ESilicon designs and manufactures custom semiconductor chips for leading electronics companies. The six-year-old company, based in Sunnyvale, Calif., ranked as #3 on the Inc. 500 in 2005. Inc. Technology spoke with Jack Harding, CEO of eSilicon, about what leaders of small and medium-sized businesses need to know about chip technology.

Inc. Technology: How do entrepreneurs prepare for tomorrow's technology today?

Harding: I think most people have a good idea of what their road map needs to look like but it's more of a question of execution. The bottom line is that it's just simply too complex and expensive for small or medium-sized companies to build chips of their own. It shows up in a couple of areas: One is complexity drives the need for more and more specialized skills. The generalist who made chips 10 years ago with two of his buddies is now surrounded by 30 people, each with a very deep and narrow expertise. And the cost of that is prohibitive. So, as a result, it's really aggregators, businesses like ours that can timeshare skills across multiple accounts, bring that value to a young company trying to plan for the future, and digest complexities that are beyond the normal marketplace today.

Inc. Technology: So you need to outsource? Is it impossible to do it yourself today?

Harding: Certainly there's exceptions to every rule. Some technologies don't necessarily improve with more complexity, although they're rare. But statistically, each chip costs more and more than the one before it, it takes longer to make it, the risks of failure are greater, and the specialized skill sets needed to execute are growing geometrically with the complexity of the chip technology itself. Not to make this a commercial for eSilicon, complexity in general is both the bane and opportunity for a young company. But the key is to focus on those things that you deem to be your core competency and outsource everything else that isn't going to be a differentiator.

Inc. Technology: Are manufacturers going to make devices smaller and smaller?

Harding: The reduction in size, or in other words, an increase in density, is inevitable for several reasons. One is the infrastructure to make denser and denser chips has already been built...the suppliers of the raw materials will basically only make the denser and smaller chip technologies, and they'll phase out the old technology, which has been the history of semiconductor business. Secondly, in order to get returns on investments, you need to reach the largest markets, and this usually means creating for consumers, too. Now we're talking about portable devices that are smaller, cheaper, higher-performance and require less power -- so the only way to achieve all these goals, which is often at odds a one another, is to have smaller and smaller geometries that consumes less and less power and cost less and less to make. Power consumption is one of the biggest issues for portable products, which is especially a challenge because of the convergence trend of multi-functionality like a cell phone that's also a MP3 player, Internet browser, camera, PDA, GPS and more.

Inc. Technology: What should a start-up company know before buying or building portable electronics in today's day and age?

Harding: If I could give one piece of advice is hardware alone is not a product and is barely a solution to anything. More and more the value is in the software, so if you intend to make a chip you better have the software as part of your package. It continually shocks who many young companies are funded to only make a piece of hardware and to leave the software development to their customers. To me this is a formula for disaster.

Inc. Technology: So, leave the hardware development to a partner.

Harding: Yes, and build your core competency and value-add around the software. I could probably draw out the seven basic architectures that constitute 95 percent of chips in the world...there's very little mystery. The integration and development of the chip should be left to people who do it all day long, like an aggregator such as eSilicon, if I may. The value add, the application knowledge must come from the customer. Because at the end of the day, if you fast forward five years, we're only going to be left with two things: one will be our application knowledge and the other will be our channel/brand...and everything else will be outsourced who do that element better and less expensively than any given company can.