We're surely living in a very interesting era. I just finished a Skype video interview with VIA's marketing vice-president, Richard Brown (pictured holding the latest Isaiah processor, increasing current C7-M processor performance by 2-4 times for the same low power).

It was 9PM on the West Coast, 1PM the next day in Taiwan where the fabless semiconductor company is located. Although the video call is not of a grand video or audio quality, it sures gives an idea of things possible with faster Internet links (I was using just a home DSL) and higher performance computers (the call was hosted on my 2GHz Macbook!).

My first motivation (but certainly not the only one!) to speak with Brown was to get his reaction on upcoming start-up Montalvo Systems that recently got talked a lot in the press (that's a story that I recently wrote for the French leading tech site, 01net) and its plans to market a low power X86 chip for the mobile internet device space. His response: it's not for the faint-hearted!

VIA continues CPU march to platformisation... to profitability

VIA designs various kind of chips: from X86 CPUs to chipsets, networking, graphics; chips for the telecom sector (CDMA) and even ARM compatible parts.

Brown no longer believes a chipset only company can be successful today without CPU, graphics/3D, networking, etc... technologies/capabilities. Put it simply, it's a platform world. Companies are trying to put more functions (graphics, I/O, multimedia, networking) into the same die. Intel and AMD are now talking of multi-core chips/systems on a chip that will integrate both the CPU and the graphics engine and soon much more.

"One of the initial challenges with the chipset business is that you're never in charge of your own destiny. You very rely on the CPU vendor and supporting on the changes and technologies that CPU vendor mandates. For awhile, it was a very good business [...] Suddenly everyone else looked at the chipset business and said, Look it's fantastic! And you have Nvidia come in, ATI, etc... Which in the end, meant the market is being commoditised aggressively", said Richard Brown.

However, VIA is taking a different view on integration than Intel and AMD. Both Californian companies are pushing for more functions in the CPU. Instead VIA would prefer to see more things put into the chipset. Of course. Ultimately, the result is the same: one integrated platform, with multiple integrated circuits (ICs) but coming from one single vendor.

"It's a challenge for companies that do not have CPU capabilities [...] Does it really make sense for them to be in the chipset business. And I would argue that it's very very difficult. More and more, it's about the ability to provide a complete platform rather than a single discrete component in that platform"

VIA's transition from chipsets to CPUs, albeit a smart one, was not without difficulties. In the process, the Taiwanese fabless design semiconductor company that employs 2,000+ person worldwide became quickly unprofitable, on about a $600-700 million turn-over per year. During our conversation, Brown anticipates that VIA will move back to the black by year's end. We'll remind him later on his prediction.

Intel's next-generation UMPC platform (Menlow) and all the buzz around start-up Montalvo Systems validates VIA's strategy

Intel push in VIA's territory validates indeed the Taiwanese strategy in the low-power X86 market but also brings a formidable competitor. Brown speculates that it will also grow the market for such low power chips/platforms much more rapidly than if VIA was the only kid on the block.

"When we first moved into the CPU business, we pretty much decided very early on it didn't make sense for us to compete in the MHz and GHz wars. And we felt at some stage that there will be a transition away from pure performance measurement like GHz. And that happened 2-3 years ago and it's now gathering pace at an even faster rate [...] If you look in terms of what a mainstream product is, obviously for many years it was the desktop, but if you look in the future, the so called mainstream becomes notebooks and beyond that become smaller ultra-mobile devices maybe with PC phone connectivity integrated [...]

With our processors, we focus on low power, high level of integration. Because ultimately we believe THAT will be the mass volume market. Wether its for notebooks, mini-notebooks or phone type devices [...] By 2010, notebooks will be a bigger market than desktops [...] So we're positioned for future mainstream markets rather than the mainstream markets of the past".

As for Montalvo, the first reason I called Brown?

"It's going to be very difficult for any company, no matter how well financed or how good its engineering is to enter the x86 market at this stage. One, the market itself is very very mature. And two, it's not just the case of simply having a decent product. It's that you need to convince customers you're going to be there for the long term, you've got to have a long term technology roadmap, you've got to have strong foundry relationships so that you can build the parts, cost effectively and respond very quickly to customer demands. And I personally think its extremely difficult [...] One of the key barrier is the IP (intellectual property) barrier", added Brown.

VIA also sees growing demand for low-power CPUs in desktops and thin clients

"There is a growing demand for energy efficient solutions for desktops and thin clients [...] Right now, we are a very small player [...] We have certainly less than 2% of the overall market [...] If you look at the UMPC space, we do have design wins with people like Samsung, OQO, and overall 40 design wins [...] HP does use our CPUs in thin clients and also in the green PC they introduced in China. So we do have penetration amongst leading OEMs".