Montalvo Systems has been seeking a hoard of cash in its latest fundraising round, but the chip startup is dangerously close to running out of money, VentureBeat has learned.
Although it raised $73 million, the company has run into problems finishing a chip that will compete with Intel in the low-power portable devices market, a market where chip giant Intel has become increasingly competitive. The company grew to nearly 300 employees in Santa Clara and India, resulting in a big burn rate and forcing the company to go back to fundraising. (our past coverage).
“There are a lot of people and burning it quickly,” said one venture capitalist who received a pitch but declined to speak on the record. “No matter how much money you raise, it won’t last long. “
Reached by phone, Montalvo CEO Matthew Perry declined comment. At this exact moment, I’m not sure of Montalvo’s status but I have heard enough to be able to say as much as I am in this story. One source told me the situation is fluid.
If Montalvo is unsuccessful in saving itself, that says a lot. It’s increasingly difficult to challenge Intel in the $30 billion x86 (Intel-compatible) microprocessor business, as Montalvo tried to do. It also shows that creating complex chips in this day and age is not an easy feat, even with tens of millions of dollars and hundreds of engineers. And it provides further evidence that venture capitalists are shying away from increasingly expensive semiconductor start-ups.
The company attempted recently to raise more than $100 million from investors but apparently has come up dry, VentureBeat has learned.
Last month, Cnet News.com’s Michael Kanellos spilled the beans on Montalvo’s chip strategy. The company was creating chips that would compete with Intel and Advanced Micro Devices microprocessors. Montalvo was creating chips with a lot of cores, or processing brains, on a single chip. But in contrast to Intel, Montalvo focused on asymmetric cores. That meant that the cores were not all the same. Some handled simple chores, while others were to handle complex tasks. The architecture made sense to a variety of chip experts I’ve spoken with. One caveat: an insider told me that not every detail described in the News.com article was correct.
“Asynchronous designs represent a challenge,” said Dean McCarron, an analyst and microprocessor expert at Mercury Research. “They can deal with processing bottlenecks and improve power.”
The beauty of this arrangement was supposed to be that the simple tasks can be dispatched by the simpler cores, which consume less power. The chip thus only consumes as much power as is needed. That could have made it a lethal competitor to Intel, whose Stepford wife-style cores were simpler to manage but not quite so versatile. But Intel is racing ahead fast on manufacturing technology, which gives its chips a performance edge. And it is also concentrating on doing smarter low-power designs. Intel is about to launch a low-power chip called Atom that will power a new class of mobile Internet devices. Every year, Intel promises to make great strides with its portable computing technology.
Besides Perry, other Montalvo executives declined comment or did not return calls for comment for this story. Current investors include Bay Partners, NEA-IndoUS Ventures, U.S Venture Partners, Leapfrog Ventures, CMEA Ventures and Adams Street Partners.
This company was a rarity in its ambitions. It was the biggest of the microprocessor start-ups, probably on par with P.A. Semi. It also resembled the once-ambitious Transmeta, the low-power microprocessor start-up that raised a bunch of money but never lived up to its promise of dethroning Intel in laptop chips.
Perry is the former CEO of Transmeta, a once high-flyer whose own struggles supplied its building neighbor Montalvo with a number of engineers. (Transmeta launched its low-power chips, if you recall, at Villa Montalvo in Saratoga, Calif.). There are a bunch of other prominent figures on board and among the investors, including Vinod Dham, now a venture capitalist and the father of the Intel Pentium chip.
The company’s chief architect was Greg Favor, who was one of the key architects at Nexgen, which was acquired by Advanced Micro Devices a decade ago. The Nexgen team helped AMD become much more competitive with Intel. The team also featured Peter Glaskowsky, former editor of the Microprocessor Report; original Montalvo (then called Memory Logix) founder Peter Song; and former Nexgen engineer Mike Yamamura.
The reason Montalvo and other start-ups have tried to fight Intel is obvious. If a start-up can capture just 1 percent of the $30.8 billion x86 microprocessor market dominated by Intel, that’s roughly $300 million in sales, according to Mercury Research.
Transmeta, as well as many other Intel-compatible, or x86, challengers that came before it, found that it couldn’t run fast enough on the “tick tock” treadmill set by Intel. It had to eventually resort to litigation against Intel to recover some of its investment.
Maybe Montalvo was the last best hope to provide a competitor to Intel and AMD. Maybe it just wasn’t the right company to do that. AMD is certainly having a tough time itself. The bar is getting higher and higher. It’s kind of a sad state of affairs and it may be tough for all ambitious chip start-ups in the times ahead. Not just because of the recession. But because the bets are too big and the execution lacking.
How bleak is the state of competition? Dave Ditzel, the founder of Transmeta and former Sun Microsystems chip architect, recently joined Intel to work on chip architecture.