fully-taxed pro forma estimate that assumes Oracle integrated the acquisitions at similar margins to Oracle, which I think is a safe assumption since these acquisitions represent about 35% of Oracle’s revenue today and margins have consistently risen in this period of time.  In Sun’s case management explicitly says that it is now running at pre-acquisition Oracle margins, despite being a hardware business.  Also all of these companies came with 60-80% gross margins, except Sun which came with 42% gross margins.  In every case except Sun the companies were growing revenue at 10-20%.

So I think history shows that Oracle has made acquisitions at reasonable valuations.  Repurchasing stock at today’s valuations is also very sensible.

Model & Valuation

In this scenario that I believe to be a reasonable base case, Oracle could do $2.85/share in FCF in the next 12 months.  And we’re paying $23.50 net of gross cash (for which this model gives no credit via interest income).  So we pay 8.2x NTM free cash flow, even without any acquisitions or buybacks.  If Oracle spent $15 billion on acquisitions next year, that could add another $.30/share in earnings power, putting us closer to 7.5x free cash flow.  Doing $15 billion of buybacks would reduce the shares outstanding by 10%, adding $.28/share. 

Without doing any acquisitions or buybacks Oracle could do $3.35/share in 2015.  I think that could easily justify a $60 stock in two years if we add in the $12/share of cash they’ll have at the end of FY2014.  Let’s pretend that Oracle spends $50 billion on acquisitions at 10x pro forma FCF.  That would add another $1/share in free cash flow.  Then I think you’d get closer to a $70 stock.

In the downside scenario, even if new license fees, service revenues, and hardware revenues are down 15-20% as I believe they were organically during the worst quarter in FY2009 (they were down much less for the full year organically), I believe that pre-G&A earnings will grow 1% if update & support revenue grows mid-single digits organically as they did in FY2009.  Additionally I think Exadata/Exalogic revenues will grow no matter what in 2013 – it’s a matter of do they grow 50% or 100% I think.  Management would also likely accelerate the rate of buybacks as they did during FY2009, providing further support for EPS.  So even in the downside scenario I think FCF will grow.  So if we go into a recession of 2009 severity and Oracle does $2.50/share and we give that 8x and add $6/share in cash the stock could go to $26, 10% downside.