At Sun, It's Chips Ahoy!

Sun wasn't too chatty about its latest acquisition. When the company snapped up the remaining IP from fabless chip startup Montalvo Systems in late April, nary a press release was issued, nor a blog written. But despite Sun's large-scale acquisition of MySQL earlier this year, there's a very significant chance that the Montalvo purchase will mean more for Sun in the long run than its grasping of the database leader.

Montalvo was a complete debacle, from start to finish. The company rose out of the ashes of famous money-sink Transmeta, a company so flush with cash, for a time, that it counted Linus Torvalds among its employees. Transmeta was legendary for its low-power chips with radically different energy management ideas. But Transmeta was plagued from the start by the ridiculous capital expenditures needed to launch a new consumer processor. After releasing a few chips to market, which made their way into some unique, compact and expensive laptops, Transmeta wandered off into the sunset with an Intel lawsuit in tow.

While Transmeta isn't entirely gone, its hopes and dreams are, basically, dashed to pieces at this point. Thus, a portion of the company's management left a few years back to form Montalvo Systems. Montalvo hoped to build low-power x86 chips, and to do so in India with rented time on ultra-violet laser etchers. The idea was to remain fabless, and as such, the company only needed a relatively small amount of capital to float. Or, that was the theory.

But after three years and over US$70 million spent, Montalvo was a failure. Enter Sun, in April, with what is said to be a pocket full of change. For a song, Sun snapped up Montalvo's IP, and ostensibly, some of its brainy processor architects.

So when I met with a room full of Sun and Intel spokespeople on Thursday, there to discuss their happy, huggy relationship, my first question was naturally related to Montalvo. Was Sun going to be producing its own low-power x86 chips?

The official word was, “what?” Sun's representative was not aware of the acquisition. Intel's multiple spokespeople, however, were. The resounding reply from them was, no comment.

Of course, there should have been a third party in the room, as well: AMD. It's understandable that representatives from that company would frown upon sharing a room with Intel, but the question would have been no less relevant there.

Is Sun preparing to move into the x86 chip market? I'm going to speculate here, something I'm not really supposed to do, as a journalist. But, heck, none of the parties involved want to contemplate the possible answer, so I'm the only one in the room who can. Is Sun hoping for low-power x86 chips?

Unequivocally: yes. But I'll add a caveat here; Sun's interest in the desktop is non-existent. For the server market, there's also no use for this chip. Where this low-power x86 chip could be most useful, however, is in cell phones and mobile devices. Sun just can't help crowing about its mobile aspirations of late, a fact which is obvious when you see all the JavaFX stuff they've been showing off at JavaOne this year.

Mix this with the company's last IP firesale acquisition, Savaje, which created a Java operating system for mobile phones, and you've got a recipe for a full-scale mobile phone platform, designed, produced and programmed entirely by Sun's big-brained engineering teams.

Perhaps the real question I should have asked is, “Can Sun actually deliver such a product to the consumer marketplace?” My initial reaction to this question is... No. Probably not.

--Alex Handy