VIA recently announced that the company will undergo a substantial division restructuring, splitting itself into three new main sections: CPU Platform, System Platform and MCE (Multimedia Consumer Electronics). Whilst this announcement may have been brushed over by most, it's more interesting than it seems. Allow me to delve a little deeper into the world of VIA, to discover what has prompted this change and how the industry signals a questionable future for a company that was once synonymous with enthusiasts.
VIA used to be a major name in motherboard chipsets. There was a time where, for performance, you'd be left with single choices on either fronts: VIA for AMD and Intel for Intel. VIA's KT range used to be
incredible and helped the Athlon processors reach the lofty heights they exist in today. Who remembers their first KT100 Pro?
"A questionable future for a company that was once synonymous with enthusiasts"
How times change. First, the graphics-based upstart NVIDIA created its own range of AMD chipsets with nForce, and proved that it was very capable of producing world-class chipsets. Leveraging the success of its GPU cards, NVIDIA has created a very desirable platform in the SLI ecosystem. Now that AMD has acquired ATI, and is destined to use its matured chipset range, VIA's largest chipset sector seems saturated.
Obviously, Intel has its own successful chipset division and is busy promoting the complete platform solution
in the forms of Viiv, Centrino and vPro, in a bid to get everyone to buy their product range from top to bottom.
In this highly competitive market, VIA has become a company plagued by release delays and an ever disappearing market, problems which have forced them to compete in the budget sector. Whilst this large market is still highly lucrative, it's very cut-throat. In previous days, VIA was able to sell budget boards off the back of its high-end performance boards - the so-called 'halo' effect. Without the performance edge, this means the name of VIA isn't publicised across the internet how it once was, and has been left to only be silently voiced by its PR to the OEM market. When was the last time you saw a review on an enthusiast website of a VIA board?
The company recently announced it's CN800 series chipset
. In the press release this chipset is paired with the same VT8237R southbridge that has been used since AMD launched its Socket 939 processors. Two SATA 150 ports? You'll be forgiven for checking your calendar isn't Two-Thousand-and-Three
So the AMD chipset market has become something of a problem for the once-dominant force in the chipset world. With the launch of its own CPUs in the late 90s in the form of the EPIA and C-series, we saw VIA attempt to diversify away from core logic into processors. Through buying IBM's Cyrix division and IDT's Centaur team, the firm acquired a lot of CPU intellectual property, including a license to the x86 architecture standard.
Seeing a new competitor attempt to enter the marketplace, Intel reacted like many successful businesses - by attempting to litigate the competition out of its space. It's a case of "He who holds all the aces"
, and no one seems to understand that better than Intel, which still claims that VIA's engineers have illegally used technology patented by Intel to make CPUs.
Forging ahead, VIA has bought its way out of that problem through more IP acquisitions, and has since used the IP to create a whole new market sub-sector in the Small Form Factor (SFF) area, with the EPIA mini and nano-ITX lines - both of which can be counted as successes for the company. Indeed, even Intel seemed to give them a break back in 2003 by allowing
VIA to buy a licence for the use of its front side bus technology until 2007 (although VIA has since claimed that this is through until 2008, a new point of contention).
Oh, how the universe is inextricably linked.
Following its own successful restructuring, Intel is now demanding VIA leave the CPU arena so it can have its own way, but is waving a bus licence renewal on the end of a stick as some form of incentive. From Intel's point of view, removing VIA's mobile processors from the equation will allow it to jump in and expand further the potentially huge UMPC market with its PX processors, off the back of the UMPC initiative it launched with Microsoft last year. The payoff would be to allow VIA to continue to market its core logic chips with limited success - a big win for Intel.
"Business has long been a paper war"
With the threat of further litigation and no obvious way out, VIA could even be tempted by the prospect, but its decision is further complicated by the fact that VIA's sister company, HTC, is deeply embedded into selling various Windows Mobile products, which makes leaving the CPU arena a terminal prospect.
Business has long been a paper war, and Intel has seen a chink in the armour and gone for the jugular. The firm has had a chip on its shoulder ever since VIA 'forced' Intel to move to DDR SDRAM over the more expensive RDRAM on its Pentium 4 processors. Now, it seems, could be payback time.
So what does VIA do? Stay in a potentially lucrative CPU market but risk being sued out of existence by Intel's lawyers, or take the bait of the bus renewal, step away from the CPUs, and last another four years on core logic until Intel decides that it doesn't want to renew the licence again?
How about something different? Taking a leaf out of the AMD book, how about an all-in-one C7 CPU, Unichrome GPU combination (with accelerated HD decoding), interlinked on-die with HyperTransport? HTT is an open architecture and this concept has already historically been done by
, and even VIA itself in the MediaGX
. VIA's nano-ITX CPU is tiny, but coupled to a northbridge with S3 graphics, in a single chip solution, it could save space in UMPC and laptop/tablet models as well as putting VIA right on par with the big boys when it comes to mobile processors - a growing market.
That's if VIA can get away with navigating the minefield of CPU and chipset patents that are growing by the day. The last thing VIA wants to do is cause more trouble for itself and potentially infringe on a "northbridge-CPU-GPU" patent that AMD undeniably has or is working on.
The company still has its Optical, Telecom and Networking solutions, and this have proved somewhat successful; but they aren't exactly the huge money earning entities VIA needs. The S3 division has consistently proved to be sub-par compared to NVIDIA and
AMD's offerings in discrete solutions, and off the map compared to Intel's integrated graphics market share.
Yes, an unenviable future indeed.