Intel last week bought for $884 million Wind River Systems, a venerable embedded operating system company — yet another of the chip giant’s recent forays into software. The reason for this purchase is both simple and grand — to help Intel vertically integrate and to further its Linux ambitions. Intel’s ultimate target with this purchase is Microsoft. It’s all about kicking Redmond out of the netbook business.
Netbooks are the big hardware success of 2009 and most are powered by Intel Atom processors. The problem with PC’s in general and netbooks in particular is that they aren’t very profitable for Intel campared to the good old days. Microsoft makes more profit from every Windows PC sold than does the PC manufacturer and LOTS more profit than Intel makes despite its massively dominant market share in microprocessors. And with Netbooks retailing under $400, compared to Microsoft Intel makes hardly any profit at all. So Microsoft has to die.
This is a huge change for Intel, which has for decades acted as Microsoft’s bitch, doing pretty much whatever Redmond demanded for fear of being written-out of the next Windows PC hardware spec in favor of AMD or even IBM. But that was the old Microsoft. The Microsoft of today isn’t nearly as powerful, whether they yet know it or not.
Netbooks and mobile handsets are the first products from Intel’s Linux strategy. But Intel’s Linux platform — called Moblin — can be extended to the desktop as a direct competitor to Windows 7 if the company finds success in the netbook space. Microsoft isn’t stupid, so of course there is a huge battle brewing.
Here’s the problem for Intel: the company lacks key software capabilities despite having more than 5000 software engineers on staff. This is true especially for system software.
Intel’s new strategy (now 3-4 years old) is to be a platform company rather than a processor company. To become a platform company, Intel needs both silicon and low level software to tie all those pieces together.
Think about Intel’s Centrino as a platform. Centrino has: 1) a low-power processor; 2) a wireless chip, and; 3) software to manage wifi/wireline connections. Intel makes more money selling these three parts together as the Centrino brand and less when they sell the individual pieces. So software to Intel is glue to connect chips together and maximize revenue.
New chips at Intel require more system software (firmware) than Intel ever anticipated. For example: Intel’s Active Management Technology (AMT) requires firmware to wake up chips and apply patches when the computer is sleeping. This is true for both virtualization and security. In recent months Intel has stumbled delivering some key platform technologies simply because the firmware wasn’t ready.
Now Intel is searching for the next wave of applications for non-PC (typically embedded) computing devices — autos, healthcare, gaming to name a few. All these non-PC based applications will require even more system software capabilities.
Intel’s efforts with the Atom processor are beginning to pay off. They have the right process technology to produce Atoms at a low price. But the market is not paying much for Atom and Intel is at risk of cannibalizing its traditional markets with OEMs designing Atom-based netbooks that consumers buy instead of more profitable notebooks. In this environment Microsoft is not helping, since Windows makes netbooks more expensive and limits the profitability of both the PC maker and Intel, which is essentially making a netbook-on-chip.
Moblin is Intel’s answer to Windows – an Open Source Linux entirely funded by Intel.
Moblin is also Intel’s answer to Google’s Android operating system. Since Android is also Open Source and free, Intel might have relied solely on Google to take on Microsoft. But Intel as a platform company is too strategic to rely on any third party, even Google — hence Moblin.
The Wind River acquisition is Intel’s $884 million acknowledgement both that Moblin is strategic for the company AND that the Linux campaign isn’t going as well as Intel would like. Wind River, as a major force in embedded software, will quickly move Moblin into a variety of non-PC devices while also giving Intel more of the low-level software expertise that it has so sorely needed. If it works, we’ll see Moblin everywhere in 18-24 months. If it works really well, we’ll see Intel challenging Microsoft on servers and desktops, too.
Intel CEO Paul Otellini is determined not to be anybody’s bitch.