Feature Article By Jeff Gould
Copyright © 2007, Peerstone Research Inc, All rights reserved.

January 23, 2007

Oracle's Linux initiative has so far failed to make a serious dent in Red Hat's business or even in its stock price. Red Hat is actually worth slightly more today than it was when Larry Ellison launched his apparently not-so-scary RHEL clone the week before Halloween.

But it is a little early to conclude that we are living in the best of all possible worlds for Red Hat. True, the company's financial results for the November quarter reassured easily-stampeded Wall Street investors who panicked in the first days after Oracle's announcement. But the fact remains that Red Hat's stock is worth 25% less today than it was a year ago. This decline reflects fundamental concerns not about the immediate threat from Oracle but about the long term value of Red Hat's business model.

Why should a company that grew 45% in the last quarter, will almost certainly exceed $500 million in sales in the coming year, and has an enterprise value of $3.85 billion be worried about its business model? Because that model is intrinsically incapable of creating as much value as that of its largest competitors, that's why. Put simply, Red Hat is the most successful open source software company in the world, it has been a public company for nearly eight years, its software is used by hundreds of thousands of companies, and yet when all and said and done it is barely worth 3% or 4% of Microsoft's Windows operating system business. This is a problem.

Where did I get that number from? Well, that's easy to explain. Microsoft doesn't actually publish a sales number for its total operating system business, but it does publish separate numbers for its Client division (desktop Windows) and its Server & Tools division, which includes Windows Server and associated middleware plus the SQL Server database. To make a long story short, if you add up these numbers and subtract a reasonable figure for the non-OS products such as SQL Server, you get a result somewhere around $20 billion. Roughly speaking, that is the annual sales number for the entire Windows OS franchise, desktop and server, excluding apps such as SQL Server, Exchange and Office (if anyone has a better number they're welcome to send it along). Red Hat by contrast is currently on track to exceed $500 million in sales this year, which would put it between 2% and 3% of Windows sales.

We get a similar Red Hat to Microsoft ratio if we look at valuation rather than revenue. According to its financial reports for the last fiscal year (ended in June 2006), Microsoft earns about 45% of its profit from the Client and Server & Tools divisions. Using these divisions as a rough proxy for Windows, and assuming that enterprise value is proportional to profits, we can deduce that the Windows business is worth approximately 45% of Microsoft's total enterprise value of $279 billion (its market capitalization minus its cash on hand). If we knock that down a bit to account for SQL Server's contribution to Server & Tools, we still arrive at a valuation for the total Windows business well north of $100 billion. By any measure, Red Hat, despite the premium it gets for its roaring growth rate which is significantly higher than Microsoft's, is worth no more than 3% or 4% of Windows.

This gaps seems a bit larger than your usual David vs. Goliath scenario. It's more like Ratbert vs. Godzilla (please ignore any emotional or aesthetic connotations you might attach to these two innocent cartoon creatures, this is purely a size comparison).

Of course one might argue that this comparison is unfair since Red Hat is mostly a server business these days and that therefore it should only be compared to the Windows Server side of Microsoft, excluding the far larger desktop Windows business. All right then, let's do that. In the last fiscal year, making adjustments for SQL Server, I estimate that Windows Server and its associated middleware generated around $6.5 billion in sales, give or take a few hundred million. That would put Red Hat's current revenue run rate at about 8% of last year's Windows Server sales. Call it Catbert vs. Godzilla.

The problem for Red Hat here is not that no one is using its product. On the contrary, hundreds of thousands of companies are, including most of the Fortune 500. In fact, the Yankee Group recently estimated that Linux holds 15% to 20% of the corporate server operating system market, compared to 65% to 70% for Windows Server, and we know that Red Hat is by far the dominant provider of corporate Linux. How Yankee produced those numbers and how accurate they are I couldn't say. But let's accept them as in the right ballpark. This allows us to see Red Hat's problem all the more clearly. It is simply this: despite its high growth rate and undoubted leadership in the enterprise Linux category, the company's revenue and valuation are proportionately much smaller than its actual share of the server OS market. Indeed, if Red Hat's revenue was proportional to its market share, it would be heading for $2.5 billion in sales this year instead of a measly $500 million, and its enterprise value would be well above $15 billion, putting it in the same league as software industry giants like Symantec, Adobe or Computer Associates.

Why does the Red Hat business model lag so far behind the traditional closed source software model in its ability to create revenue and shareholder value? To put it bluntly, it's because Red Hat gives away revenue and value to other players in the IT ecosystem. The beneficiaries of the generous Red Hat value give-away obviously include the users who choose not to pay for its software. According to Marc Fleury, only 10% of the Red Hat user base actually pays for the product. The rest just download Fedora (including Google, apparently). Now the fact that so many individuals and organizations can get their hands on such a high quality operating system for free is a tremendous benefit for the IT ecosystem as a whole and is by any account the crowning achievement of the open source movement. Countless web sites and applications take advantage of – or even owe their existence to – the availability of Red Hat's free software. But this achievement not only doesn't translate into increased value for Red Hat, it actually depresses the company's ability to create value.

Another group that receives a de facto subsidy from Red Hat consists of the IT heavyweights like IBM, Intel and Oracle who profit from the commodification of the operating system layer of the enterprise stack. Customers who spend less on the OS have more to spend on things like server hardware, database software and – last but not least – expensive professional services. So it is hardly any wonder that Larry Ellison and IBM Global Services consultants trip all over themselves to recommend Linux! What the one hand giveth, the other taketh away.

While Oracle's in-your-face move into Linux distribution doesn't seem likely at this point to inflict major short term damage on Red Hat (see my previous post), it does perhaps point out where the ceiling lies on that company's growth prospects. With competitors like Oracle able to clone its products freely, it appears highly unlikely that Red Hat's revenue or value can ever exceed a small fraction of those generated by the Windows franchise. In the most optimistic of scenarios we might imagine Red Hat reaching 10% of Microsoft's operating system business, which would give the company a valuation of more than $10 billion. This would certainly represent a spectacular gain for investors who bought RHAT at today's prices. But it would barely suffice to bring Red Hat within slingshot range of the software giant from Redmond. And in fact no one on Wall Street today is actually betting that Red Hat will reach such heights in the foreseeable future (if they were, they would be willing to pay a lot more for the stock).

Whether or not Red Hat remains an independent company, the Red Hat distribution and enterprise Linux in general are here to stay, there is no doubt about that. Even Microsoft now acknowledges this fundamental fact, as the recent Novell deal shows. But it is becoming increasingly apparent that the commercial open source business model pioneered by Red Hat suffers from severe structural handicaps. Without the financial and strategic support of IT market kingpins like IBM, Intel, HP and Oracle – and now perhaps even Microsoft – enterprise-ready Linux would not be economically viable. It is proving itself an extremely effective platform for serving the long tail of applications where IT ecosystem members don't want to or can't afford to pay the cost of closed source software. But, contrary to the prevailing ideology on the subject, it doesn't look like commercial open source can thrive as an independent ecosystem in its own right.