Relax! Survivors from the treacherous waters of the past months do exist. A clutch of companies, including Red Hat and VA Linux Systems, was lucky to enter the public arena early enough to amass a hefty war chest. Having managed their resources wisely, these early entrant companies have the time to formulate more innovative business strategies to go beyond building larger and larger professional SWAT teams to aid and assist the Global 1000. Others, like the young, privately funded OpenCola, Indrema, or Intel spinoff DevelopOnline,have the luxury of formulating new Open Source-based business models far away from the scrutiny of the public markets.
In addition, a fortunate few went public during the fall of 2000. One of the most interesting of these is Transmeta, which has the temerity to compete head-to-head with AMD and Intel. What's more, not only is Transmeta taking on the giants, it is beginning to win the game.
Transmeta is a real pioneer in more ways than one. Its x86-compatible Crusoe chip solution is a combination of a Very Long Instruction Word (VLIW) engine and Code Morphing software (see "Why Transmeta has the killer chip in its corner," October 2000). Essentially, this industry-disrupting processor is a very intelligent VLIW software compiler surrounding a VLIW hardware engine. The difference is that "source code" for this VLIW compiler is an x86 executable module, which is "morphed" into the native VLIW instruction set of the Crusoe chip. As a result, x86 versions of Linux, Windows, and applications targeting these systems can run on a system built with a Transmeta Crusoe CPU.
In developing their innovative technology, the Transmeta team may have leaped forward to create something new on the business side as well. The company is providing design specifications and professional services free of charge to hardware vendors. It has based its revenues upon the sale of the VLIW CPU and the combined Code Morphing software, which is proprietary.
That's right, Transmeta is hailing back to the models from the mainframe in which the iron was sold for revenue and the software was preinstalled to run those giant systems. The one major differentiator is that Transmeta is also giving away the majority of the service necessary for integration of the chips. The business model uses what was learned in the past from the mainframe and ushers in a whole new generation of Open Source and proprietary solutions for hardware and software platforms.
This is a revolutionary approach to the design, implementation, and sales of a hardware/software solution targeting mobile and low-cost home-appliance computing. First, Transmeta pushes the envelope in chip design by offering a hybrid solution driven by proprietary software. As a result, its VLIW chip can be deployed to run either Windows or Linux and all of the accompanying software for these operating systems.
More importantly, Transmeta's found a way to embrace the Open Source development community, fully leverage its business, and protect its revenue stream. Playing the Open Source card for Transmeta is none other than Linus Torvalds, who is charged with slimming down Linux for optimization on low-cost home-appliance and laptop devices. In particular, the new Mobile Linux has a substantially smaller kernel-a mere 8 MB-along with other features geared specifically for mobile Internet appliances:
- The ability to run without a hard disk
- The ability to operate solely in flash
- A compressed Ram File System (CRAMFS) that reduces storage and flash requirements
- Advanced power management
I still have some doubts about giving away the majority of services necessary for integration without carving out another discrete revenue stream. Maybe they know something I don't. Perhaps with repeated installations, this service becomes less of a value-added proposition for key hardware OEMs. Perhaps Transmeta's commitment to Open Source design specifications makes better sense than charging at the beginning and then tapering off.
Nonetheless, the bottom line is that Transmeta intends to make money on selling microprocessor chips to large, qualified Internet-appliance vendors. And when that CPU draws little power, runs very cool, and thereby provides for very long battery life, a long line of vendors interested in mobile and home-appliance Internet computing starts to line up, including the likes of Fujitsu, Hitachi, and Sony.
| TRANSMETA BUSINESS MODEL |
- Sell x86-compatible microprocessors to large hardware vendors for low-cost Internet appliances for home and mobile use.
- Make design specifications open.
- Provide integration services free of charge.
- Develop Mobile Linux under the traditional GPL.
- Hold the critical code-morphing software proprietary.
|
Even more interesting is a deal that has gone down with AOL and Gateway. After Microsoft-in particular MSN-teamed up with Compaq Computer to break into the home Internet-appliance market with the iPaq, AOL needed to counter this move or risk losing eyeballs for all of that rich Time-Warner content, not to mention burgeoning instant-messaging traffic.
AOL responded by teaming up with Gateway to provide a touch-screen home Internet appliance dubbed the Touch Pad, which employs the Transmeta Crusoe chip, runs Mobile Linux, and uses Netscape's Gecko engine for Web browsing.
Transmeta is breaking new ground in both its technology and its solid business model. This model is respectful of the Open Source developer community and understandable to its corporate partners. The Open Source developers, led by Linus, are working on a cutting-edge OS for this new Internet platform. The corporate partners, in turn, are increasing their value by selling hardware optimized for an Internet appliance. Meanwhile, Transmeta makes money on the sale of chips combined with software. Everyone appears to be winning.
| 
