Blog Post

Why Microsoft should just pack it in and buy Red Hat

Stay on Top of Enterprise Technology Trends

Get updates impacting your industry from our GigaOm Research Community
Join the Community!

Over the past 40 years Microsoft has built its mega-billion-dollar business around the Windows operating system and Windows applications. But as the company seeks to preserve its enterprise power in the cloud era, where Windows is not as huge a deal, here’s a bold proposal: Microsoft(s msft) should buy Red Hat(s rhat).

It’s a counterintuitive move for a company whose former CEO likened Linux to cancer, but here’s why I think Microsoft might want to do it. First, Microsoft is an enterprise software company that hopes to transition its strength in that segment into the cloud-and-mobile era. And for Microsoft’s new CEO Satya Nadella, who will orchestrate that transition, an acquisition of Red Hat would be a bold stroke. Talk about thinking outside the box.

Source: IDC

Microsoft is pitching Windows Azure as the public cloud of choice for businesses. A major stumbling block there is the fact that Red Hat Enterprise Linux, the de facto standard Linux in businesses, does not run on Azure. Why not? No one’s really saying, but my take is that Red Hat is dragging its feet because RHEL competes with Windows Server. Red Hat has no such compunction about putting RHEL on AWS, which does run RHEL, because Amazon is not an OS competitor per se — at least as the market is defined now.

The lack of RHEL on Azure is a big issue for Microsoft. Given that Microsoft has bent over backwards to make Azure hospitable to third-party OSes and tool sets (Ubuntu Linux and SUSE Linux run on it), RHEL’s absence is particularly glaring because when you talk enterprise applications from Oracle(s orcl) and other vendors, SAP(s sap) et al., RHEL is Linux (sorry Oracle Linux, you’re not there yet.)

Neither Red Hat nor Microsoft would comment on why RHEL won’t run on Azure but sources close to Microsoft said customers ask the company about Red Hat support all the time and it has to refer customers back to Red Hat. Their thinking is that nothing can pressure a company to action more than customer pressure, but still, no dice.

Red Hat is affordable

If you look solely at the numbers, Microsoft could do this deal easily. The market value of Red Hat stock is about $11.2 billion. Subracting its $1.3 billion in cash, Red Hat would cost $9.9 billion. A 20 percent premium would put the purchase price at more like $12 billion. As one of my Wall Street friends put it, since Microsoft has about $84 billion cash in the bank, it could do this acquisition with its eyes closed.

Antitrust hurdles are big but not insurmountable

Analysts, including Gigaom Research’s own MSV Janakiram, said no such transaction would pass regulatory muster. Taken together, Red Hat(s rhat) Enterprise Linux (RHEL) and Windows Server(s msft) would comprise 87 percent of the server operating environment market, according to IDC analyst Al Gillen — who was also dubious about my premise.

That number would give regulators pause if they define the market as server operating systems running inside company data centers. But if they were to factor in what’s running in Google(s goog) and Amazon(s amzn) Web Services and other clouds, the picture could be much different. And we all know that lawyers can redefine the affected market when it painting a prospective acquisition in a rosy, competitive light.

In that universe, Windows Server plus RHEL might not be dominant at all. According to figures from The Cloud Market, nearly 60 percent of all Amazon Machine Images (AMIs) run Ubuntu Linux, while 6.3 percent run Windows and 5.9 percent run Red Hat. And I’m pretty sure Google runs zero Red Hat Linux or Windows.

Source: The Cloud Market

And, if regulators can approve the US Airways and American Airline or Comcast and NBC Universal mergers, I think all bets are off. I’ve seen strange things happen in the software world, including Oracle’s acquisitions of Siebel Systems and PeopleSoft in blockbuster deals that consolidated much of the enterprise applications market.

Those transactions taught me to never say never. So go ahead, use the comments below to (politely, please) to let me know what you think.

Feature photo courtesy of Shutterstock user Jorge Salcedo

Note: this story was updated at 9:19 a.m. PST March 7 to reflect that most SAP applications do not run on RHEL.

34 Responses to “Why Microsoft should just pack it in and buy Red Hat”

  1. Alexandre Bordeleau

    Let’s have some virus in the cloud :P

    Red hat is already well establish in the cloud and vmware too. Mircorosoft should begin by releasing product that are not still in alpha state and call their beta R2 some time later. Windows 8.1 is a nightmare with driver and obviously we also have the common security hole that come pre bundle and free of charge in every Microsoft Product

    Have fun!

    Oh our hyper-v core edition is now stable, it crash once a week.

  2. adamwill

    (Red Hat employee here)

    Red Hat is, in descending order of importance:

    1) Its engineers
    2) Its sales relationships

    7632) The code (since it’s all open source anyway)

    If Microsoft bought Red Hat, it would only be buying about 3% of 1), because the other 97% would take their redundancy packages and the money they made on the acquisition and quit just as fast as we could mail in the resignation letter.

    It’d probably be able to keep more of 2) in the *short* term as not all of RH’s sales types are that ideologically involved. However, some of them certainly are, and they’d be in the stampede out of the door on day 0. A lot of the others, being sales types, are dispassionate self-interest types, and would probably be reading the tea leaves and sending out resumes in short order. Whatever personnel were left around, most of RH’s sales relationships are ultimately built on the quality of RH engineering, and even the dimmest customer would note *fairly* quickly that all the engineers had abandoned ship. RH’s sales are just about all done on a subscription basis, with fairly short term subscriptions; we have to keep proving that we’re a hot shot bunch of engineers with a sales department attached in order for our customers to keep renewing their subscriptions. No-one is very deeply locked in to Red Hat software. It’s not Windows XP.

    So…the large proprietary software vendor who buys Red Hat buys trash, if you see what I mean. This sort of ‘ideological engineer poison pill’ is perhaps not well understood by all market analysts, but it’s a large reason RH hasn’t been bought out so far.

  3. indexof

    I don’t think it is a good idea to buy RedHat. It certainly will kill off RedHat, but it will hurt Microsoft more. If microsoft becomes a Linux vendor, their enterprise customers will not want to go with Windows. It means microsoft is not confident with Windows and their propriotory model of business. Going with Linux opens oppertunities for other vendors (Ubuntu), because a lot of enterprise customer don’t like Microsoft name attached to Linux.

  4. Ahmed Mohamed Maawy

    Well… Azure is cloud, and God knows how that is hosted. Its not a server side OS. Windows as a server side OS still does something right for the big boys :)

  5. Hi Barb,

    I suspect you’re being provocative here, but I’ll bite. I’m not convinced there would be value for Microsoft. Red Hat’s business model is built around support subscriptions. FY ending Feb 2013 for RH was 1.33B. Microsoft’s FY earnings were 77B. Adding a billion in revenue isn’t anything to sneeze at, but I’d have to wonder about what the M&A costs would look like and what the actual revenue earned after the merger would be. As folks here have pointed out, many customers could consider a move to another linux distro.

    On the product side, I think it’s interesting to consider the Red Hat product portfolio. Customers can install open source versions of virtually the entire Red Hat portfolio and enjoy the same hardware/software compatibility, feature set and performance as Red Hat supported versions. There is a core of engineering talent employed by Red Hat for developing the next versions of product, but trying to aqui-hire open source developers would be akin to trying to drink water with a fork. Not many folks would stick around. The majority of Red Hat’s product portfolio has been released open source or is scheduled for release in the near future. If Microsoft wanted any of that IP they could take the open source code re-distribute for costs lower than the acquisition of Red Hat.

    If we consider only the Operating System – which I think may be myopic given the current public cloud offerings you are well versed in – Redmond has the engineering talent, they could build their own distribution. They could even build a distribution specifically for Azure IaaS and PaaS solutions, with a downstream open source distribution for folks to peruse and abuse. That would be a very interesting development. But I can’t see them purchasing Red Hat proper. Bridging cultures is the most complex part of any merger, in this case I believe the culture at Microsoft and the culture in North Carolina are completely incompatible. I’m not sure an acquisition would ever be successful.

  6. Jim Wilcox

    Microsoft must hedge against the risk of becoming a monopoly. If there’s anything the browser monopoly lawsuit taught Microsoft, it’s that MS must prop up its competition or the US DOJ will restrict the hell out of it, or even rip it apart… this is why MS invested in Apple and involved itself in improving Linux. The purchase would probably be allowed by the FTC, but it would not be worth the risk for MS.

  7. Redhat employees would love it. Maybe they would finally get paid more than hippy love feelings. Microsoft pays about twice as much per engineer. If Redhat wasn’t on a university campus sucking up kids, no one would work there for what they pay.

    Where is this statistic about installations coming from? There are a ton of them, but they are almost all CentOS. Ever tried getting their hugely expensive support from Redhat? There is a correlation to pay vs. performance and again they pay very low even for NC. It isn’t like the package maintainers are going to jump on the phone. It is better to contact them directly, and then why the heck are you paying some ‘3rd party’ for support?

    I do agree not supporting Azure is completely Redhat’s fault. Azure runs Ubuntu fine. While we are on Ubuntu, not that it or Debian are perfect… but at least they aren’t defeated by your own lazy admins. Step 1 of 99% of all Redhat admin installs is to disable selinux because its hard to work with. So…. why they heck would you ever pick that rpm spec file random external repo required crap fest that is Redhat? Ubuntu/Debian at least provide apparmor and Ubuntu admins are smart enough to keep it on.

    So sure, buy Redhat. It will make as much sense as most Microsoft mergers. After all, the smart thing would be to just hire the CentOS guys for a million a piece and runaway with what actually matters.

  8. Another thought, what use is either company to Microsoft if the open source community abandoned ship after the acquisition?

    Meaning, these are not “typical” enterprise software companies; that’s what I’ve been led to believe. Their value proposition is closely connected to the developer talent ecosystem that chose to align with a “movement” that they believe in deeply.

  9. Barb, you said “Given that Microsoft has bent over backwards to make Azure hospitable to third-party OSes and tool sets (Ubuntu Linux)”

    So, wouldn’t Canonical be a much better bet, instead of Red Hat? What’s you thinking here?

  10. i_love_aws

    Who cares about RHEL and SUSE in the cloud. Ubuntu is the leader because developers like it. Proof of that is the fact that Ubuntu is the leader even where RHEL is officially available, AWS.

  11. You should take the “Cloud Market” numbers with a large spoonful of salt.

    They don’t measure _running_ images, but rather all images that exist. That includes nightly test builds of Ubuntu pushed into the system by Canonical, and another big chunk is Bitnami doing roughly the same thing. (Look at the “by owner” chart.)

    But this metric really isn’t useful at all. If a million people are running from a given image, it’s only counted once — while a million images run by no one at all are counted as a million.

      • I wish I knew a better source, but I don’t think the really useful information is available to anyone outside of Amazon. I agree that that leaves us somewhat in the dark, but I don’t think that means that _bad_ numbers are more useful than no numbers. They’re still bad numbers and really can’t be taken to mean something they don’t.

        I certainly don’t doubt that Ubuntu is popular in EC2, but these numbers don’t really show the truth either way.

        • Joe Fernandes

          Thank-you mattdm. I’ve seen so many tech writers misinterpret & misuse these same statistics, it’s frustrating. Counting available AMI’s is not the same as counting actual running instances!

  12. I honestly don’t believe that 67% of servers are running Windows. How was this data collected? Because W3Techs did a study on the top 10 million (public) websites, retrieving OS-specific packet responses from each server and came up with: 67.4% Unix-based (38.6% Linux), and 32.6% Windows.

    But if you want to include all websites, IDC’s data (which is who you were citing) actually shows that in Q3 of 2013 only 50.3% of servers were running Windows, and 28% on Linux.

  13. Jeff Putz

    You are completely wrong. Red Hat does run on Azure. Google it (or Bing it, if you prefer). Just because it doesn’t have a pre-built image doesn’t mean it won’t run there. It’s bring-your-own-license. This is another one of your weird link-baiting “articles” subtly dissing Microsoft.

  14. dngreengas

    Don’t they already have a stake in Suse due to their relationship with Attachmate? Also, the critical success factor of RHEL is their ability to attract top talent. If Microsoft were to acquire them, they would lose much of their top talent due both to culture shock and the reputation that Microsoft has in the software space. With the inability to keep top talent, Red Hat would implode. The only reason to acquire them would be to destroy competition and that would definitely not be allowed.

    If, instead, Microsoft focused on the existing distributions that already run on Azure like Suse and Ubuntu, they could increase their influence in the cloud computing arena. Then they should purchase Xamarin and make Visual Studio and Windows Desktop the best solution for developing mobile applications that connect to Azure.

  15. sounds like a good idea, microsoft could become like IBM of the late eighties, too grand in itself to believe in the need to spread it’s wings, which is good, in that in has encouraged a whole lot of other companies to get a foothold ithey would never have enjoyed had mircosoft dug in early in that sphere, i suppose after missing the boat on the internet, it thinks the way Tiger woods did before the big fall out, untouchable, giving customers access to service, no matter where it comes from, is smart business in a world that is only a finger tip away.

  16. Ankit Upadhyay

    Interesting proposition to say the least and definitely out of the box. I am sure the headlines will follow that Microsoft doesn’t care about consumers and buying and becoming fully the enterprise company.