RHT - Red Hat, Inc

[Palantir]

Red Hat is a provider of virtualization solutions to businesses and have had impressive growth over the last few years. It's about 10% off its all time high, so it might be worth a deeper look. What's interesting about their business model is that they distribute open source software, aka software anybody can copy/modify/do as they please with it. Red Hat however, makes money by selling services tied to the software packages. The reason I find this interesting is because I like to look for businesses that have a strong annuity stream, and with open source software, there are no major upfront costs for customers, but strong and repeating maintenance and service fees.

 

However, this is not really in my circle of competence (Not sure I have a circle of competence to begin with). Does anyone have any insight into their market position? If anybody has comments on my assessment of the firm, I'm happy to hear it.

[Guest rimm_never_sleeps]

this to me has always defied gravity. I could never figure out what the attraction was of company that basically provides service ontop of a free OS. It has always traded at a high multiple. Has it come down?

[Green King]

where did their earnings go ?

I don't see that big of a growth in book  value or dividend distribution.

[ItsAValueTrap]

*Disclaimer: I have no domain knowledge in IT for large businesses or in mainframes.

 

To me, it seems that Red Hat should be thought of as a services company.  The NYSE uses the Red Hat OS for their servers.  Because their trading servers are mission-critical, it's worth it for them to pay for Red Hat's support services even though they could potentially do it in-house cheaper (because it's open source and they can simply compile their own OS).

 

2- The valuation seems a little ridiculous.

[Guest valueInv]

I have held RHT for a long time. I sold most of my holdings in the 50s as I think it is too expensive. I would buy more again if it got cheaper.

 

It is a solid business with one of the widest moats in the technology. In fact, it is one of the few tech companies that I consider to have a solid moat.While they sell services , what people buy from RHAT is a "throat to choke" or "insurance" when things go wrong. When you are running mission critical applications on Linux, you can't afford to take the risk of things going down for extended periods of time. Thats where RHT comes in. Their services provide "hot fixes" for Linux if things were to  go wrong (among other things). A CIO risks losing his job if he is not able to get things back up quickly and hence, doesn't mind buying "insurance" from RHT.

 

They are a subscription business and have a renewal rate in the high 90% range. For this reason, there is very little volatility in earnings. Haven't checked recently, but they used to have 80% + marketshare in the paid Linux market. Not sure how their Java application server business is doing these days.

 

Moat: For deployment in enterprises, applications and hardware need to be certified. Otherwise, companies won't support them. For example, if you deploy Oracle on a free Linux version that you downloaded. Something goes wrong and you call Oracle up. They will refuse to provide support, you're on your own. Thats why certifications are important - they provide the insurance or throat to choke.

 

Companies will certify only on popular platforms, otherwise, it increases support costs greatly. Since RHT has 80% marketshare, few are willing to certify competitors' Linux versions as it is not worth it for them. This applies to both hardware and software applications . Without certifications, enterprise are not willing to deploy competitors' applications, so they cannot gain marketshare. So its a virtual lock in the paid Linux marketplace for RHT. Since the ecosystem is clustered around RHT, it is very hard for a competitor to break in.

 

There are some risks:

- Cloud companies like Google have developed their own Linux expertise and don't buy RHT. This trend may rise.

- Ubuntu is RHTs closest competitor and has been nipping at their heels. But the last time I checked, they weren't making much headway.

- Their Java business may not have a great future.

 

 

 

[Palantir]

This all is music to my ears. I'm definitely looking at them as a services company with a nice renewable fee model. I'm not too worried about large companies developing their own versions of Linux as I feel that the trend is always towards more specialization and companies would rather use the same cost to hire an outside guy.

 

Valuation ultimately depends on growth rate. If you feel the growth will continue, then it may be justified. They've grown at nearly 20% over the last few years. I'd say a growth rate of 10% in the model would be conservative and reasonable....

 

A GR of 10% gives me an IV of 62, whereas one of 15% gives me one of 85. I'll stick with the lower one, and given that I want to buy 40% below the IV, I'd set a buy target at roughly 44...

[Palantir]

So question for those more knowledgeable than me -

 

 

What exactly stops a current Red Hat customer who is using their OS to switch over to another support provider? That is, say I use Red Hat Enterprise Linux, but feel that Oracle can give better support, what stops me from switching?

[Guest valueInv]

So question for those more knowledgeable than me -

 

 

What exactly stops a current Red Hat customer who is using their OS to switch over to another support provider? That is, say I use Red Hat Enterprise Linux, but feel that Oracle can give better support, what stops me from switching?

 

Because there is a risk. Oracle's expertise in Linux is not as good as RHTs. RHT contributes a lot of code to Linux and hence, has better knowledge of its works. There could also be certification issues.

Third party hardware and software vendors are likely to cooperate more with RHT for support. Oracle competes with them in both hardware and software. If you have a all Oracle stack, it may make sense though.

 

Oracle has been trying to steal RHT customers for a while now. Judging by RHTs revenues, they have not succeeded.

[rkbabang]

I've been using Linux at work for over 10 years (before that I was using Sun Microsystems "Solaris" Unix) and in all that time at two different companies (Both large S&P500 chip companies) I've never seen anything other than Red Hat.  Our expertise is in semiconductors, I don't believe my company will ever start rolling its own Linux.  Red Hat offers excellent support and all the EDA tools just simply work on Red Hat and are only supported on Red Hat.  I can imagine a Ubuntu gaining some market share with small businesses who need to cut costs or companies who don't need 100% uptime, but for large design operations where we are paying millions of dollars per year in EDA tool licenses there is no substitute for Red Hat.  I've owned RHT in the past, but even with its moat, and I agree with ValueInv's assessment that it is one of the largest moats in tech, it always seems too richly valued to me. 

 

The one thing that gives me pause, and stops me from owning RHT, is that Sun Microsystems used to have as strong a lock on the design industry as Red Hat now does and that seemed like a large moat right up until it wasn't.  Linux could be run on cheaper, but just as fast, intel based PC hardware, and Solaris needed to be run on more expensive proprietary Sun workstations.  I think Sun open-sourced Solaris and had an Intel based version at the end, but it was too late.  Red Hat had an opening and exploited it fully.  Will the same type of thing happen to Red Hat someday?  It surely will, as nothing in technology lasts forever, but I don't see it happening soon, but I probably won't see it is happening until it is too late to profit or get out of RHT without a loss.  Investing in tech is hard, even for those of us who are in the industry.

[Palantir]

I've been using Linux at work for over 10 years (before that I was using Sun Microsystems "Solaris" Unix) and in all that time at two different companies (Both large S&P500 chip companies) I've never seen anything other than Red Hat.  Our expertise is in semiconductors, I don't believe my company will ever start rolling its own Linux.  Red Hat offers excellent support and all the EDA tools just simply work on Red Hat and are only supported on Red Hat.  I can imagine a Ubuntu gaining some market share with small businesses who need to cut costs or companies who don't need 100% uptime, but for large design operations where we are paying millions of dollars per year in EDA tool licenses there is no substitute for Red Hat.  I've owned RHT in the past, but even with its moat, and I agree with ValueInv's assessment that it is one of the largest moats in tech, it always seems too richly valued to me. 

 

The one thing that gives me pause, and stops me from owning RHT, is that Sun Microsystems used to have as strong a lock on the design industry as Red Hat now does and that seemed like a large moat right up until it wasn't.  Linux could be run on cheaper, but just as fast, intel based PC hardware, and Solaris needed to be run on more expensive proprietary Sun workstations.  I think Sun open-sourced Solaris and had an Intel based version at the end, but it was too late.  Red Hat had an opening and exploited it fully.  Will the same type of thing happen to Red Hat someday?  It surely will, as nothing in technology lasts forever, but I don't see it happening soon, but I probably won't see it is happening until it is too late to profit or get out of RHT without a loss.  Investing in tech is hard, even for those of us who are in the industry.

 

Great points. From my (amateur) perspective, it doesn't seem that Red Hat is dependent upon a technological edge but rather they're willing to give away stuff for free while selling themselves as a services firm. If people are willing to pay them....their business model seems more bulletproof than others.

 

I like companies that compete against their clients' own internal departments as it seems that it is natural to outsource that work to a specialized provider....think Paychex etc.

[Palantir]

Moat: For deployment in enterprises, applications and hardware need to be certified. Otherwise, companies won't support them. For example, if you deploy Oracle on a free Linux version that you downloaded. Something goes wrong and you call Oracle up. They will refuse to provide support, you're on your own. Thats why certifications are important - they provide the insurance or throat to choke.

 

Companies will certify only on popular platforms, otherwise, it increases support costs greatly. Since RHT has 80% marketshare, few are willing to certify competitors' Linux versions as it is not worth it for them. This applies to both hardware and software applications . Without certifications, enterprise are not willing to deploy competitors' applications, so they cannot gain marketshare. So its a virtual lock in the paid Linux marketplace for RHT. Since the ecosystem is clustered around RHT, it is very hard for a competitor to break in.

 

When you say "certify" or "certifications", could you clarify what you mean? Does it simply mean an agreement to support the product you're using?

[rkbabang]

Moat: For deployment in enterprises, applications and hardware need to be certified. Otherwise, companies won't support them. For example, if you deploy Oracle on a free Linux version that you downloaded. Something goes wrong and you call Oracle up. They will refuse to provide support, you're on your own. Thats why certifications are important - they provide the insurance or throat to choke.

 

Companies will certify only on popular platforms, otherwise, it increases support costs greatly. Since RHT has 80% marketshare, few are willing to certify competitors' Linux versions as it is not worth it for them. This applies to both hardware and software applications . Without certifications, enterprise are not willing to deploy competitors' applications, so they cannot gain marketshare. So its a virtual lock in the paid Linux marketplace for RHT. Since the ecosystem is clustered around RHT, it is very hard for a competitor to break in.

 

When you say "certify" or "certifications", could you clarify what you mean? Does it simply mean an agreement to support the product you're using?

 

I think the issue is support.  Software companies support their products only when being run on certain platforms.  You will not get much support if you roll your own.  The largest semiconductor EDA software company (the industry I'm most familiar with) is Synopsys.  Here is a list of Synopsys tools.  If you go through it you will notice that while many tools are supported on multiple platforms, all of them are supported on Red Hat.  You can't say that for Windows, SUSE Linux, or any other platform. 

 

Also I think many are underestimating the scope of the work it would take to put together and support your own version of Linux, and then to get all the tools running on it, and keep it running 24 hours a day 7 days per week.  I regularly run jobs that take 4+ days to complete, even an unplanned down hour in the middle of the night on the weekend would be a major problem.  Don't get me wrong, creating/supporting your own Linux wouldn't be too bad a task for a company of software engineers like Google to take on But for a hardware company that has few if any software people, it is just too far outside the circle of competence to be worth the investment. And much too risky to be worth the inevitable growing pains that getting it up and running would certainly involve.

 

EDIT:  I forget to mention that for any large company your custom Linux would have to be supported across multiple sites all over the world.  You wouldn't need just one Linux team, but an internal Linux support team located at each design center.  Where Red Hat is not only very stable to begin with, but well supported both by Red Hat itself, and by all of the software venders who support their software on the platform world wide.

[Guest valueInv]

Moat: For deployment in enterprises, applications and hardware need to be certified. Otherwise, companies won't support them. For example, if you deploy Oracle on a free Linux version that you downloaded. Something goes wrong and you call Oracle up. They will refuse to provide support, you're on your own. Thats why certifications are important - they provide the insurance or throat to choke.

 

Companies will certify only on popular platforms, otherwise, it increases support costs greatly. Since RHT has 80% marketshare, few are willing to certify competitors' Linux versions as it is not worth it for them. This applies to both hardware and software applications . Without certifications, enterprise are not willing to deploy competitors' applications, so they cannot gain marketshare. So its a virtual lock in the paid Linux marketplace for RHT. Since the ecosystem is clustered around RHT, it is very hard for a competitor to break in.

 

When you say "certify" or "certifications", could you clarify what you mean? Does it simply mean an agreement to support the product you're using?

Yes. So when you release a software or a hardware product, you test it with particular versions of the OS. For example, RHEL 5.0, Solaris 10.4,etc. Then, you "certify" that it works and that you will support it when something goes wrong. So, if you multiply the Linux companies with the versions they have released, this becomes a big problem vey fast. This will increase your testing and support costs and make for a logistical nightmare.

 

Remember, each Linux vendor makes slight changes to the OS and follows different release cycles. So if you are currently supporting 3 versions of your own product on say, the last 3 versions of RHEL and you use Oracle (and support 4 versions of it), you have a big problem if you want to add Ubuntu into the mix.

[mikazo]

I'm surprised that no one has mentioned CentOS yet. CentOS is essentially the free version of RHEL. However, I've heard that it takes much longer for essential patches and updates to RHEL to make it into CentOS. So on top of support, clients are paying Red Hat Inc. for fast security updates, etc.

 

I haven't looked at Red Hat Inc.'s financials yet, but I think I might after having read through this thread.

[rkbabang]

CentOS isn't afieated with Red Hat at all.  They just put together a bunch of open source packages to create a Linux distro based on RHEL.  There are tons of such projects.  Red Hat's free version is called Fedora.  There isn't anything special about CentOS I doubt you'll see enterprise customers leave Red Hat for any of these.  Although free distros are fun to play with at home.  I ran Fedora on my home PC for a number of years before switching to Ubuntu's free home version.

 

From: (List of Lixux distributions):

Red Hat Enterprise Linux-based

 

See also: Red Hat Enterprise Linux derivatives and Commercial products based on Red Hat Enterprise Linux

 

Distribution Description

 

Asianux - A distribution co-developed between Red Flag Software Co., Ltd., Miracle Linux Corp. and Haansoft, INC., focused on Chinese, Japanese and Korean support.

 

CentOS - Community-supported distribution that aims to be compatible with Red Hat Enterprise Linux without the inclusion of proprietary software.

 

ClearOS - Small Business Server. File, Print, Messaging, UTM, VPN.

 

Fermi Linux LTS - Based on Scientific Linux.[13]

 

Miracle Linux - Developed by Japanese software vendor Miracle Linux Co., Ltd

 

Oracle Linux - Supported by Oracle. Aims to be fully compatible with Red Hat Enterprise Linux.

 

Red Flag Linux - A distribution developed in China and optimized for the Chinese market. Based on Asianux.

 

Rocks Cluster Distribution - A distribution for building a High-Performance Computing computer cluster, with a recent release supporting Cloud computing. It is based on Red Hat Enterprise Linux but with extensions to support large multi-node heterogeneous systems for clusters (HPC), Cloud, and Data Warehousing (in development).

 

Scientific Linux - A distribution co-developed by Fermi National Accelerator Laboratory and the European Organization for Nuclear Research (CERN), which aims to be compatible with and based on Red Hat Enterprise Linux.

 

SME Server - Based on CentOS and targeting Small and Medium Enterprises.

 

TrixBox - A Voice Over IP (VoIP) distribution based on CentOS.

[rkbabang]

I don't think the danger to Red Hat will come from anywhere within the Linux industry.  I think the danger will be another paradigm shift which can't be predicted right now, away from Linux somehow maybe. 

 

The last paradigm shift was the industry moving away from proprietary Unix systems (proprietary hardware and software sold together as a package) and towards using open source software (Linux) running on commodity hardware (x86 PC hardware).  This shift wasn't a minor cost savings it was a gigantic cost savings, and because Intel passed everyone else in uP performance, it was a performance boost as well. Which is why you don't see any thriving proprietary Unix workstation companies today.  Right now I don't see anything coming down the pike to replace this current Linux/x86 setup for scientific and/or engineering workstations.

[ItsAValueTrap]

What's uP performance?

[rkbabang]

Sorry, uP was just me abbreviating microprocessor quickly.  "u" as a short way of typing "μ"(the Greek letter Mu) for micro and "P" for processor.

[Palantir]

Red Hat has drifted down from an all time high of 60 to 47.5. I plan on initiating a position at 45 or below....this firm seems to have a really good model confirmed by some of the knowledgeable posters on this rhread.

[Palantir]

I am long Red Hat at about 48.36. I put in a stop order at about 46, in case this thing goes to hell, at which point I will try to reinvest at around 42. But basically, I think this is undervalued at a price of 50 or lower, with a strong market position, niche, and high switching costs.......lets's see how it goes!

[writser]

Please, never use stop orders..

[compoundinglife]

I'm surprised that no one has mentioned CentOS yet. CentOS is essentially the free version of RHEL. However, I've heard that it takes much longer for essential patches and updates to RHEL to make it into CentOS. So on top of support, clients are paying Red Hat Inc. for fast security updates, etc.

 

I haven't looked at Red Hat Inc.'s financials yet, but I think I might after having read through this thread.

 

RedHat has a packaging system call RPM, which is defined format for packaging up software for redhat (and other) systems. RPMs usually have two versions you can download the compiled binary program you want in an RPM package, and then a "source" RPM which has the source code for the software and all the meta data required to compile and build a binary RPM.

 

CentOS basically takes the source RPMs from redhat and builds a redhat clone. However it is community supported meaning you get support from mailing lists and forums that include other CentoOS users and usually some of CentOS developers/maintainers.

 

CentOS and Debian are probably the two most popular non-commercial linux distros on the server side and Ubuntu is probably the most popular on the desktop side. However Ubuntu is gaining in the server and cloud space. Ubuntu has a slightly different model in that they provide the OS AND updates for free and you just pay for support if you want it. The last time I used RHEL you could install it but you would never get an update without a license.

 

I don't think CentOS is a threat to RHEL. Most of the enterprises and start ups I know of are either in one of two camps 1. They want support and buy RHEL or 2. They refuse to pay for a license to use a linux distro and they use either CentOS, Debian or Ubuntu. There are a lot of businesses that use free Linux distros for production systems that will probably never switch to RHEL. RedHat's current customers will never leave Redhat to go to CentOS.

 

At this point I would not rule out Canonical (makers of Ubuntu) as a competitor. Most of the sales in the past have been for suppor licenses on Ubuntu desktop but they have been actively trying to push their server/cloud distro. I think part of the problem Ubuntu suffers in the server space is that there are a lot of companies with highly technical sysadmins and staff who refuse to pay for support on a Linux distro so they use Ubuntu or Debian but Canonical never sees a dime.

 

I think some companies that buy RHEL are starting to consider Canonical/Ubuntu as a potential option, however I think at this point Canonical as a company is not quite polished enough to outsell RHEL in most large existing RHEL customers. As previously mentioned if you are using off the shelf commercial apps support for anything other than RHEL is going to spotty.

 

FWIW, I have worked and consulted at many places that run large highly available commerce sites, HPC clusters and other kinds of large scale systems, most of the ones I have worked with do NOT use RHEL and instead use CentOS, debian or Ubuntu. In fact some really large companies you might find it differs by the org.

 

The only jobs or gigs where I have actively advocated purchasing RHEL is when we had to run Oracle. Oracle does not support any of the free distros (at least the last time I checked, its been a few years) and you can get it to work on some of them but its a nightmare. Now that Oracle has its own OS and hardware I see that as potential threat to RHEL. But I should not be considered as a data sample of the average enterprise customer.

 

 

 

[compoundinglife]

So question for those more knowledgeable than me -

 

 

What exactly stops a current Red Hat customer who is using their OS to switch over to another support provider? That is, say I use Red Hat Enterprise Linux, but feel that Oracle can give better support, what stops me from switching?

 

There is brand loyalty among System Engineers and System Administrators. These are people who argue over what the best terminal based text editor, see http://en.wikipedia.org/wiki/Editor_war . People are generally very distro loyal and while all the distros run the "linux" kernel they are all setup somewhat differently and some of the core things you do such installing software/patches, building customer software packages vary from distro to distro.

 

As ValueInv mentioned, RedHat's support is like insurance for a lot of IT managers. Linux based operating systems are great, but when something doesn't work trouble shooting for the average IT person can be very hard. Also companies like RedHat tend to make it easier to integrate their systems with things like Active Directory and other systems that exist in the enterprise.

 

For the average enterprise running Linux switching is a big deal. You have already tested and proved that all of your internal systems and sofware work on this distro, moving can be a major pain and take a long time. However I am talking about large non-tech companies. Companies like Amazon, Google, Facebook and Yahoo have people on staff that are Linux kernel developers and they would get no value what so ever from the support of a company like Redhat and hence don't use it.

 

The other question is who would you switch to? There is Canonical (ubuntu), Novell (Suse) and now Oracle. Canonical is not a very Enterprise focused company. Oracle is fairly new in the game so it will take some time to see how that shakes out. Suse has a good customer base in Europe from what I understand and at one point it was the suggest platform for running SAP on linux. RedHat is the clear leader in commercial enterprise Linux.

 

Ubuntu could be a threat IMO if they really executed well and we shall see what happens with Oracle, I have not been following them too closely.

[compoundinglife]

I've been using Linux at work for over 10 years (before that I was using Sun Microsystems "Solaris" Unix) and in all that time at two different companies (Both large S&P500 chip companies) I've never seen anything other than Red Hat.  Our expertise is in semiconductors, I don't believe my company will ever start rolling its own Linux.  Red Hat offers excellent support and all the EDA tools just simply work on Red Hat and are only supported on Red Hat.  I can imagine a Ubuntu gaining some market share with small businesses who need to cut costs or companies who don't need 100% uptime, but for large design operations where we are paying millions of dollars per year in EDA tool licenses there is no substitute for Red Hat.  I've owned RHT in the past, but even with its moat, and I agree with ValueInv's assessment that it is one of the largest moats in tech, it always seems too richly valued to me. 

 

The one thing that gives me pause, and stops me from owning RHT, is that Sun Microsystems used to have as strong a lock on the design industry as Red Hat now does and that seemed like a large moat right up until it wasn't.  Linux could be run on cheaper, but just as fast, intel based PC hardware, and Solaris needed to be run on more expensive proprietary Sun workstations.  I think Sun open-sourced Solaris and had an Intel based version at the end, but it was too late.  Red Hat had an opening and exploited it fully.  Will the same type of thing happen to Red Hat someday?  It surely will, as nothing in technology lasts forever, but I don't see it happening soon, but I probably won't see it is happening until it is too late to profit or get out of RHT without a loss.  Investing in tech is hard, even for those of us who are in the industry.

 

I have been using Linux for around the same amount of time. I think the first time I installed it I was using Slackware in the mid the late 90s. I remember telling coworkers running Solaris systems that someday they would likely all need to learn Linux at some point to keep there skills marketable and they laughed at me :) After being in the game all this time, most tech companies end up going in the too hard pile. Your right even as an industry insider its hard to invest in tech. The only tech company I own at the moment is Dell because I felt it was far too cheap and I have MSFT in an account I manage for a family relative.

[Palantir]

Excellent posts above. I'm getting validation of my investment thesis.....I think that is a bias. Eh who cares.